6 days ago · 2b27ee6a34
--- a/cache/2024/463a5fd436bb887f90f9e01b01b7716b/index.html
+++ b/cache/2024/463a5fd436bb887f90f9e01b01b7716b/index.html
@@ -0,0 +1,370 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="en">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>What's next for Kagi? (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://blog.kagi.com/what-is-next-for-kagi#9">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>What's next for Kagi?</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://blog.kagi.com/what-is-next-for-kagi#9" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <p>Two years ago, on June 1st, 2022, <a href="https://blog.kagi.com/kagi-orion-public-beta">Kagi introduced</a> a search engine that challenged the ad-supported version of the web. Kagi Search instead works for you, the user, and not an advertiser paying for your attention. At the time of launch, we did not know if anyone would pay for their search engine and web browser, but luckily, here we are two years later at the forefront of a movement to humanize the internet’s most-used products and put the user back in the driving seat.</p>

 <p>This blog post is meant to provide an update to our community and future members on where we are and what our plans for the future are.</p>

 <h2><a name="0"></a>Our commitment to you</h2>

 <ul>
 <li>We provide the best search results in a clear and easy-to-understand way, with no ads, amplifying your knowledge.</li>
 <li>We are committed to your privacy. We allow no tracking or ads, and you remain in control of your data.</li>
 <li>We commit to a user-centric, “pay for the product”, business model with no other sources of revenue.</li>
 <li>We ensure access to timely and high-quality search information. Our refined algorithm ranks information by quality rather than by potential monetization.</li>
 <li>We serve our users’ needs with speed, efficiency, and accuracy.</li>
 <li>We are committed to providing leading-edge technological solutions in service of users.</li>
 <li>We maintain the integrity and quality of our system.</li>
 </ul>

 <h2><a name="1"></a>Growth and sustainability</h2>

 <p>Our journey over the past two years has been marked by continuous growth and improvement. We have expanded our paid member base (which is now <a href="https://kagi.com/stats">more populous than 14 world countries!</a>), introduced a <a href="https://kagi.com/changelog">slew of innovative features</a>, and maintained a steadfast commitment to <a href="https://kagi.com/privacy">privacy</a> and the <a href="https://x.com/vladquant/likes">quality</a> of the search experience.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716361973-858686-screenshot-2024-05-22-at-001246.png"></center>
 <center><em>Kagi member growth over time. Note the expansion after the $10/mo unlimited plan was introduced.</em></center></p>
 <p>We are also thrilled to report that we have achieved profitability. This significant milestone is a testament to our sustainable growth and fiscal responsibility. It demonstrates that our approach of offering a premium, ad-free search experience resonates with users who support a service aligning with their values. Becoming profitable allows us to reinvest in the business, further enhancing our offerings and ensuring that we can continue to provide a top-notch search experience.</p>

 <p>We are grateful for the support of our users who have played a crucial role in reaching this milestone. Together, we will continue to build a better, more user-centric web.</p>

 <p>As we look to the future, our focus remains on sustaining this momentum. If we continue on the path of current growth, which so far has been completely organic, we will reach 1 million paying users in the next 36 months.</p>

 <h2><a name="2"></a>Increased recognition</h2>

 <p>Kagi is getting noticed by experts like <a href="https://twit.tv">Leo Laporte</a>, <a href="https://www.404media.co/friendship-ended-with-google-now-kagi-is-my-best-friend/">Jason Koebler</a>, <a href="https://daringfireball.net/linked/2024/01/19/bray-google-kagi">John Gruber</a>, and <a href="https://pluralistic.net/2024/04/04/teach-me-how-to-shruggie/">Cory Doctorow</a>, as well as high quality media organizations such as <a href="https://www.theverge.com/23896415/kagi-search-google-meta-quest-3-chatgpt-macos-sonoma-installer-newsletter">The Verge</a> and <a href="https://www.theregister.com/2024/01/30/ai_is_changing_search/">The Register</a>, publications like <a href="https://monocle.com/radio/shows/the-entrepreneurs/eureka-389/play/">Monocle</a> and <a href="https://www.zeit.de/index">Die Zeit</a>. We thank them for their coverage and helping spread the word.</p>

 <p></p>
 <p><center><a href="https://monocle.com/radio/shows/the-entrepreneurs/eureka-389/play/"><img src="https://kagifeedback.org/assets/files/2024-05-22/1716357691-215959-screenshot-2024-05-21-at-230124.png"></a></center>
 <center><em>Recent interview in Monocle radio</em></center></p>
 <p>Additionally, we are grateful to content creators like <a href="https://youtu.be/YnSv8ylLfPw?t=523">Linus Tech Tips</a> and <a href="https://www.youtube.com/watch?v=K297opaBp8E&amp;t=7s">Ask Leo</a>, numerous <a href="https://dannb.org/blog/2023/how-kagi-beats-google/">bloggers</a>, and the vibrant discussions by Kagi members on Hacker News, X, Mastodon, Reddit, Blue Sky, LinkedIN and other social media for helping <a href="https://x.com/vladquant/likes">spread awareness</a>.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716359315-445982-screenshot-2024-05-21-at-232829.png"></center>
 <center><em>Public comment from a member on Reddit 🙏</em></center></p>
 <p>Support like this not only means a lot to us but has been the main driver of Kagi adoption so far, and for that we are forever thankful!</p>

 <h2><a name="3"></a>New shareholders and team members</h2>

 <p>We are excited to announce that so far, 93 Kagi members have joined Kagi as investors through our two investment rounds, most recent closing a few weeks ago. The investment not only brought in capital (which we will use for hiring, R&amp;D and search index expansion) but also strengthens our community-centric approach, as these investors are also users who deeply believe in our mission. The rest of the company remains owned by the founder, employees, and advisors, ensuring that those who are directly involved in building and sustaining Kagi have a significant stake in its success.</p>

 <p>This diverse ownership structure fosters a sense of shared purpose and collaboration and aligns the incentives of everyone involved. Having users as investors provides invaluable insights and feedback (and sometimes key leadership talent!), helping us better understand and meet their needs. It also reinforces our commitment to transparency and accountability, as we are directly answerable to a community that is both invested in and supportive of our long-term vision. While we have no immediate plans to raise additional funding, if you are interested to be notified about any future fundraising, <a href="https://forms.gle/1Try2v6JtXbKSjKx9">get in touch</a>.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716360672-296077-screenshot-2024-05-21-at-235107.png"></center>
 <center><em>Kagi team members around the globe</em></center></p>
 <p>Our team has expanded to 37 people around the globe, working together to bring a friendly version of the web to people, families, and organizations worldwide. We are actively <a href="https://help.kagi.com/kagi/company/hiring-kagi.html">hiring</a>.</p>

 <h2><a name="4"></a>Kagi becomes a Public Benefit Corporation (PBC)</h2>

 <p>The Delaware public benefit corporation (PBC) is an entity created by amendment to the Delaware General Corporation Law in 2013, and is structurally identical to the familiar Delaware C corporation with one major difference: company directors and officers are given explicit permission to consider a purpose beyond simply maximizing shareholder value.</p>

 <p>Kagi transitioned to a Delaware Public Benefit Corporation (PBC), marking a significant milestone in our journey. This structural change allows the legal entity to consider a purpose beyond simply maximizing shareholder value, balancing it with the best interests of all stakeholders, including the communities we serve, along with the specific public benefit.</p>

 <p>In particular, the public benefit statement entered into Kagi’s charter is: “<strong>Kagi is committed to creating a more human-centric and sustainable web that benefits individuals, communities, and society as a whole, with a transparent business model that aligns the incentives of everyone involved.</strong>”</p>

 <p><img src="https://kagifeedback.org/assets/files/2024-05-22/1716360472-761510-screenshot-2024-05-21-at-234747.png" alt="Kagi PBC"></p>

 <p>This transition ensures that our mission-driven approach is not just a guiding principle but a legal obligation, protecting us from external pressures that might otherwise divert us from our path. The PBC designation also resonates with our community of users, who value ethical business practices and corporate responsibility. It provides a clear signal that Kagi is dedicated to making a positive impact on the world, aligning our business practices with the broader goal of societal benefit.</p>

 <p>Some examples of companies structured as PBC  out there include Laureate, Plum Organics, Kickstarter and Anthropic. We’d like to thank <a href="https://en.wikipedia.org/wiki/Eric_Ries">Eric Ries</a> for introducing us to the PBC structure.  You can read more about it <a href="https://kagi.com/search?q=public+benefit+corporation+PBC&amp;r=us&amp;sh=FuSMLpCLQlOTugxwnnQo6g">on Kagi</a>.</p>

 <h2><a name="5"></a>Kagi’s unique approach to search</h2>

 <p>From day one, Kagi Search was built with one goal in mind — show the world that a high-quality, user-centric search engine is possible. Why was this not possible before? In short, in every market economy, there has to be a customer. For search, that customer can be either you or an advertiser, and this simple choice determines all incentives and outcomes in the search experience. With Kagi, for the first time in history of search engines, you are the customer and everything is built around you and your needs alone.</p>

 <p>And we know what our customers want because it is the same thing we want: no ads, no spam, no low-quality content, and no insult to our intelligence from the search engine we use. We want results that include human discussions, not LLM-generated content. We value opinions and expertise from personal websites and blogs over ad-filled, and tracking-infested content silos.</p>

 <p>We want to explore and discover the web, versus being told what to buy. We want to connect to and understand other humans and their work, not be consumers of the content created with the only purpose of monetizing our attention and wasting our time and brain cycles in the process.</p>

 <p>And this is what Kagi strives to deliver.</p>

 <p>For instance, when you ask your search engine a programming question, you want to see official documentation, industry experts, personal blogs, and other authentic sources of information.</p>

 <p></p>
 <p><center><a href="https://kagifeedback.org/assets/files/2024-05-22/1716392268-405998-screenshot-2024-05-22-at-083623.png"><img src="https://kagifeedback.org/assets/files/2024-05-22/1716392268-405998-screenshot-2024-05-22-at-083623.png"></a></center>
 <center><em>User-centric vs ad-supported, click to enlarge</em></center></p>
 <p>Similarly, searching for “best headphones” should yield genuine recommendations, not headphone ads from the highest bidder. What you often find in Kagi is information from people who are truly knowledgeable and passionate about their topic.</p>

 <p></p>
 <p><center><a href="https://kagifeedback.org/assets/files/2024-05-22/1716392725-691336-screenshot-2024-05-22-at-084426.png"><img src="https://kagifeedback.org/assets/files/2024-05-22/1716392725-691336-screenshot-2024-05-22-at-084426.png"></a></center>
 <center><em>User-centric vs ad-supported, click to enlarge</em></center></p>
 <p>Whenever possible we want to make sure that facts shown in Kagi are accurate and computable. This is why we <a href="https://blog.kagi.com/kagi-wolfram">integrate with Wolfram|Alpha</a>.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-03-05/1709654440-735786-screenshot-2024-03-05-at-080035.png"></center></p>
 <p>We are not trying to influence users’ search results; instead, we aim to provide users with tools and features to <a href="https://help.kagi.com/kagi/features/website-info-personalized-results.html">personalize</a> their search experience according to their preferences. For example, with Kagi, users can <a href="https://kagi.com/stats?stat=leaderboard">block sites</a> they do not like and promote those they prefer.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716393012-475138-screenshot-2024-05-22-at-085004.png"></center></p>
 <p>Use <a href="https://help.kagi.com/kagi/features/lenses.html">Kagi lenses</a> to narrow down searches to the academic realm or one of the many predefined lenses - or create your own search lens with your favorite sites.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716393399-721245-screenshot-2024-05-22-at-085620.png"></center>
 <center><em>Search results with “Academic” lens enabled</em></center></p>
 <p>The continuous shipping of features and improvements has been largely driven by our active member community and would not have been possible without it. The Kagi Search <a href="https://kagifeedback.org">feedback forum</a> currently hosts over <strong>3,500</strong> discussions, while the Kagi’s Orion Browser <a href="https://orionfeedback.org">feedback forum</a> boasts over <strong>6,400</strong> discussions!</p>

 <p>These massive public forums are among the most active web search and browser product communities on the web, reflecting the incredible level of engagement and valuable insights provided by our users. This is also why <strong>we do not have or need any telemetry in our products</strong> - we can barely keep up with what our users are directly asking us to do!</p>

 <h2><a name="6"></a>Orion - Web browser that has your back</h2>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-30/1717079482-339174-04.png"></center>
 <center><em>Orion is a lightweight Mac browser, with native support for horizontal, vertical and compact tabs</em></center></p>
 <p>Most people may not know this, but Kagi started developing <a href="https://kagi.com/orion">Orion browser</a> even before the search engine (and Orion’s feedback forum has almost twice as many discussion than the one for Kagi Search!). It is clear to us that no successful search engine could thrive without a companion web browser built on the same set of core values, including the alignment of all incentives.</p>

 <p>We are pleased to see so many people excited about the prospect of Orion and willing to <a href="https://kagi.com/orion/orionplus.html">pay for it</a>, which may sound like an even more “outrageous” idea than paying for search.</p>

 <p>But think about it: a web browser is one of the most important and intimate tools we use on our computers and devices every single day, yet we as a society, have somehow allowed most browsers in use to be directly or indirectly paid for and controlled by advertisers! It is like the situation with search, just worse. This is insanity to me and this is why we are building Orion, a user-centric web browser with a paid business model and no other sources of revenue, to bring sanity back and help support our mission to humanize the web.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-30/1717079301-260116-screenshot-2024-05-30-at-072813.png"></center>
 <center><em>Native support for web extensions, profiles, and much more…</em></center></p>
 <p>Orion is WebKit-based, built from scratch for high performance, zero-telemetry by default, <a href="https://kagi.com/orion/#privacy">truly privacy respecting</a>, with built-in ad and tracking blocking like no other web browser in the world. On top of that, we brought the same Web Extensions API used by Chrome and Firefox to WebKit, allowing Orion to run both Chrome and Firefox extensions natively out of the box (you can follow <a href="https://docs.google.com/spreadsheets/d/14IgSRVop4psUTgtLZlvYJYrAArhvL3WvRlUdzdQbIoQ/edit?usp=sharing">our progress</a>, with many extensions already working, including on iOS!).</p>

 <p>For the Mac fans out there, you will be pleased to know that Orion has ported the latest WebKit to macOS Mojave and Catalina, even though Apple stopped updating it years ago. Yes, it is incredibly hard to do this and maintain it with our limited resources, but we believe it should be done, and we are known for doing things the hard way.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-30/1717083278-428587-screen-shot-2024-05-30-at-83125-am.png"></center>
 <center><em>Orion on macOS Catalina using WebKit that is newer than not only WebKit in Safari on Catalina, but also the one used in Safari on Sonoma. Kudos for the Orion dev team.</em></center></p>
 <p>And the best thing - Orion is free (truly free, not like ad-supported browser “free”) and you can choose to <a href="https://kagi.com/orion/orionplus.html">pay for it</a> and support your browser directly.</p>

 <p>We continue to work very hard on Orion, aiming to exit beta status by the beginning of 2025. It is currently <a href="https://kagi.com/orion/#download_sec">available</a> for Mac and iPhone/iPad, with plans (resource permitting) to expand to Windows, Android, and Linux.</p>

 <h2><a name="7"></a>Chapter 2: Setting up for success</h2>

 <p>The relentless pursuit of excellence ensures that Kagi users receive a seamless and trustworthy search experience, setting us apart in an industry driven by ads, tracking, and misaligned incentives.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716359822-625955-screenshot-2024-05-21-at-233656.png"></center>
 <center><em>Note from a Kagi member, shown with their permission</em></center></p>
 <p>As we look behind us, we are proud of what we have built together over the last two years, with our incredible member community.</p>

 <p>We realize that we are now entering the next chapter of Kagi’s journey, Chapter 2, where we are intensifying our focus on craftsmanship and delighting users, supported by a more mature organization.</p>

 <p>We have hired key leadership positions to lead our efforts in these critical areas.</p>

 <p></p>
 <p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716358485-924099-img-7462.jpg"></center>
 <center><em>Kagi leadership meeting, New York, April 2024.<br> From left to right: Zac Nowicki, Vladimir Prelovac, Hugh O’Brien, Luis Da Silva.</em></center></p>
 <p>Additionally, we are reorganizing our internal structure, creating specialized teams, establishing new processes, and setting ourselves up for success over the next two years.</p>

 <h3><a name="8"></a>Moving forward: Relentless focus on core search experience</h3>

 <p>In a rapidly evolving digital landscape where “AI” continues to dominate the conversation, especially in search (it is almost as if the rest of the industry forgot that the purpose of search engines is to focus on relevant search results), Kagi’s unwavering commitment to refining the core search experience has never been more critical.</p>

 <p>We are dedicated to delivering the most accurate, privacy-respecting, and user-centric search results, reinforcing our belief that the human element remains essential in the search experience. We will be investing in expanding our index, launching an API so that everyone can build on top of Kagi’s search quality, and improving the UX of search itself with new widgets, customization and quality of life improvements.</p>

 <p>We recognize that the search experience should be intuitive, fast, reliable, and most importantly, of the highest possible quality. As such, our development efforts are centered around improving search algorithms, expanding our own index, reducing latency, and ensuring that our platform remains free of low-quality and LLM-generated content.</p>

 <h3><a name="9"></a>What about “AI”?</h3>

 <p>Although “AI” has been used in search for decades, it has recently resurfaced in public discourse with the proliferation of large language models (LLMs). We never integrated features (including AI) for the sake of having them, but did so <a href="https://blog.kagi.com/kagi-ai-search#philosophy">as thoughtfully as we could</a> to enhance the user experience without compromising our core values. The job of technology is not to make humans do more; the job of technology is to make our journey in this world more enjoyable.</p>

 <p>I have been using LLMs as a tool for coding (acknowledging their <a href="https://x.com/vladquant/status/1791341643351880050">limitations</a>). I am using them for factual summarization through Kagi’s <a href="https://help.kagi.com/kagi/ai/quick-answer.html">Quick Answer</a> and <a href="https://help.kagi.com/kagi/ai/assistant.html">Assistant</a>, which are on-demand features that work well thanks to grounding in Kagi Search. And as someone diagnosed with osteoarthritis in my hands, I will rely increasingly on voice-to-text AI for typing and interaction. I believe that other AI modalities, such as image understanding, will greatly improve the quality of life for people with disabilities like visual impairment.</p>

 <p>LLM research has produced advanced models, and some companies have paved the way for the public to accept paying $20-$30 per month to use top LLMs. We see an opportunity to compete in this market with Kagi Assistant. We will offer access to world’s best LLMs, with answers grounded in Kagi Search, all in one place and included in a single $25/month Kagi Ultimate subscription. Kagi Assistant is currently in <a href="https://help.kagi.com/kagi/ai/assistant.html">open beta</a>, with the full version expected to launch in July.</p>

 <p>It is clear that the quality of any AI integration built on top of search will only be as good as the underlying search engine. Our primary goal is to keep providing the best search results in the world.</p>

 <h3><a name="10"></a>Roadmap ahead</h3>

 <p>Our roadmap includes several ambitious projects:</p>

 <ul>
 <li>Expansion of Kagi’s own index, aiming to cover 100 million of the highest quality pages on the web. (we will keep focusing on quality vs quantity.)</li>
 <li>Doing more with the wonderful <a href="https://blog.kagi.com/small-web">Kagi Small Web</a> initiative that has grown to over 12,000 websites and is <a href="https://github.com/kagisearch/smallweb">open-source</a>.</li>
 <li>A complete revamp of our search widgets, including stock, sports, weather, and calculator widgets, including <a href="https://kagifeedback.org/d/36-custom-widgetscomputer-plugins">user-defined widgets</a> - all right inside search results.</li>
 <li>Launching Kagi mobile apps for both iOS and Android platforms.</li>
 <li>Launching <a href="https://help.kagi.com/kagi/ai/assistant.html">Kagi Assistant</a>: Access the world’s best LLMs, grounded in Kagi Search.</li>
 <li>Kagi Maps, based on Mapbox and OpenStreetMap.</li>
 <li><a href="https://kagi.com/orion">Orion Browser</a> exiting public beta and offering high-performance (WebKit based), zero-telemetry, ad-free, and tracking-free browsing. Also, similar to Kagi Search, you can <a href="https://kagi.com/orion/orionplus.html">pay for your browser</a>.</li>
 <li>Seamless integration between Kagi products, for great user experience inside our ecosystem.</li>
 <li>Launch of Kagi Search APIs to enable others to build upon our legendary search results.</li>
 <li>Expanding our offering to enterprise and education sectors with a tailored solution. (We’d love to see Kagi in schools, and if you would as well, <a href="mailto:vlad@kagi.com">let us know</a> or mention Kagi to your kid’s teacher.)</li>
 </ul>

 <p><img src="https://kagifeedback.org/assets/files/2024-05-30/1717077415-686041-screenshot-2024-05-30-at-065617.png" alt="Kagi’s Future Roadmap">
 </p>
 <p><center><em>Kagi product roadmap. Dates are aspirational.</em></center></p>
 <p>These initiatives are designed to enhance our core search offering, introduce other products with the same core values to our ecosystem, enlarge the number of people who will benefit from Kagi, and provide even more value to our member community included in their subscription. We are excited about the future and committed to delivering on these promises, one step at a time.</p>

 <p>Thank you for being part of our journey. Here’s to the next chapter of Kagi.</p>

 <p><em>Vladimir Prelovac</em><br>
 <em>CEO, Kagi</em><br>
 <em>vlad@kagi.com</em><br>
 <em>(and please feel free to write to me directly about anything Kagi related)</em></p>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/463a5fd436bb887f90f9e01b01b7716b/index.md
+++ b/cache/2024/463a5fd436bb887f90f9e01b01b7716b/index.md
@@ -0,0 +1,204 @@
 title: What's next for Kagi?
 url: https://blog.kagi.com/what-is-next-for-kagi#9
 hash_url: 463a5fd436bb887f90f9e01b01b7716b
 archive_date: 2024-06-24
 og_image: data:image/svg+xml,<svg xmlns=%22http://www.w3.org/2000/svg%22 viewBox=%220 0 100 100%22><text y=%22.9em%22 font-size=%2290%22>%f0%9f%97%9d%ef%b8%8f</text></svg>
 description: Two years ago, on June 1st, 2022, Kagi introduced ( https://blog.kagi.com/kagi-orion-public-beta ) a search engine that challenged the ad-supported version of the web.
 favicon: data:image/svg+xml,<svg xmlns=%22http://www.w3.org/2000/svg%22 viewBox=%220 0 100 100%22><text y=%22.9em%22 font-size=%2290%22>%f0%9f%97%9d%ef%b8%8f</text></svg>
 language: en_US

 <p>Two years ago, on June 1st, 2022, <a href="https://blog.kagi.com/kagi-orion-public-beta">Kagi introduced</a> a search engine that challenged the ad-supported version of the web. Kagi Search instead works for you, the user, and not an advertiser paying for your attention. At the time of launch, we did not know if anyone would pay for their search engine and web browser, but luckily, here we are two years later at the forefront of a movement to humanize the internet’s most-used products and put the user back in the driving seat.</p>

 <p>This blog post is meant to provide an update to our community and future members on where we are and what our plans for the future are.</p>


 <h2><a name="0"></a>Our commitment to you</h2>

 <ul>
 <li>We provide the best search results in a clear and easy-to-understand way, with no ads, amplifying your knowledge.</li>
 <li>We are committed to your privacy. We allow no tracking or ads, and you remain in control of your data.</li>
 <li>We commit to a user-centric, “pay for the product”, business model with no other sources of revenue.</li>
 <li>We ensure access to timely and high-quality search information. Our refined algorithm ranks information by quality rather than by potential monetization.</li>
 <li>We serve our users’ needs with speed, efficiency, and accuracy.</li>
 <li>We are committed to providing leading-edge technological solutions in service of users.</li>
 <li>We maintain the integrity and quality of our system.</li>
 </ul>

 <h2><a name="1"></a>Growth and sustainability</h2>

 <p>Our journey over the past two years has been marked by continuous growth and improvement. We have expanded our paid member base (which is now <a href="https://kagi.com/stats">more populous than 14 world countries!</a>), introduced a <a href="https://kagi.com/changelog">slew of innovative features</a>, and maintained a steadfast commitment to <a href="https://kagi.com/privacy">privacy</a> and the <a href="https://x.com/vladquant/likes">quality</a> of the search experience.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716361973-858686-screenshot-2024-05-22-at-001246.png"></center>
 <center><em>Kagi member growth over time. Note the expansion after the $10/mo unlimited plan was introduced.</em></center>

 <p>We are also thrilled to report that we have achieved profitability. This significant milestone is a testament to our sustainable growth and fiscal responsibility. It demonstrates that our approach of offering a premium, ad-free search experience resonates with users who support a service aligning with their values. Becoming profitable allows us to reinvest in the business, further enhancing our offerings and ensuring that we can continue to provide a top-notch search experience.</p>

 <p>We are grateful for the support of our users who have played a crucial role in reaching this milestone. Together, we will continue to build a better, more user-centric web.</p>

 <p>As we look to the future, our focus remains on sustaining this momentum. If we continue on the path of current growth, which so far has been completely organic, we will reach 1 million paying users in the next 36 months.</p>

 <h2><a name="2"></a>Increased recognition</h2>

 <p>Kagi is getting noticed by experts like <a href="https://twit.tv">Leo Laporte</a>, <a href="https://www.404media.co/friendship-ended-with-google-now-kagi-is-my-best-friend/">Jason Koebler</a>, <a href="https://daringfireball.net/linked/2024/01/19/bray-google-kagi">John Gruber</a>, and <a href="https://pluralistic.net/2024/04/04/teach-me-how-to-shruggie/">Cory Doctorow</a>, as well as high quality media organizations such as <a href="https://www.theverge.com/23896415/kagi-search-google-meta-quest-3-chatgpt-macos-sonoma-installer-newsletter">The Verge</a> and <a href="https://www.theregister.com/2024/01/30/ai_is_changing_search/">The Register</a>, publications like <a href="https://monocle.com/radio/shows/the-entrepreneurs/eureka-389/play/">Monocle</a> and <a href="https://www.zeit.de/index">Die Zeit</a>. We thank them for their coverage and helping spread the word.</p>

 <p></p><center><a href="https://monocle.com/radio/shows/the-entrepreneurs/eureka-389/play/"><img src="https://kagifeedback.org/assets/files/2024-05-22/1716357691-215959-screenshot-2024-05-21-at-230124.png"></a></center>
 <center><em>Recent interview in Monocle radio</em></center>

 <p>Additionally, we are grateful to content creators like <a href="https://youtu.be/YnSv8ylLfPw?t=523">Linus Tech Tips</a> and <a href="https://www.youtube.com/watch?v=K297opaBp8E&amp;t=7s">Ask Leo</a>, numerous <a href="https://dannb.org/blog/2023/how-kagi-beats-google/">bloggers</a>, and the vibrant discussions by Kagi members on Hacker News, X, Mastodon, Reddit, Blue Sky, LinkedIN and other social media for helping <a href="https://x.com/vladquant/likes">spread awareness</a>.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716359315-445982-screenshot-2024-05-21-at-232829.png"></center>
 <center><em>Public comment from a member on Reddit 🙏</em></center>

 <p>Support like this not only means a lot to us but has been the main driver of Kagi adoption so far, and for that we are forever thankful!</p>

 <h2><a name="3"></a>New shareholders and team members</h2>

 <p>We are excited to announce that so far, 93 Kagi members have joined Kagi as investors through our two investment rounds, most recent closing a few weeks ago. The investment not only brought in capital (which we will use for hiring, R&amp;D and search index expansion) but also strengthens our community-centric approach, as these investors are also users who deeply believe in our mission. The rest of the company remains owned by the founder, employees, and advisors, ensuring that those who are directly involved in building and sustaining Kagi have a significant stake in its success.</p>

 <p>This diverse ownership structure fosters a sense of shared purpose and collaboration and aligns the incentives of everyone involved. Having users as investors provides invaluable insights and feedback (and sometimes key leadership talent!), helping us better understand and meet their needs. It also reinforces our commitment to transparency and accountability, as we are directly answerable to a community that is both invested in and supportive of our long-term vision. While we have no immediate plans to raise additional funding, if you are interested to be notified about any future fundraising, <a href="https://forms.gle/1Try2v6JtXbKSjKx9">get in touch</a>.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716360672-296077-screenshot-2024-05-21-at-235107.png"></center>
 <center><em>Kagi team members around the globe</em></center>

 <p>Our team has expanded to 37 people around the globe, working together to bring a friendly version of the web to people, families, and organizations worldwide. We are actively <a href="https://help.kagi.com/kagi/company/hiring-kagi.html">hiring</a>.</p>

 <h2><a name="4"></a>Kagi becomes a Public Benefit Corporation (PBC)</h2>

 <p>The Delaware public benefit corporation (PBC) is an entity created by amendment to the Delaware General Corporation Law in 2013, and is structurally identical to the familiar Delaware C corporation with one major difference: company directors and officers are given explicit permission to consider a purpose beyond simply maximizing shareholder value.</p>

 <p>Kagi transitioned to a Delaware Public Benefit Corporation (PBC), marking a significant milestone in our journey. This structural change allows the legal entity to consider a purpose beyond simply maximizing shareholder value, balancing it with the best interests of all stakeholders, including the communities we serve, along with the specific public benefit.</p>

 <p>In particular, the public benefit statement entered into Kagi’s charter is: “<strong>Kagi is committed to creating a more human-centric and sustainable web that benefits individuals, communities, and society as a whole, with a transparent business model that aligns the incentives of everyone involved.</strong>”</p>

 <p><img src="https://kagifeedback.org/assets/files/2024-05-22/1716360472-761510-screenshot-2024-05-21-at-234747.png" alt="Kagi PBC"></p>

 <p>This transition ensures that our mission-driven approach is not just a guiding principle but a legal obligation, protecting us from external pressures that might otherwise divert us from our path. The PBC designation also resonates with our community of users, who value ethical business practices and corporate responsibility. It provides a clear signal that Kagi is dedicated to making a positive impact on the world, aligning our business practices with the broader goal of societal benefit.</p>

 <p>Some examples of companies structured as PBC  out there include Laureate, Plum Organics, Kickstarter and Anthropic. We’d like to thank <a href="https://en.wikipedia.org/wiki/Eric_Ries">Eric Ries</a> for introducing us to the PBC structure.  You can read more about it <a href="https://kagi.com/search?q=public+benefit+corporation+PBC&amp;r=us&amp;sh=FuSMLpCLQlOTugxwnnQo6g">on Kagi</a>.</p>

 <h2><a name="5"></a>Kagi’s unique approach to search</h2>

 <p>From day one, Kagi Search was built with one goal in mind — show the world that a high-quality, user-centric search engine is possible. Why was this not possible before? In short, in every market economy, there has to be a customer. For search, that customer can be either you or an advertiser, and this simple choice determines all incentives and outcomes in the search experience. With Kagi, for the first time in history of search engines, you are the customer and everything is built around you and your needs alone.</p>

 <p>And we know what our customers want because it is the same thing we want: no ads, no spam, no low-quality content, and no insult to our intelligence from the search engine we use. We want results that include human discussions, not LLM-generated content. We value opinions and expertise from personal websites and blogs over ad-filled, and tracking-infested content silos.</p>

 <p>We want to explore and discover the web, versus being told what to buy. We want to connect to and understand other humans and their work, not be consumers of the content created with the only purpose of monetizing our attention and wasting our time and brain cycles in the process.</p>

 <p>And this is what Kagi strives to deliver.</p>

 <p>For instance, when you ask your search engine a programming question, you want to see official documentation, industry experts, personal blogs, and other authentic sources of information.</p>

 <p></p><center><a href="https://kagifeedback.org/assets/files/2024-05-22/1716392268-405998-screenshot-2024-05-22-at-083623.png"><img src="https://kagifeedback.org/assets/files/2024-05-22/1716392268-405998-screenshot-2024-05-22-at-083623.png"></a></center>
 <center><em>User-centric vs ad-supported, click to enlarge</em></center>

 <p>Similarly, searching for “best headphones” should yield genuine recommendations, not headphone ads from the highest bidder. What you often find in Kagi is information from people who are truly knowledgeable and passionate about their topic.</p>

 <p></p><center><a href="https://kagifeedback.org/assets/files/2024-05-22/1716392725-691336-screenshot-2024-05-22-at-084426.png"><img src="https://kagifeedback.org/assets/files/2024-05-22/1716392725-691336-screenshot-2024-05-22-at-084426.png"></a></center>
 <center><em>User-centric vs ad-supported, click to enlarge</em></center>

 <p>Whenever possible we want to make sure that facts shown in Kagi are accurate and computable. This is why we <a href="https://blog.kagi.com/kagi-wolfram">integrate with Wolfram|Alpha</a>.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-03-05/1709654440-735786-screenshot-2024-03-05-at-080035.png"></center>

 <p>We are not trying to influence users’ search results; instead, we aim to provide users with tools and features to <a href="https://help.kagi.com/kagi/features/website-info-personalized-results.html">personalize</a> their search experience according to their preferences. For example, with Kagi, users can <a href="https://kagi.com/stats?stat=leaderboard">block sites</a> they do not like and promote those they prefer.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716393012-475138-screenshot-2024-05-22-at-085004.png"></center>

 <p>Use <a href="https://help.kagi.com/kagi/features/lenses.html">Kagi lenses</a> to narrow down searches to the academic realm or one of the many predefined lenses - or create your own search lens with your favorite sites.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716393399-721245-screenshot-2024-05-22-at-085620.png"></center>
 <center><em>Search results with “Academic” lens enabled</em></center>

 <p>The continuous shipping of features and improvements has been largely driven by our active member community and would not have been possible without it. The Kagi Search <a href="https://kagifeedback.org">feedback forum</a> currently hosts over <strong>3,500</strong> discussions, while the Kagi’s Orion Browser <a href="https://orionfeedback.org">feedback forum</a> boasts over <strong>6,400</strong> discussions!</p>

 <p>These massive public forums are among the most active web search and browser product communities on the web, reflecting the incredible level of engagement and valuable insights provided by our users. This is also why <strong>we do not have or need any telemetry in our products</strong> - we can barely keep up with what our users are directly asking us to do!</p>

 <h2><a name="6"></a>Orion - Web browser that has your back</h2>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-30/1717079482-339174-04.png"></center>
 <center><em>Orion is a lightweight Mac browser, with native support for horizontal, vertical and compact tabs</em></center>

 <p>Most people may not know this, but Kagi started developing <a href="https://kagi.com/orion">Orion browser</a> even before the search engine (and Orion’s feedback forum has almost twice as many discussion than the one for Kagi Search!). It is clear to us that no successful search engine could thrive without a companion web browser built on the same set of core values, including the alignment of all incentives.</p>

 <p>We are pleased to see so many people excited about the prospect of Orion and willing to <a href="https://kagi.com/orion/orionplus.html">pay for it</a>, which may sound like an even more “outrageous” idea than paying for search.</p>

 <p>But think about it: a web browser is one of the most important and intimate tools we use on our computers and devices every single day, yet we as a society, have somehow allowed most browsers in use to be directly or indirectly paid for and controlled by advertisers! It is like the situation with search, just worse. This is insanity to me and this is why we are building Orion, a user-centric web browser with a paid business model and no other sources of revenue, to bring sanity back and help support our mission to humanize the web.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-30/1717079301-260116-screenshot-2024-05-30-at-072813.png"></center>
 <center><em>Native support for web extensions, profiles, and much more…</em></center>

 <p>Orion is WebKit-based, built from scratch for high performance, zero-telemetry by default, <a href="https://kagi.com/orion/#privacy">truly privacy respecting</a>, with built-in ad and tracking blocking like no other web browser in the world. On top of that, we brought the same Web Extensions API used by Chrome and Firefox to WebKit, allowing Orion to run both Chrome and Firefox extensions natively out of the box (you can follow <a href="https://docs.google.com/spreadsheets/d/14IgSRVop4psUTgtLZlvYJYrAArhvL3WvRlUdzdQbIoQ/edit?usp=sharing">our progress</a>, with many extensions already working, including on iOS!).</p>

 <p>For the Mac fans out there, you will be pleased to know that Orion has ported the latest WebKit to macOS Mojave and Catalina, even though Apple stopped updating it years ago. Yes, it is incredibly hard to do this and maintain it with our limited resources, but we believe it should be done, and we are known for doing things the hard way.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-30/1717083278-428587-screen-shot-2024-05-30-at-83125-am.png"></center>
 <center><em>Orion on macOS Catalina using WebKit that is newer than not only WebKit in Safari on Catalina, but also the one used in Safari on Sonoma. Kudos for the Orion dev team.</em></center>

 <p>And the best thing - Orion is free (truly free, not like ad-supported browser “free”) and you can choose to <a href="https://kagi.com/orion/orionplus.html">pay for it</a> and support your browser directly.</p>

 <p>We continue to work very hard on Orion, aiming to exit beta status by the beginning of 2025. It is currently <a href="https://kagi.com/orion/#download_sec">available</a> for Mac and iPhone/iPad, with plans (resource permitting) to expand to Windows, Android, and Linux.</p>

 <h2><a name="7"></a>Chapter 2: Setting up for success</h2>

 <p>The relentless pursuit of excellence ensures that Kagi users receive a seamless and trustworthy search experience, setting us apart in an industry driven by ads, tracking, and misaligned incentives.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716359822-625955-screenshot-2024-05-21-at-233656.png"></center>
 <center><em>Note from a Kagi member, shown with their permission</em></center>

 <p>As we look behind us, we are proud of what we have built together over the last two years, with our incredible member community.</p>

 <p>We realize that we are now entering the next chapter of Kagi’s journey, Chapter 2, where we are intensifying our focus on craftsmanship and delighting users, supported by a more mature organization.</p>

 <p>We have hired key leadership positions to lead our efforts in these critical areas.</p>

 <p></p><center><img src="https://kagifeedback.org/assets/files/2024-05-22/1716358485-924099-img-7462.jpg"></center>
 <center><em>Kagi leadership meeting, New York, April 2024.<br> From left to right: Zac Nowicki, Vladimir Prelovac, Hugh O’Brien, Luis Da Silva.</em></center>

 <p>Additionally, we are reorganizing our internal structure, creating specialized teams, establishing new processes, and setting ourselves up for success over the next two years.</p>

 <h3><a name="8"></a>Moving forward: Relentless focus on core search experience</h3>

 <p>In a rapidly evolving digital landscape where “AI” continues to dominate the conversation, especially in search (it is almost as if the rest of the industry forgot that the purpose of search engines is to focus on relevant search results), Kagi’s unwavering commitment to refining the core search experience has never been more critical.</p>

 <p>We are dedicated to delivering the most accurate, privacy-respecting, and user-centric search results, reinforcing our belief that the human element remains essential in the search experience. We will be investing in expanding our index, launching an API so that everyone can build on top of Kagi’s search quality, and improving the UX of search itself with new widgets, customization and quality of life improvements.</p>

 <p>We recognize that the search experience should be intuitive, fast, reliable, and most importantly, of the highest possible quality. As such, our development efforts are centered around improving search algorithms, expanding our own index, reducing latency, and ensuring that our platform remains free of low-quality and LLM-generated content.</p>

 <h3><a name="9"></a>What about “AI”?</h3>

 <p>Although “AI” has been used in search for decades, it has recently resurfaced in public discourse with the proliferation of large language models (LLMs). We never integrated features (including AI) for the sake of having them, but did so <a href="https://blog.kagi.com/kagi-ai-search#philosophy">as thoughtfully as we could</a> to enhance the user experience without compromising our core values. The job of technology is not to make humans do more; the job of technology is to make our journey in this world more enjoyable.</p>

 <p>I have been using LLMs as a tool for coding (acknowledging their <a href="https://x.com/vladquant/status/1791341643351880050">limitations</a>). I am using them for factual summarization through Kagi’s <a href="https://help.kagi.com/kagi/ai/quick-answer.html">Quick Answer</a> and <a href="https://help.kagi.com/kagi/ai/assistant.html">Assistant</a>, which are on-demand features that work well thanks to grounding in Kagi Search. And as someone diagnosed with osteoarthritis in my hands, I will rely increasingly on voice-to-text AI for typing and interaction. I believe that other AI modalities, such as image understanding, will greatly improve the quality of life for people with disabilities like visual impairment.</p>

 <p>LLM research has produced advanced models, and some companies have paved the way for the public to accept paying $20-$30 per month to use top LLMs. We see an opportunity to compete in this market with Kagi Assistant. We will offer access to world’s best LLMs, with answers grounded in Kagi Search, all in one place and included in a single $25/month Kagi Ultimate subscription. Kagi Assistant is currently in <a href="https://help.kagi.com/kagi/ai/assistant.html">open beta</a>, with the full version expected to launch in July.</p>

 <p>It is clear that the quality of any AI integration built on top of search will only be as good as the underlying search engine. Our primary goal is to keep providing the best search results in the world.</p>

 <h3><a name="10"></a>Roadmap ahead</h3>

 <p>Our roadmap includes several ambitious projects:</p>

 <ul>
 <li>Expansion of Kagi’s own index, aiming to cover 100 million of the highest quality pages on the web. (we will keep focusing on quality vs quantity.)</li>
 <li>Doing more with the wonderful <a href="https://blog.kagi.com/small-web">Kagi Small Web</a> initiative that has grown to over 12,000 websites and is <a href="https://github.com/kagisearch/smallweb">open-source</a>.</li>
 <li>A complete revamp of our search widgets, including stock, sports, weather, and calculator widgets, including <a href="https://kagifeedback.org/d/36-custom-widgetscomputer-plugins">user-defined widgets</a> - all right inside search results.</li>
 <li>Launching Kagi mobile apps for both iOS and Android platforms.</li>
 <li>Launching <a href="https://help.kagi.com/kagi/ai/assistant.html">Kagi Assistant</a>: Access the world’s best LLMs, grounded in Kagi Search.</li>
 <li>Kagi Maps, based on Mapbox and OpenStreetMap.</li>
 <li><a href="https://kagi.com/orion">Orion Browser</a> exiting public beta and offering high-performance (WebKit based), zero-telemetry, ad-free, and tracking-free browsing. Also, similar to Kagi Search, you can <a href="https://kagi.com/orion/orionplus.html">pay for your browser</a>.</li>
 <li>Seamless integration between Kagi products, for great user experience inside our ecosystem.</li>
 <li>Launch of Kagi Search APIs to enable others to build upon our legendary search results.</li>
 <li>Expanding our offering to enterprise and education sectors with a tailored solution. (We’d love to see Kagi in schools, and if you would as well, <a href="mailto:vlad@kagi.com">let us know</a> or mention Kagi to your kid’s teacher.)</li>
 </ul>

 <p><img src="https://kagifeedback.org/assets/files/2024-05-30/1717077415-686041-screenshot-2024-05-30-at-065617.png" alt="Kagi’s Future Roadmap">
 </p><center><em>Kagi product roadmap. Dates are aspirational.</em></center>

 <p>These initiatives are designed to enhance our core search offering, introduce other products with the same core values to our ecosystem, enlarge the number of people who will benefit from Kagi, and provide even more value to our member community included in their subscription. We are excited about the future and committed to delivering on these promises, one step at a time.</p>

 <p>Thank you for being part of our journey. Here’s to the next chapter of Kagi.</p>

 <p><em>Vladimir Prelovac</em><br>
 <em>CEO, Kagi</em><br>
 <em>vlad@kagi.com</em><br>
 <em>(and please feel free to write to me directly about anything Kagi related)</em></p>
--- a/cache/2024/486e49501e495e42b2e40320969c373f/index.html
+++ b/cache/2024/486e49501e495e42b2e40320969c373f/index.html
@@ -0,0 +1,184 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="en">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>Did anyone ask for these AI features? • Cory Dransfeldt (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://coryd.dev/posts/2024/did-anyone-ask-for-these-ai-features/">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>Did anyone ask for these AI features? • Cory Dransfeldt</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://coryd.dev/posts/2024/did-anyone-ask-for-these-ai-features/" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <p>I wasn't going to say anything about WWDC. I wasn't going to say anything about WWDC. But, did anyone actually ask for these AI features?</p>
 <p>I mean — other than investors. Did users demonstrate an interest in any of this? Have they demonstrated an interest in any of the myriad applications generative tech has been shoehorned into? <a href="https://www.baldurbjarnason.com/2024/bubble-lifecycles/">It seems pretty clear that we're well into a bubble</a>, but when the money dries up is anyone's guess.</p>
 <p>I'm genuinely curious as to who wants this outside of investors looking for the next big thing. Anecdotally, my more technical friends are generally lukewarm about new AI features and less technical friends either don't pay attention or are annoyed by the changes.</p>
 <p>What Apple announced looks like more of the same that's been offered lately. Help writing emails that don't need to be written, bad generative images that look little better than anything else on the market and a dubious integration with OpenAI which feels <em>super weird</em> given the former's much marketed stance on privacy and the latter's dubious respect for <em>anyone's</em> data<sup class="footnote-ref"><a href="#fn1" id="fnref1">[1]</a></sup>.</p>
 <p>All this yields...a more responsive Siri? One that's more confidently wrong at least 20% of the time? Photo editing so that your images no longer reflect reality? Maybe they'll sort emails more effectively, but that's a pretty sensitive data set for them to be trawling through. I wouldn't expect their AI additions to Xcode to be any more helpful than Copilot (which is to say, not very).</p>
 <p>Apple's been big on touting their environmental record in past presentations but left it out of this one. Maybe it's hard to hype fancy autocorrect when we're spinning carbon-intensive power plants back up to power it.</p>
 <p>Less privacy. More environmental impact. More hype. <em>Cool.</em></p>
 <p><strong class="highlight-text">More than anything it all goes to show that the one thing you can count on companies to do over the long term is disappoint you.</strong></p>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/486e49501e495e42b2e40320969c373f/index.md
+++ b/cache/2024/486e49501e495e42b2e40320969c373f/index.md
@@ -0,0 +1,10 @@
 title: Did anyone ask for these AI features? • Cory Dransfeldt
 url: https://coryd.dev/posts/2024/did-anyone-ask-for-these-ai-features/
 hash_url: 486e49501e495e42b2e40320969c373f
 archive_date: 2024-06-24
 og_image: https://cdn.coryd.dev/assets/avatar.png
 description: I wasn't going to say anything about WWDC. I wasn't going to say anything about WWDC. But, did anyone actually ask for these AI features?
 favicon: https://cdn.coryd.dev/assets/icons/favicon.svg?v=19.5.32
 language: en_US

 <p>I wasn't going to say anything about WWDC. I wasn't going to say anything about WWDC. But, did anyone actually ask for these AI features?</p><p>I mean — other than investors. Did users demonstrate an interest in any of this? Have they demonstrated an interest in any of the myriad applications generative tech has been shoehorned into? <a href="https://www.baldurbjarnason.com/2024/bubble-lifecycles/">It seems pretty clear that we're well into a bubble</a>, but when the money dries up is anyone's guess.</p><p>I'm genuinely curious as to who wants this outside of investors looking for the next big thing. Anecdotally, my more technical friends are generally lukewarm about new AI features and less technical friends either don't pay attention or are annoyed by the changes.</p><p>What Apple announced looks like more of the same that's been offered lately. Help writing emails that don't need to be written, bad generative images that look little better than anything else on the market and a dubious integration with OpenAI which feels <em>super weird</em> given the former's much marketed stance on privacy and the latter's dubious respect for <em>anyone's</em> data<sup class="footnote-ref"><a href="#fn1" id="fnref1">[1]</a></sup>.</p><p>All this yields...a more responsive Siri? One that's more confidently wrong at least 20% of the time? Photo editing so that your images no longer reflect reality? Maybe they'll sort emails more effectively, but that's a pretty sensitive data set for them to be trawling through. I wouldn't expect their AI additions to Xcode to be any more helpful than Copilot (which is to say, not very).</p><p>Apple's been big on touting their environmental record in past presentations but left it out of this one. Maybe it's hard to hype fancy autocorrect when we're spinning carbon-intensive power plants back up to power it.</p><p>Less privacy. More environmental impact. More hype. <em>Cool.</em></p><p><strong class="highlight-text">More than anything it all goes to show that the one thing you can count on companies to do over the long term is disappoint you.</strong></p>
--- a/cache/2024/54ee1030f6fef778ec10e59726bc5731/index.html
+++ b/cache/2024/54ee1030f6fef778ec10e59726bc5731/index.html
@@ -0,0 +1,178 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="en">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>I've had this thought but never put it into words. (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://micro.anniegreens.lol/2024/05/06/ive-had-this.html">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>I've had this thought but never put it into words.</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://micro.anniegreens.lol/2024/05/06/ive-had-this.html" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <p>The modern tech company or capitalistic enterprise is taking cues from fascism and dictators in how they introduce a small piece of controversy, like a testing of the waters. Wait for an “overreaction” by the masses/their audience/core users, wait for that to burnout and normalize, before flooding the zone with too much to fight or argue with all at once. This is the same way fascism causes burnout and complacency and feelings of helplessness, like there’s nothing to be done to fight the thing.</p>
 <p>This is what is being done with AI.</p>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/54ee1030f6fef778ec10e59726bc5731/index.md
+++ b/cache/2024/54ee1030f6fef778ec10e59726bc5731/index.md
@@ -0,0 +1,11 @@
 title: I've had this thought but never put it into words.
 url: https://micro.anniegreens.lol/2024/05/06/ive-had-this.html
 hash_url: 54ee1030f6fef778ec10e59726bc5731
 archive_date: 2024-06-24
 og_image: https://micro.blog/anniegreens/avatar.jpg
 description: The modern tech company or capitalistic enterprise is taking …
 favicon: https://micro.blog/anniegreens/favicon.png
 language: en_US

 <p>The modern tech company or capitalistic enterprise is taking cues from fascism and dictators in how they introduce a small piece of controversy, like a testing of the waters. Wait for an “overreaction” by the masses/their audience/core users, wait for that to burnout and normalize, before flooding the zone with too much to fight or argue with all at once. This is the same way fascism causes burnout and complacency and feelings of helplessness, like there’s nothing to be done to fight the thing.</p>
 <p>This is what is being done with AI.</p>
--- a/cache/2024/63a6d75aee63c072c6d3b174c918740e/index.html
+++ b/cache/2024/63a6d75aee63c072c6d3b174c918740e/index.html
@@ -0,0 +1,205 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="fr">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>Le panneau des préjugés (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://mimo.blog/le-panneau-des-prejuges">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>Le panneau des préjugés</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://mimo.blog/le-panneau-des-prejuges" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <p>Selon Louise Champoux-Paillé et Anne-Marie Croteau de l'École de gestion John-Molson, Université Concordia, les femmes sont moins intéressées par l'IA que les hommes, donc sont plus à risque de perdre leur emploi dans l’avenir. En fait, est-ce vraiment un manque d'intérêt ou ne serait-ce pas plutôt la difficulté de se tailler une place parce que les hommes ont la préférence?</p>

 <p>Un des passages du texte de Louise Champoux-Paillé et Anne-Marie Croteau invite les femmes à prendre leur place dans les métiers où l'IA va s'imposer.</p>

 <blockquote><p>Avec l’appui des dirigeants d’organisations privées et publiques, les femmes auront à déployer des stratégies innovantes et audacieuses afin de s’assurer que l’intégration de l’IA leur permette de poursuivre leur élan, et non le freine, sur le chemin sinueux menant vers la parité.
 Les femmes sont moins intéressées par l'IA que les hommes. Elles auraient pourtant tout avantage à en exploiter le potentiel</p></blockquote>

 <p>Ce passage m'a interpellée parce qu'il laisse entendre que les femmes seraient à blâmer du fait qu'elles résistent à l'utilisation de l'intelligence artificielle.</p>

 <p>Or, <a href="https://www.welcometothejungle.com/fr/articles/femmes-utilisent-moins-ia-mettent-en-danger-leur-carriere">dans un autre article</a> portant également sur le fait que les femmes seraient moins portées à utiliser l'intelligence artificielle, on lit que «les métiers porteurs du marché de l’IA aujourd’hui sont dominés par les hommes.»</p>

 <p>Est-ce que le fait que ces métiers sont dominés par les hommes n'expliquerait pas que plus d'hommes que de femmes utilisent l'IA?</p>

 <p>Plutôt que de vérifier cette donnée, on préfère conclure que les femmes, ne voulant pas apprendre à utiliser l'IA, n'auront qu'elles à blâmer si elles sont incapables d'exercer des métiers, ou professions qui vont demander cette compétence.</p>

 <p>Ce n'est pas la première fois que des changements entraînent des pertes d'emplois. Je ne vois pas pourquoi les femmes seraient moins aptes que les hommes à s'adapter (ou non) aux changements amenés pas l'IA dans l'exercice d'un certain nombre d'emplois.</p>

 <p>Ce qu'il faut, c'est se demander quel soutien auront besoin celles (et ceux) qui oeuvrent dans des emplois transformés par l'IA pour pouvoir s'adapter au changement.</p>

 <p>Les gestionnaires qui vont embaucher dans l'avenir devront pour leur part se demander si c'est le bon profil qui est recherché, à l'exemple de cette entreprise du domaine de l'assurance qui a choisi de modifier ses plans de recrutement.</p>

 <blockquote><p>«<em>Si nous cherchons certes des profils qui aient un socle de savoirs techniques, ce qui nous intéresse surtout ce sont des personnalités dotées de soft skills, capables de s’adapter et d’apprendre car les métiers techniques évoluent en permanence</em>».
 <a href="https://www.lefigaro.fr/decideurs/management/les-femmes-trouveront-elles-leur-place-dans-les-metiers-d-avenir-20240523">Les femmes trouveront-elles leur place dans les métiers d’avenir?</a></p></blockquote>

 <p>L'entreprise en question a non seulement recruté davantage de femmes, mais en pls elle dispose d'une main d'œuvrent qui va pouvoir continuer d'évoluer.</p>

 <p>Je ne dis pas que les femmes sont forcément meilleures que les hommes face à ce défi d'exercer un métier qui nécessite de constamment s'adapter, mais qui sait s'il n'y aurait pas là une caractéristique qui serait à notre avantage?</p>

 <p>Pourvu que les gestionnaires ne tombent pas dans le panneau des préjugés envers les femmes.</p>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/63a6d75aee63c072c6d3b174c918740e/index.md
+++ b/cache/2024/63a6d75aee63c072c6d3b174c918740e/index.md
@@ -0,0 +1,39 @@
 title: Le panneau des préjugés
 url: https://mimo.blog/le-panneau-des-prejuges
 hash_url: 63a6d75aee63c072c6d3b174c918740e
 archive_date: 2024-06-24
 og_image: https://i.snap.as/NRRhOg4f.jpg
 description: Selon Louise Champoux-Paillé et Anne-Marie Croteau de l'École de gestion John-Molson, Université Concordia, les femmes sont moins intéres...
 favicon: https://i.snap.as/CoWgXgFY.png
 language: fr_FR

 <p>Selon Louise Champoux-Paillé et Anne-Marie Croteau de l'École de gestion John-Molson, Université Concordia, les femmes sont moins intéressées par l'IA que les hommes, donc sont plus à risque de perdre leur emploi dans l’avenir. En fait, est-ce vraiment un manque d'intérêt ou ne serait-ce pas plutôt la difficulté de se tailler une place parce que les hommes ont la préférence?</p>


 <p>Un des passages du texte de Louise Champoux-Paillé et Anne-Marie Croteau invite les femmes à prendre leur place dans les métiers où l'IA va s'imposer.</p>

 <blockquote><p>Avec l’appui des dirigeants d’organisations privées et publiques, les femmes auront à déployer des stratégies innovantes et audacieuses afin de s’assurer que l’intégration de l’IA leur permette de poursuivre leur élan, et non le freine, sur le chemin sinueux menant vers la parité.
 Les femmes sont moins intéressées par l'IA que les hommes. Elles auraient pourtant tout avantage à en exploiter le potentiel</p></blockquote>

 <p>Ce passage m'a interpellée parce qu'il laisse entendre que les femmes seraient à blâmer du fait qu'elles résistent à l'utilisation de l'intelligence artificielle.</p>

 <p>Or, <a href="https://www.welcometothejungle.com/fr/articles/femmes-utilisent-moins-ia-mettent-en-danger-leur-carriere">dans un autre article</a> portant également sur le fait que les femmes seraient moins portées à utiliser l'intelligence artificielle, on lit que «les métiers porteurs du marché de l’IA aujourd’hui sont dominés par les hommes.»</p>

 <p>Est-ce que le fait que ces métiers sont dominés par les hommes n'expliquerait pas que plus d'hommes que de femmes utilisent l'IA?</p>

 <p>Plutôt que de vérifier cette donnée, on préfère conclure que les femmes, ne voulant pas apprendre à utiliser l'IA, n'auront qu'elles à blâmer si elles sont incapables d'exercer des métiers, ou professions qui vont demander cette compétence.</p>

 <p>Ce n'est pas la première fois que des changements entraînent des pertes d'emplois. Je ne vois pas pourquoi les femmes seraient moins aptes que les hommes à s'adapter (ou non) aux changements amenés pas l'IA dans l'exercice d'un certain nombre d'emplois.</p>

 <p>Ce qu'il faut, c'est se demander quel soutien auront besoin celles (et ceux) qui oeuvrent dans des emplois transformés par l'IA pour pouvoir s'adapter au changement.</p>

 <p>Les gestionnaires qui vont embaucher dans l'avenir devront pour leur part se demander si c'est le bon profil qui est recherché, à l'exemple de cette entreprise du domaine de l'assurance qui a choisi de modifier ses plans de recrutement.</p>

 <blockquote><p>«<em>Si nous cherchons certes des profils qui aient un socle de savoirs techniques, ce qui nous intéresse surtout ce sont des personnalités dotées de soft skills, capables de s’adapter et d’apprendre car les métiers techniques évoluent en permanence</em>».
 <a href="https://www.lefigaro.fr/decideurs/management/les-femmes-trouveront-elles-leur-place-dans-les-metiers-d-avenir-20240523">Les femmes trouveront-elles leur place dans les métiers d’avenir?</a></p></blockquote>

 <p>L'entreprise en question a non seulement recruté davantage de femmes, mais en pls elle dispose d'une main d'œuvrent qui va pouvoir continuer d'évoluer.</p>

 <p>Je ne dis pas que les femmes sont forcément meilleures que les hommes face à ce défi d'exercer un métier qui nécessite de constamment s'adapter, mais qui sait s'il n'y aurait pas là une caractéristique qui serait à notre avantage?</p>

 <p>Pourvu que les gestionnaires ne tombent pas dans le panneau des préjugés envers les femmes.</p>
--- a/cache/2024/8b0a0eb43814f315e2b296aa5bdd5eee/index.html
+++ b/cache/2024/8b0a0eb43814f315e2b296aa5bdd5eee/index.html
@@ -0,0 +1,223 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="fr">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>ChatGPT et l’indifférence à la vérité (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://www.arthurperret.fr/blog/2024-06-21-chatgpt-et-l-indifference-a-la-verite.html">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>ChatGPT et l’indifférence à la vérité</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://www.arthurperret.fr/blog/2024-06-21-chatgpt-et-l-indifference-a-la-verite.html" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <p>Vous connaissez peut-être le mot anglais <em>bullshit</em>, qu’on peut traduire en français par connerie ou foutaise. Le philosophe américain Harry Frankfurt en a fait un concept tout à fait sérieux<span class="citation" data-cites="frankfurt2005"><span><label for="sn-0" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-0" class="margin-toggle"><span class="sidenote"><a href="#ref-frankfurt2005" role="doc-biblioref"><em>On bullshit</em>, 2005</a>.<br>
 </span></span></span>. Pour lui, le <em>bullshit</em> n’est pas le mensonge mais l’indifférence à la vérité : les <em>bullshitters</em>, qu’on trouve notamment dans la publicité ou la politique, ne se préoccupent pas de savoir si ce qu’ils disent est vrai ou faux ; ce qui compte, ce sont les effets que leurs énoncés produisent.</p>
 <p>Dans un article récent, que j’ai découvert <a href="https://etnadji.fr/veille/c3139f0d-40c8-4e50-b149-f587901f5e93.html">via Étienne Nadji</a>, trois auteurs appliquent ce concept de <em>bullshit</em> à ChatGPT<span class="citation" data-cites="hicks2024"><span><label for="sn-1" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-1" class="margin-toggle"><span class="sidenote"><a href="#ref-hicks2024" role="doc-biblioref">Hicks, Humphries et Slater, <span>« ChatGPT is bullshit »</span>, 2024</a>.<br>
 </span></span></span>. Pour eux, c’est une <em>bullshit machine</em> : un générateur de conneries. Autre traduction, plus littérale et plus vulgaire mais tout aussi parlante : ChatGPT dit de la merde.</p>
 <p>Pourquoi ce débat terminologique est-il important ? Parce qu’il nous donne des ressources pour critiquer le terme « hallucination », utilisé pour décrire le fait que ChatGPT génère parfois de fausses informations – un terme impropre et dont il faudrait se débarrasser si on veut avoir un vrai débat sur ce que des outils comme ChatGPT risquent de faire à nos sociétés.</p>
 <hr>
 <p>Dire que ChatGPT « hallucine » est un anthropomorphisme : c’est une métaphore, une représentation calquée sur une réalité que nous vivons en tant qu’humains ; en l’occurrence, on hallucine quand on perçoit quelque chose qui n’est pas vraiment là.</p>
 <p>Or la réalité du fonctionnement de ChatGPT, qui est basé sur un grand modèle de langue (<em>large language model</em>, LLM) est très différente :</p>
 <blockquote>
 <p><span class="marginnote"><em>“LLMs do not perceive, so they surely do not “mis-perceive”. […] What occurs in the case of an LLM delivering false utterances is not an unusual or deviant form of the process it usually goes through […] the very same process occurs when its outputs happen to be true. […] It falsely indicates that ChatGPT is, in general, attempting to convey accurate information in its utterances. Where it does track truth, it does so indirectly, and incidentally. […] It is bullshitting, even when it’s right.”</em></span> « Les LLMs ne perçoivent pas, ils ne peuvent donc pas “mal percevoir”. […] Lorsqu’un LLM énonce quelque chose de faux, ce n’est pas dû à un accident de parcours […] il suit exactement le même processus lorsqu’il énonce quelque chose de vrai. […] Dire que ChatGPT hallucine, c’est sous-entendre qu’en temps normal il essaye de transmettre des informations justes. Or c’est faux : le fait que ChatGPT tombe juste n’est qu’un résultat indirect de son fonctionnement. […] ChatGPT se fout de ce qu’il raconte, y compris quand il a raison<span class="citation" data-cites="hicks2024"><span><label for="sn-2" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-2" class="margin-toggle"><span class="sidenote"><a href="#ref-hicks2024" role="doc-biblioref"><em>Ibid.</em>, p. 38‑39</a>.<br>
 </span></span></span>. »</p>
 </blockquote>
 <p>C’est essentiel à comprendre : en tant que programme, la fonction de ChatGPT n’est pas de nous informer mais de simuler un interlocuteur humain. Pour cela, il assemble des éléments de langage de manière probabiliste. Les messages qu’il produit de cette manière n’ont de fond que pour la forme : ils doivent être cohérents d’un point de vue linguistique, et c’est ce qui définit leur contenu sémantique ; que ce contenu soit vrai ou faux, peu importe.</p>
 <p>Alors pourquoi parle-t-on d’« hallucinations » ? Pour les concepteurs d’outils, comme OpenAI pour ChatGPT, c’est assez intéressant : au lieu de reconnaître leur responsabilité dans la création d’un générateur de conneries, ils rejettent la faute sur les modèles de langue qui ne seraient pas assez aboutis. La solution est alors toute trouvée : il faut continuer à améliorer les modèles. Jouer sur le langage permet à ces acteurs économiques de détourner notre attention et d’esquiver des épreuves de justification : c’est la mécanique de l’impensé<span class="citation" data-cites="robert2020"><span><label for="sn-3" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-3" class="margin-toggle"><span class="sidenote"><a href="#ref-robert2020" role="doc-biblioref">Robert (dir.), <em>L’impensé numérique - Tome 2</em>, 2020</a>.<br>
 </span></span></span>.</p>
 <p>Pour compléter, on peut reprendre le vocabulaire de <span class="citation" data-cites="mcluhan1968">McLuhan<span><label for="sn-4" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-4" class="margin-toggle"><span class="sidenote"><a href="#ref-mcluhan1968" role="doc-biblioref"><em>Pour comprendre les médias</em>, 2015 [1968]</a>.<br>
 </span></span></span> et analyser ChatGPT en tant que média : on cherche ici à attirer notre attention sur le contenu de ChatGPT plutôt que sur son caractère. On se focalise sur la nature vraie ou fausse des propos générés ; ce qui nous pousse dans une logique familière, celle de la lutte contre la désinformation qui passe notamment par la vérification des faits. Ce faisant, on n’examine pas la nature de ChatGPT en tant que média, son caractère ; or c’est ce dernier qui influence le plus nos activités, et qui peut affecter durablement l’évolution de nos sociétés.</p>
 <p>Ce qui caractérise ChatGPT, nous l’avons vu, c’est qu’il se fout de ce qu’il raconte : il communique pour communiquer. Je crois qu’il y a lieu de s’inquiéter des effets d’un tel média sur nos sociétés. Comme le dénonçait notamment Norbert Wiener, lorsqu’il n’y a de communication que pour elle-même, elle n’a plus de valeur<span class="citation" data-cites="wiener2014a"><span><label for="sn-5" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-5" class="margin-toggle"><span class="sidenote"><a href="#ref-wiener2014a" role="doc-biblioref"><span>« Rôle de l’intellectuel et du savant »</span>, 2014 [1954]</a>.<br>
 </span></span></span>.</p>
 <p>Quelle culture de l’information ChatGPT pourrait-il produire ? Récemment, Julien Gossa signalait <a href="https://social.sciences.re/@juliengossa/112597638789699369">sur Mastodon</a> <a href="https://open.devinci.fr/ressource/etude-2024-impact-ia-generatives-etudiants/">un sondage</a> effectué au sein des écoles privées du pôle Léonard de Vinci, et qui suggère que l’utilisation de ChatGPT par des étudiants se traduit le plus souvent par une dépendance à l’outil sans réelle maîtrise de celui-ci : ChatGPT augmente la productivité mais réduit la littératie. L’IA générative est-elle en train de fabriquer de nouvelles formes d’usines, comme le suggère Olivier Ertzscheid dans <a href="https://cfeditions.com/ia-cyberespace/">son dernier ouvrage</a>, avec un « cognitariat<span class="citation" data-cites="newfield2009"><span><label for="sn-6" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-6" class="margin-toggle"><span class="sidenote"><a href="#ref-newfield2009" role="doc-biblioref">Newfield, <span>« Structure et silence du cognitariat »</span>, 2009</a>.<br>
 </span></span></span> » docile ?</p>
 <p>Les crises auxquelles l’humanité fait face nécessitent un sursaut d’intelligence collective ; la rhétorique de l’« hallucination » masque un projet de société qui va dans le sens inverse. Même si <a href="https://www.arthurperret.fr/veille/2024-04-04-implosion-de-la-bulle-ia.html">la bulle IA implose</a>, ses effets peuvent pourrir plusieurs générations d’apprenants, donc il faut réagir. L’un des axes stratégiques, c’est l’éducation aux médias et à l’information. Comme je l’ai esquissé dans ce billet, les SIC peuvent produire des analyses critiques de ChatGPT en tant que média, pour alimenter cette réponse. Au boulot !</p>
 <section id="bibliographie" class="level1">
 <h1>Bibliographie</h1>
 <aside>
 Pour une synthèse de la littérature scientifique critique sur l’intelligence artificielle générative, on peut consulter le travail de Baldur Bjarnason, que j’avais signalé <a href="2023-08-22-ia-llm-chatgpt-quelques-analyses.html">dans une note de veille</a>, notamment son ouvrage <a href="https://illusion.baldurbjarnason.com"><em>The Intelligence Illusion</em></a>.
 </aside>
 <div id="refs" class="references csl-bib-body" role="list">
 <p id="ref-frankfurt2005" class="csl-entry" role="listitem">
 <span class="smallcaps">Frankfurt</span>, Harry G. <em>On bullshit</em>. Princeton University Press, 2005. 978-0-691-12294-6.
 </p>
 <div id="ref-hicks2024" class="csl-entry" role="listitem">
 <span class="smallcaps">Hicks</span><p>, Michael Townsen, </p><span class="smallcaps">Humphries</span><p>, James et </p><span class="smallcaps">Slater</span><p>, Joe. </p><span>« ChatGPT is bullshit »</span><p>. </p><em>Ethics and Information Technology</em><p>. 2024, Vol. 26, n° 2, p. 38. </p><a href="https://doi.org/10.1007/s10676-024-09775-5">https://doi.org/10.1007/s10676-024-09775-5</a><p>.
 </p></div>
 <p id="ref-mcluhan1968" class="csl-entry" role="listitem">
 <span class="smallcaps">McLuhan</span>, Marshall. <em>Pour comprendre les médias: les prolongements technologiques del’homme</em>. Trad. par Jean Paré. Seuil, 2015 [1968]. 978-2-7578-5014-5.
 </p>
 <div id="ref-newfield2009" class="csl-entry" role="listitem">
 <span class="smallcaps">Newfield</span><p>, Christopher. </p><span>« Structure et silence du cognitariat »</span><p>. </p><em>Multitudes</em><p>. Trad. par Yves Citton. 2009, Vol. 39, n° 4, p. 68‑78. </p><a href="https://doi.org/10.3917/mult.039.0068">https://doi.org/10.3917/mult.039.0068</a><p>.
 </p></div>
 <div id="ref-robert2020" class="csl-entry" role="listitem">
 <span class="smallcaps">Robert</span><p> (dir.). </p><em>L’impensé numérique - Tome 2. Interprétations critiques et logiques pragmatiques de l’impensé</em><p>. Editions des archives contemporaines, 2020. 978-2-8130-0357-7. </p><a href="https://doi.org/10.17184/eac.9782813003577">https://doi.org/10.17184/eac.9782813003577</a><p>.
 </p></div>
 <p id="ref-wiener2014a" class="csl-entry" role="listitem">
 <span class="smallcaps">Wiener</span>, Norbert. <span>« Rôle de l’intellectuel et du savant »</span>. Dans : <em>Cybernétique et société: l’usage humain des êtres humains</em>. Trad. par Ronan Le Roux et Pierre Yves Mistoulon. Seuil, 2014 [1954], p. 158‑162. 978-2-7578-4278-2.
 </p>
 </div>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/8b0a0eb43814f315e2b296aa5bdd5eee/index.md
+++ b/cache/2024/8b0a0eb43814f315e2b296aa5bdd5eee/index.md
@@ -0,0 +1,56 @@
 title: ChatGPT et l’indifférence à la vérité
 url: https://www.arthurperret.fr/blog/2024-06-21-chatgpt-et-l-indifference-a-la-verite.html
 hash_url: 8b0a0eb43814f315e2b296aa5bdd5eee
 archive_date: 2024-06-24
 og_image: https://www.arthurperret.fr/img/scriptor.jpg
 description: Site web dâArthur Perret, enseignant-chercheur en SIC.
 favicon: https://www.arthurperret.fr/blog/favicon-32x32.png
 language: fr_FR

 <p>Vous connaissez peut-être le mot anglais <em>bullshit</em>, qu’on peut traduire en français par connerie ou foutaise. Le philosophe américain Harry Frankfurt en a fait un concept tout à fait sérieux<span class="citation" data-cites="frankfurt2005"><span><label for="sn-0" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-0" class="margin-toggle"><span class="sidenote"><a href="#ref-frankfurt2005" role="doc-biblioref"><em>On bullshit</em>, 2005</a>.<br>
 </span></span></span>. Pour lui, le <em>bullshit</em> n’est pas le mensonge mais l’indifférence à la vérité : les <em>bullshitters</em>, qu’on trouve notamment dans la publicité ou la politique, ne se préoccupent pas de savoir si ce qu’ils disent est vrai ou faux ; ce qui compte, ce sont les effets que leurs énoncés produisent.</p>
 <p>Dans un article récent, que j’ai découvert <a href="https://etnadji.fr/veille/c3139f0d-40c8-4e50-b149-f587901f5e93.html">via Étienne Nadji</a>, trois auteurs appliquent ce concept de <em>bullshit</em> à ChatGPT<span class="citation" data-cites="hicks2024"><span><label for="sn-1" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-1" class="margin-toggle"><span class="sidenote"><a href="#ref-hicks2024" role="doc-biblioref">Hicks, Humphries et Slater, <span>« ChatGPT is bullshit »</span>, 2024</a>.<br>
 </span></span></span>. Pour eux, c’est une <em>bullshit machine</em> : un générateur de conneries. Autre traduction, plus littérale et plus vulgaire mais tout aussi parlante : ChatGPT dit de la merde.</p>
 <p>Pourquoi ce débat terminologique est-il important ? Parce qu’il nous donne des ressources pour critiquer le terme « hallucination », utilisé pour décrire le fait que ChatGPT génère parfois de fausses informations – un terme impropre et dont il faudrait se débarrasser si on veut avoir un vrai débat sur ce que des outils comme ChatGPT risquent de faire à nos sociétés.</p>
 <hr>
 <p>Dire que ChatGPT « hallucine » est un anthropomorphisme : c’est une métaphore, une représentation calquée sur une réalité que nous vivons en tant qu’humains ; en l’occurrence, on hallucine quand on perçoit quelque chose qui n’est pas vraiment là.</p>
 <p>Or la réalité du fonctionnement de ChatGPT, qui est basé sur un grand modèle de langue (<em>large language model</em>, LLM) est très différente :</p>
 <blockquote>
 <p><span class="marginnote"><em>“LLMs do not perceive, so they surely do not “mis-perceive”. […] What occurs in the case of an LLM delivering false utterances is not an unusual or deviant form of the process it usually goes through […] the very same process occurs when its outputs happen to be true. […] It falsely indicates that ChatGPT is, in general, attempting to convey accurate information in its utterances. Where it does track truth, it does so indirectly, and incidentally. […] It is bullshitting, even when it’s right.”</em></span> « Les LLMs ne perçoivent pas, ils ne peuvent donc pas “mal percevoir”. […] Lorsqu’un LLM énonce quelque chose de faux, ce n’est pas dû à un accident de parcours […] il suit exactement le même processus lorsqu’il énonce quelque chose de vrai. […] Dire que ChatGPT hallucine, c’est sous-entendre qu’en temps normal il essaye de transmettre des informations justes. Or c’est faux : le fait que ChatGPT tombe juste n’est qu’un résultat indirect de son fonctionnement. […] ChatGPT se fout de ce qu’il raconte, y compris quand il a raison<span class="citation" data-cites="hicks2024"><span><label for="sn-2" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-2" class="margin-toggle"><span class="sidenote"><a href="#ref-hicks2024" role="doc-biblioref"><em>Ibid.</em>, p. 38‑39</a>.<br>
 </span></span></span>. »</p>
 </blockquote>
 <p>C’est essentiel à comprendre : en tant que programme, la fonction de ChatGPT n’est pas de nous informer mais de simuler un interlocuteur humain. Pour cela, il assemble des éléments de langage de manière probabiliste. Les messages qu’il produit de cette manière n’ont de fond que pour la forme : ils doivent être cohérents d’un point de vue linguistique, et c’est ce qui définit leur contenu sémantique ; que ce contenu soit vrai ou faux, peu importe.</p>
 <p>Alors pourquoi parle-t-on d’« hallucinations » ? Pour les concepteurs d’outils, comme OpenAI pour ChatGPT, c’est assez intéressant : au lieu de reconnaître leur responsabilité dans la création d’un générateur de conneries, ils rejettent la faute sur les modèles de langue qui ne seraient pas assez aboutis. La solution est alors toute trouvée : il faut continuer à améliorer les modèles. Jouer sur le langage permet à ces acteurs économiques de détourner notre attention et d’esquiver des épreuves de justification : c’est la mécanique de l’impensé<span class="citation" data-cites="robert2020"><span><label for="sn-3" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-3" class="margin-toggle"><span class="sidenote"><a href="#ref-robert2020" role="doc-biblioref">Robert (dir.), <em>L’impensé numérique - Tome 2</em>, 2020</a>.<br>
 </span></span></span>.</p>
 <p>Pour compléter, on peut reprendre le vocabulaire de <span class="citation" data-cites="mcluhan1968">McLuhan<span><label for="sn-4" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-4" class="margin-toggle"><span class="sidenote"><a href="#ref-mcluhan1968" role="doc-biblioref"><em>Pour comprendre les médias</em>, 2015 [1968]</a>.<br>
 </span></span></span> et analyser ChatGPT en tant que média : on cherche ici à attirer notre attention sur le contenu de ChatGPT plutôt que sur son caractère. On se focalise sur la nature vraie ou fausse des propos générés ; ce qui nous pousse dans une logique familière, celle de la lutte contre la désinformation qui passe notamment par la vérification des faits. Ce faisant, on n’examine pas la nature de ChatGPT en tant que média, son caractère ; or c’est ce dernier qui influence le plus nos activités, et qui peut affecter durablement l’évolution de nos sociétés.</p>
 <p>Ce qui caractérise ChatGPT, nous l’avons vu, c’est qu’il se fout de ce qu’il raconte : il communique pour communiquer. Je crois qu’il y a lieu de s’inquiéter des effets d’un tel média sur nos sociétés. Comme le dénonçait notamment Norbert Wiener, lorsqu’il n’y a de communication que pour elle-même, elle n’a plus de valeur<span class="citation" data-cites="wiener2014a"><span><label for="sn-5" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-5" class="margin-toggle"><span class="sidenote"><a href="#ref-wiener2014a" role="doc-biblioref"><span>« Rôle de l’intellectuel et du savant »</span>, 2014 [1954]</a>.<br>
 </span></span></span>.</p>
 <p>Quelle culture de l’information ChatGPT pourrait-il produire ? Récemment, Julien Gossa signalait <a href="https://social.sciences.re/@juliengossa/112597638789699369">sur Mastodon</a> <a href="https://open.devinci.fr/ressource/etude-2024-impact-ia-generatives-etudiants/">un sondage</a> effectué au sein des écoles privées du pôle Léonard de Vinci, et qui suggère que l’utilisation de ChatGPT par des étudiants se traduit le plus souvent par une dépendance à l’outil sans réelle maîtrise de celui-ci : ChatGPT augmente la productivité mais réduit la littératie. L’IA générative est-elle en train de fabriquer de nouvelles formes d’usines, comme le suggère Olivier Ertzscheid dans <a href="https://cfeditions.com/ia-cyberespace/">son dernier ouvrage</a>, avec un « cognitariat<span class="citation" data-cites="newfield2009"><span><label for="sn-6" class="margin-toggle sidenote-number"></label><input type="checkbox" id="sn-6" class="margin-toggle"><span class="sidenote"><a href="#ref-newfield2009" role="doc-biblioref">Newfield, <span>« Structure et silence du cognitariat »</span>, 2009</a>.<br>
 </span></span></span> » docile ?</p>
 <p>Les crises auxquelles l’humanité fait face nécessitent un sursaut d’intelligence collective ; la rhétorique de l’« hallucination » masque un projet de société qui va dans le sens inverse. Même si <a href="https://www.arthurperret.fr/veille/2024-04-04-implosion-de-la-bulle-ia.html">la bulle IA implose</a>, ses effets peuvent pourrir plusieurs générations d’apprenants, donc il faut réagir. L’un des axes stratégiques, c’est l’éducation aux médias et à l’information. Comme je l’ai esquissé dans ce billet, les SIC peuvent produire des analyses critiques de ChatGPT en tant que média, pour alimenter cette réponse. Au boulot !</p>
 <section id="bibliographie" class="level1">
 <h1>Bibliographie</h1>
 <aside>
 Pour une synthèse de la littérature scientifique critique sur l’intelligence artificielle générative, on peut consulter le travail de Baldur Bjarnason, que j’avais signalé <a href="2023-08-22-ia-llm-chatgpt-quelques-analyses.html">dans une note de veille</a>, notamment son ouvrage <a href="https://illusion.baldurbjarnason.com"><em>The Intelligence Illusion</em></a>.
 </aside>
 <div id="refs" class="references csl-bib-body" role="list">
 <p id="ref-frankfurt2005" class="csl-entry" role="listitem">
 <span class="smallcaps">Frankfurt</span>, Harry G. <em>On bullshit</em>. Princeton University Press, 2005. 978-0-691-12294-6.
 </p>
 <div id="ref-hicks2024" class="csl-entry" role="listitem">
 <span class="smallcaps">Hicks</span><p>, Michael Townsen, </p><span class="smallcaps">Humphries</span><p>, James et </p><span class="smallcaps">Slater</span><p>, Joe. </p><span>« ChatGPT is bullshit »</span><p>. </p><em>Ethics and Information Technology</em><p>. 2024, Vol. 26, n° 2, p. 38. </p><a href="https://doi.org/10.1007/s10676-024-09775-5">https://doi.org/10.1007/s10676-024-09775-5</a><p>.
 </p></div>
 <p id="ref-mcluhan1968" class="csl-entry" role="listitem">
 <span class="smallcaps">McLuhan</span>, Marshall. <em>Pour comprendre les médias: les prolongements technologiques del’homme</em>. Trad. par Jean Paré. Seuil, 2015 [1968]. 978-2-7578-5014-5.
 </p>
 <div id="ref-newfield2009" class="csl-entry" role="listitem">
 <span class="smallcaps">Newfield</span><p>, Christopher. </p><span>« Structure et silence du cognitariat »</span><p>. </p><em>Multitudes</em><p>. Trad. par Yves Citton. 2009, Vol. 39, n° 4, p. 68‑78. </p><a href="https://doi.org/10.3917/mult.039.0068">https://doi.org/10.3917/mult.039.0068</a><p>.
 </p></div>
 <div id="ref-robert2020" class="csl-entry" role="listitem">
 <span class="smallcaps">Robert</span><p> (dir.). </p><em>L’impensé numérique - Tome 2. Interprétations critiques et logiques pragmatiques de l’impensé</em><p>. Editions des archives contemporaines, 2020. 978-2-8130-0357-7. </p><a href="https://doi.org/10.17184/eac.9782813003577">https://doi.org/10.17184/eac.9782813003577</a><p>.
 </p></div>
 <p id="ref-wiener2014a" class="csl-entry" role="listitem">
 <span class="smallcaps">Wiener</span>, Norbert. <span>« Rôle de l’intellectuel et du savant »</span>. Dans : <em>Cybernétique et société: l’usage humain des êtres humains</em>. Trad. par Ronan Le Roux et Pierre Yves Mistoulon. Seuil, 2014 [1954], p. 158‑162. 978-2-7578-4278-2.
 </p>
 </div>
--- a/cache/2024/8f9d64c455a9246d5e23810ed10e3fe2/index.html
+++ b/cache/2024/8f9d64c455a9246d5e23810ed10e3fe2/index.html
@@ -0,0 +1,268 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="en">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>Turning the Tables on AI (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://ia.net/topics/turning-the-tables-on-ai">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>Turning the Tables on AI</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://ia.net/topics/turning-the-tables-on-ai" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <p>Artificial Intelligence has assimilated Microsoft, Facebook, Google, and now Apple. There is no escape: As soon as you open your computer you’re invited to skip thinking and let the machine fill the void. But what if we used AI not to think less but <em>more</em>?</p>
 <p>AI has one trait: It makes everything the same. Every tech company now offers more or less the same service: “Think less.” How about Apple? Think different? Not anymore.</p>
 <p>Tech companies big and small sell AI as something that thinks for us. It <em>does</em> replace thought with statistics—but it is not intelligent. No one knows what the future will bring. But is a future without thought a better future?</p>
 <p>Now, with a tool that might help us think… How about using AI not to think less but <em>more</em>?</p>
 <p><em>But I don’t know where to start. But I really like that ChatGPT sentence there. But I can’t say it better. But I’m stuck. But I’m worried I wrote something stupid. But I just need to correct some commas and typos. But I need a second opinion.</em></p>
 <p>Let’s turn the tables on our excuses.</p>
 <h2>1. Don’t ask AI, let AI ask you</h2>
 <h3>a) But I don’t know where to start</h3>
 <p>Give ChatGPT an outline and let it write for you. Let’s turn the tables and have ChatGPT prompt <em>us</em>. Tell AI to ask you questions about what you’re writing. Push yourself to express in clear terms what you really want to say. Like this, for example:</p>
 <blockquote>
 <p>I want to write [format] about [topic]. Ask me questions one at a time that force me to explain my idea.</p>
 </blockquote>
 <p>Keep asking until your idea is clear to you. Here’s how such a dialog where ChatGPT prompts <em>you</em> would look:</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/write-for-me.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Write for me:</b> This is how AI is sold to us. But why write if you have nothing to say? Who would read it, and why?</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/ask-ChatGPT.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Ask me instead:</b> Make AI ask *you* questions about your text. Copy-paste your answers. First draft done.</figcaption>
 <p></p>
 <p>Once you have enough to work with, paste <em>your</em> answers into your text editor. You have just written your first draft without cheating.</p>
 <p>ChatGPT may have asked or said some clever things along the way. Feeling tempted to steal it? Don’t. Rather use it as a chance to think and express what you feel.</p>
 <h2>2. Don’t sell stolen goods—make your own</h2>
 <h3>b) But I really like that ChatGPT sentence there</h3>
 <p><em>What if the ChatGPT generated something useful that I want to keep?</em> Paste it as a note <em>Marked as AI.</em> Use quotes, use markup, and note its origin.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/paste-as-AI-from-ChatGPT.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Paste as AI:</b> Want to use AI-generated output? In Writer, right-click and select Paste as AI.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/text-marked-as-ai.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>External sources in Writer:</b> iA Writer greys out text that you marked as AI so you can always discern what is yours and what isn’t.</figcaption>
 <p></p>
 <h2>3. Don’t pretend. Create</h2>
 <h3>c) But I can’t say it better</h3>
 <p><em>What if Artificial Intelligence says exactly what I wanted to say?</em> Rethink and rewrite it to make it your own. Ask if what has been generated is really true. At first glance, it may look tight, but if you really think it through, you’ll find mistakes. Feel it, mean it, think it through. Say it your way.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/cheating-with-ia-writer.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>It’s not yours (yet):</b> Writer brings your own words to the foreground as you type over AI-generated text. But don’t just change a word and claim it as yours. Rethink and rewrite until you own it.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/don-t-cheat-think-2.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>This is yours:</b> Don’t sweat it if you have one or two common words left over from ChatGPT (here it’s “cheat,” in grey). What matters is that you thought it through by yourself.</figcaption>
 <p></p>
 <p>Continue working on your first draft by cutting and ordering your thoughts. Focus on what you want to say and get yourself into the flow. And you can forget ChatGPT for the next few hours. The story is yours now.</p>
 <h2>4. Editing: Cut, clarify, simplify</h2>
 <p>Idea, structure, first draft, editing, and publishing seem to be separate steps. In reality, they’re connected. Your idea takes form as you write. Writing and editing go hand in hand until the idea wants to be shared. And even while you’re editing, Gemini, ChatGPT, and similar tools can help again.</p>
 <h3>d) But I’m stuck</h3>
 <p><em>Can’t find the right way to say something?</em> Tell ChatGPT to regenerate parts of your document, make it shorter and in the style of your favorite writer to help contrast your writing with somebody else’s.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/rewrite-as-tom-wolfe.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>The same in Tom Wolfe’s Style:</b> Your thoughts in a different style. Don’t copy-paste. Compare and improve.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/rewrite-as-george-carlin.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>The same in George Carlin’s Style:</b> Using comedians to see your thoughts from another angle is a nice way to catch a break, especially when you write about dark matters.</figcaption>
 <p></p>
 <h3>e) But I’m worried I wrote something stupid</h3>
 <p><em>Worried you missed something obvious?</em> You can use ChatGPT as an editor to list potential flaws, like long words, clichés, or factual errors.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/list-long-words-with-alternatives.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Find long words:</b> List long words and suggest shorter alternatives. Don’t simply find and replace. Rethink the sentences.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/rhetoric_devices.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Find rhetoric devices:</b> More food for thought.</figcaption>
 <p></p>
 <p>Use “list” rather than rewrite. Rewriting can lead to all sorts of unwanted changes. They can be hard to spot because ChatGPT seems so smooth, grammatically and orthographically.</p>
 <h3>f) But I just need to correct some commas and typos</h3>
 <p><em>But what if it’s really just commas or typos?</em> Then paste ChatGPT’s text over your own as edits. With Authorship enabled, iA Writer will show you what ChatGPT changed. This way you’ll make sure that it doesn’t rewrite things you said.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/fix-typos-with-chatgpt.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Paste edits as:</b> You can paste the edited text over the text you have written.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/fixed-comma-2.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>AI changes are grey:</b> In this case, ChatGPT has added a comma. Make sure you agree with every change, no matter how small.</figcaption>
 <p></p>
 <h3>g) But I need a second opinion</h3>
 <p><em>What weaknesses have I missed in my writing?</em> Ask ChatGPT for progressively harsher critiques of your writing. That can be unpleasant, but that’s how you find and fix your blind spots, without the shame.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/criticize-as-hard-as-you-can.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Criticize my text:</b> Ask for harsh feedback, otherwise ChatGPT will be too forgiving.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/far-from-done.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Far from done:</b> ChatGPT will find weaknesses. But rather than blindly obeying, rethink weaknesses that you agree with. Use iA Writer’s Syntax Highlight to find more flaws (did you spot the two “now’s”?).</figcaption>
 <p></p>
 <h2>Why bother?</h2>
 <p>Enthusiasts present AI as a magic wand that can solve humanity’s biggest problems. In the meantime, it uses an exponential amount of energy to make everything the same.</p>
 <p>With every thought we outsource, we miss out on a chance to grow. Love it or hate it, AI is here to stay. However we use it, we need to think more, not less.</p>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/8f9d64c455a9246d5e23810ed10e3fe2/index.md
+++ b/cache/2024/8f9d64c455a9246d5e23810ed10e3fe2/index.md
@@ -0,0 +1,101 @@
 title: Turning the Tables on AI
 url: https://ia.net/topics/turning-the-tables-on-ai
 hash_url: 8f9d64c455a9246d5e23810ed10e3fe2
 archive_date: 2024-06-24
 og_image: https://ia.net/wp-content/uploads/2024/06/AI-Turn-Table-Think-Different.png
 description: AI is sold to make us think less... How about using AI not to think less but *more*?
 favicon: https://ia.net/wp-content/themes/iA-library/assets/hotlink-ok/favicon-192x192.png
 language: en_US

 <p>Artificial Intelligence has assimilated Microsoft, Facebook, Google, and now Apple. There is no escape: As soon as you open your computer you’re invited to skip thinking and let the machine fill the void. But what if we used AI not to think less but <em>more</em>?</p>
 <p>AI has one trait: It makes everything the same. Every tech company now offers more or less the same service: “Think less.” How about Apple? Think different? Not anymore.</p>
 <p>Tech companies big and small sell AI as something that thinks for us. It <em>does</em> replace thought with statistics—but it is not intelligent. No one knows what the future will bring. But is a future without thought a better future?</p>
 <p>Now, with a tool that might help us think… How about using AI not to think less but <em>more</em>?</p>
 <p><em>But I don’t know where to start. But I really like that ChatGPT sentence there. But I can’t say it better. But I’m stuck. But I’m worried I wrote something stupid. But I just need to correct some commas and typos. But I need a second opinion.</em></p>
 <p>Let’s turn the tables on our excuses.</p>
 <h2>1. Don’t ask AI, let AI ask you</h2>
 <h3>a) But I don’t know where to start</h3>
 <p>Give ChatGPT an outline and let it write for you. Let’s turn the tables and have ChatGPT prompt <em>us</em>. Tell AI to ask you questions about what you’re writing. Push yourself to express in clear terms what you really want to say. Like this, for example:</p>
 <blockquote>
 <p>I want to write [format] about [topic]. Ask me questions one at a time that force me to explain my idea.</p>
 </blockquote>
 <p>Keep asking until your idea is clear to you. Here’s how such a dialog where ChatGPT prompts <em>you</em> would look:</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/write-for-me.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Write for me:</b> This is how AI is sold to us. But why write if you have nothing to say? Who would read it, and why?</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/ask-ChatGPT.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Ask me instead:</b> Make AI ask *you* questions about your text. Copy-paste your answers. First draft done.</figcaption>
 <p></p>
 <p>Once you have enough to work with, paste <em>your</em> answers into your text editor. You have just written your first draft without cheating.</p>
 <p>ChatGPT may have asked or said some clever things along the way. Feeling tempted to steal it? Don’t. Rather use it as a chance to think and express what you feel.</p>
 <h2>2. Don’t sell stolen goods—make your own</h2>
 <h3>b) But I really like that ChatGPT sentence there</h3>
 <p><em>What if the ChatGPT generated something useful that I want to keep?</em> Paste it as a note <em>Marked as AI.</em> Use quotes, use markup, and note its origin.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/paste-as-AI-from-ChatGPT.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Paste as AI:</b> Want to use AI-generated output? In Writer, right-click and select Paste as AI.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/text-marked-as-ai.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>External sources in Writer:</b> iA Writer greys out text that you marked as AI so you can always discern what is yours and what isn’t.</figcaption>
 <p></p>
 <h2>3. Don’t pretend. Create</h2>
 <h3>c) But I can’t say it better</h3>
 <p><em>What if Artificial Intelligence says exactly what I wanted to say?</em> Rethink and rewrite it to make it your own. Ask if what has been generated is really true. At first glance, it may look tight, but if you really think it through, you’ll find mistakes. Feel it, mean it, think it through. Say it your way.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/cheating-with-ia-writer.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>It’s not yours (yet):</b> Writer brings your own words to the foreground as you type over AI-generated text. But don’t just change a word and claim it as yours. Rethink and rewrite until you own it.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/don-t-cheat-think-2.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>This is yours:</b> Don’t sweat it if you have one or two common words left over from ChatGPT (here it’s “cheat,” in grey). What matters is that you thought it through by yourself.</figcaption>
 <p></p>
 <p>Continue working on your first draft by cutting and ordering your thoughts. Focus on what you want to say and get yourself into the flow. And you can forget ChatGPT for the next few hours. The story is yours now.</p>
 <h2>4. Editing: Cut, clarify, simplify</h2>
 <p>Idea, structure, first draft, editing, and publishing seem to be separate steps. In reality, they’re connected. Your idea takes form as you write. Writing and editing go hand in hand until the idea wants to be shared. And even while you’re editing, Gemini, ChatGPT, and similar tools can help again.</p>
 <h3>d) But I’m stuck</h3>
 <p><em>Can’t find the right way to say something?</em> Tell ChatGPT to regenerate parts of your document, make it shorter and in the style of your favorite writer to help contrast your writing with somebody else’s.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/rewrite-as-tom-wolfe.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>The same in Tom Wolfe’s Style:</b> Your thoughts in a different style. Don’t copy-paste. Compare and improve.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/rewrite-as-george-carlin.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>The same in George Carlin’s Style:</b> Using comedians to see your thoughts from another angle is a nice way to catch a break, especially when you write about dark matters.</figcaption>
 <p></p>
 <h3>e) But I’m worried I wrote something stupid</h3>
 <p><em>Worried you missed something obvious?</em> You can use ChatGPT as an editor to list potential flaws, like long words, clichés, or factual errors.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/list-long-words-with-alternatives.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Find long words:</b> List long words and suggest shorter alternatives. Don’t simply find and replace. Rethink the sentences.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/rhetoric_devices.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Find rhetoric devices:</b> More food for thought.</figcaption>
 <p></p>
 <p>Use “list” rather than rewrite. Rewriting can lead to all sorts of unwanted changes. They can be hard to spot because ChatGPT seems so smooth, grammatically and orthographically.</p>
 <h3>f) But I just need to correct some commas and typos</h3>
 <p><em>But what if it’s really just commas or typos?</em> Then paste ChatGPT’s text over your own as edits. With Authorship enabled, iA Writer will show you what ChatGPT changed. This way you’ll make sure that it doesn’t rewrite things you said.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/fix-typos-with-chatgpt.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Paste edits as:</b> You can paste the edited text over the text you have written.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/fixed-comma-2.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>AI changes are grey:</b> In this case, ChatGPT has added a comma. Make sure you agree with every change, no matter how small.</figcaption>
 <p></p>
 <h3>g) But I need a second opinion</h3>
 <p><em>What weaknesses have I missed in my writing?</em> Ask ChatGPT for progressively harsher critiques of your writing. That can be unpleasant, but that’s how you find and fix your blind spots, without the shame.</p>
 <p><figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/criticize-as-hard-as-you-can.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Criticize my text:</b> Ask for harsh feedback, otherwise ChatGPT will be too forgiving.</figcaption>
 <p>
 <figure class="compare">
 <img decoding="async" src="https://ia.net/wp-content/uploads/2024/06/far-from-done.png" alt class="macOS littleshadow"></figure></p>
 <figcaption><b>Far from done:</b> ChatGPT will find weaknesses. But rather than blindly obeying, rethink weaknesses that you agree with. Use iA Writer’s Syntax Highlight to find more flaws (did you spot the two “now’s”?).</figcaption>
 <p></p>
 <h2>Why bother?</h2>
 <p>Enthusiasts present AI as a magic wand that can solve humanity’s biggest problems. In the meantime, it uses an exponential amount of energy to make everything the same.</p>
 <p>With every thought we outsource, we miss out on a chance to grow. Love it or hate it, AI is here to stay. However we use it, we need to think more, not less.</p>
--- a/cache/2024/92ca1fc6eb390997db6b76afaf742d63/index.html
+++ b/cache/2024/92ca1fc6eb390997db6b76afaf742d63/index.html
@@ -0,0 +1,207 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="fr">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>Cinq idées reçues lors de l’école d’été sur les humanités numériques (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://www.lobrassard.net/carnet/2024-06-14-idees-recues-ecole-ete-humanites-numeriques.html">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>Cinq idées reçues lors de l’école d’été sur les humanités numériques</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://www.lobrassard.net/carnet/2024-06-14-idees-recues-ecole-ete-humanites-numeriques.html" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <p><a href="https://www.michaelsinatra.org/teaching/ete-2024/">L’école d’été « humanités numériques »</a> coordonnée par le professeur <a href="https://www.michaelsinatra.org/">Michael Sinatra</a> réunit chaque année une heureuse ribambelle de professeur·e·s chercheur·euse·s qui présentent, à raison de deux conférences par jour pendant six jours, une facette des <em>humanités numériques</em>.</p>
 <h2 id="une-ia-moins-humaine-pourrait-elle-être-plus-juste-ollivier-dyens">1. Une IA <em>moins humaine</em> pourrait-elle être <em>plus juste</em> ? – Ollivier Dyens</h2>
 <p>L’une des grandes peurs entourant le déploiement des systèmes d’intelligence artificielle (<abbr title="intelligence artificielle">IA</abbr>) concerne ses potentielles « dérives », quand celle-ci pourrait « déraper ». Sauf que, rappelle Ollivier Dyens, l’humanité n’a pas eu besoin de l’<abbr title="intelligence artificielle">IA</abbr> pour commettre des atrocités – camps de concentration, bombe atomique, massacres à la machette et à la baïonnette, etc. <em>Pourquoi devrions-nous avoir davantage peur de l’<abbr title="intelligence artificielle">IA</abbr> ?</em></p>
 <p>L’intelligence artificielle, en tant que reflet de « nous-mêmes », pourrait en fait nous servir de miroir, à nous faire voir ce qui nous caractérise (en tant qu’êtres humains, individuellement et collectivement), à nous montrer nos propres biais – et en particulier ceux que nous n’arrivons pas à discerner nous-mêmes.</p>
 <p>En guise de réflexion, Ollivier Dyens pose la question suivante : <mark>est-ce qu’il n’y aurait pas une possibilité de « sortir » l’humain de la boucle de décision éthique, pour qu’elle soit justement « moins humaine »</mark> – comportant donc moins de biais, moins de cruauté –, ce qui en ferait justement un système peut-être <em>plus juste</em> ?</p>
 <h2 id="toute-expression-artistique-nest-pas-protégeable-par-le-droit-dauteur-au-canada-du-moins-olivier-charbonneau">2. Toute expression artistique n’est pas protégeable par le droit d’auteur <em>(au Canada du moins)</em> – Olivier Charbonneau</h2>
 <p><mark>La danse n’est pas protégée par le droit d’auteur.</mark> En fait, au Canada, elle n’est pas <em>protégeable</em>. Pourquoi ? Parce qu’une chorégraphie est quelque chose d’<em>intangible</em>, de non fixé (une idée n’est d’ailleurs pas protégeable en vertu du droit d’auteur).</p>
 <p>Ce qui est protégeable, en revanche, c’est une <em>fixation de l’œuvre</em>, comme une captation : les droits d’auteur s’appliquent alors sur <em>le fichier vidéo</em> qu’une personne pourrait produire, et les droits reviennent non pas le danseur ou la chorégraphe, mais bien la personne derrière la caméra, qui est responsable de la fixation de l’œuvre.</p>
 <p>Sur quoi se base-t-on pour établir le droit d’auteur ? Sur le concept d’<em>originalité juridique</em>, dont la définition formelle <a href="https://www.culturelibre.ca/2023/06/19/droit-dauteur-sur-les-donnees-en-humanites-numeriques/">se résume à une formule étonnamment courte</a> :</p>
 <blockquote>
 <p>Talent, jugement et effort dans la sélection et l’arrangement des éléments<span class="citation" data-cites="Charbonneau2023"></span>.</p>
 </blockquote>
 <p>Une <a href="https://www.canlii.org/fr/ca/caf/doc/1997/1997canlii6378/1997canlii6378.html">compilation</a> (données, analyse, listes d’œuvres, etc.) pourrait ou non être protégée en vertu du droit d’auteur, en fonction de son originalité – et toujours à condition d’être fixée sur un support quelconque.</p>
 <h2 id="lévaluation-par-les-pairs-nest-pas-un-gage-de-fiabilité-vincent-larivière">3. L’évaluation par les pairs n’est pas un gage de fiabilité – Vincent Larivière</h2>
 <p>L’évaluation par les pairs constitue le principal critère de « scientificité » d’un document, ce qui le distingue par exemple d’un billet d’humeur publié un blogue personnel. C’est un filtre, mais elle n’empêche pas la publication : à terme, environ 90% des articles soumis seront publiés, peut-être simplement dans des revues moins prestigieuses.</p>
 <p><mark>La prépublication est plus fiable qu’on pourrait le penser</mark> : 80% des articles ne sont à peu près pas changés après la révision par les pairs ; l’évaluation n’apporte généralement pas de différences significatives en matière de contenu. Il ne s’agit pas de dire que l’évaluation par les pairs est inutile, mais que la mise à disposition avant la publication dite « officielle » (qui peut prendre des années après l’obtention des résultats de recherche) est peut-être <em>moins essentielle</em> qu’on le pense en matière de diffusion.</p>
 <h2 id="déformer-une-œuvre-pour-la-lire-autrement-marta-boni">4. Déformer une œuvre pour la lire autrement – Marta Boni</h2>
 <p>Professeure spécialisée dans les études télévisuelles, Marta Boni a montré qu’une lecture performative permet de « lire » une œuvre vidéo d’une tout autre manière. Par « l’exploration vidéo », des chercheurs comme Jason Mittell utilisent les techniques de montage pour déformer, puis re-former une œuvre (comme l’analyse des personnages de <cite>Breaking Bad</cite> faite par Mittell<span class="citation" data-cites="Mittell2024"></span>).</p>
 <div class="margin-note">
 <p>Parmi les exemples de « déformation<span class="citation" data-cites="Mittell2021"></span> » et de « lecture oppositionnelle<span class="citation" data-cites="Hall1980"></span> » (<em>cf.</em> Stuart Hall), Mittel a superposé tous les plans d’un film afin d’en produire <a href="http://www.digitalhumanities.org/dhq/vol/15/1/000521/000521.html">une image inédite</a>.</p>
 </div>
 <p><mark>L’exploration vidéo peut ainsi être utilisée pour faire des découvertes imprévues</mark> ; elle ouvre la voie à de nouvelles façons d’entrer en contact avec les œuvres, de manière performative. Surtout, elle permet de dépasser la notion de « faire du sens » pour en arriver à « faire autre chose ».</p>
 <h2 id="qui-contrôle-la-mémoire-visuelle-julien-shuh">5. Qui contrôle la mémoire visuelle ? – Julien Shuh</h2>
 <p>Avec l’IA, de nouveaux acteurs entrent dans l’équation de la <em>mémoire culturelle collective</em> : ce qu’on conserve (et ce qu’on oublie), comment on le conserve (ce qu’on numérise ou non), ce qu’on choisit de montrer et de diffuser, etc. Constat : ce qu’on peut voir sur une plateforme de visionnement en ligne comme Netflix finit par être beaucoup plus regardé, et les films plus anciens « disparaissent » avec les algorithmes de recommandation. On regarde davantage ce qui est présenté, et on regarde moins ce qui ne l’est pas. <mark>Les outils qui contrôlent « ce qui est visible » ont aujourd’hui un impact important sur la société</mark>, et en particulier sur la mémoire visuelle collective.</p>
 <p>Autre exemple : les Google Doodle, qui rendent saillants (en fonction du pays) des événements historiques « dignes de mention ». Cela rend <em>davantage visibles</em> certains événements et en <em>invisibilise</em> d’autres. Un clic sur le Doodle du jour mène à une requête sur le moteur de recherche, laquelle présente un fait historique d’une certaine manière (la page de résultats de Google). Le premier résultat est la page Wikipédia ; les gens ne vont donc pas cliquer au hasard sur Internet, mais sur des liens présélectionnés par Google. Et la page Wikipédia (de surcroît la plus visitée) contient désormais du texte <a href="https://fr.wikipedia.org/wiki/Wikip%C3%A9dia:Illustrations_g%C3%A9n%C3%A9r%C3%A9es_par_une_intelligence_artificielle">et de nombreuses images</a> générés par <abbr title="Acronyme: intelligence artificielle">IA</abbr>, dont les contextes de génération, de sélection et de diffusion ne sont pas tous transparents…</p>
 <p>Par exemple : une <a href="https://commons.wikimedia.org/wiki/File:Gainsboug-Gall-1965-Italie.png">photographie de France Gall</a> téléversée sur Wikimedia Commons a été colorisée, mais aussi retouchée pour effacer des sujets dans l’arrière-plan (voir la <a href="https://commons.wikimedia.org/wiki/File_talk:Gainsboug-Gall-1965-Italie.png#Des_%C3%A9l%C3%A9ments_de_l'image_d'origine_ont_%C3%A9t%C3%A9_effac%C3%A9s">section discussion</a>). La version retouchée figure d’ailleurs sur la <a href="https://fr.wikipedia.org/wiki/France_Gall">page Wikipédia de France Gall</a>. Enfin, la page web citée comme provenance comporte plutôt une image rognée qui n’est fort probablement pas la source du fichier.</p>
 <figure>
 <img src="https://www.lobrassard.net/img/2024-06-14-photographie-france-gall-comparaison.jpg" alt="Comparaison entre une numérisation de la photogarphie d’origine (obtenue sur le site Gala.fr) et une image retouchée, mise en ligne sur Wikimedia Commons.">
 <figcaption aria-hidden="true">Comparaison entre une numérisation de la photogarphie d’origine (obtenue sur le site <a href="https://photo.gala.fr/photos-serge-gainsbourg-les-femmes-de-sa-vie-44322">Gala.fr</a>) et une image retouchée, mise en ligne sur Wikimedia Commons.</figcaption>
 </figure>
 <p>Un tel cas de figure fait surgir plusieurs questions : d’où provient ou telle image ? pourquoi a-t-elle été sélectionnée plutôt qu’une autre ? par qui ? au nom de quelle autorité ? quelle proportion d’une image retouchée renvoie-t-elle à l’orginal, et quelle est la part de transformation ? quel est le sens de la circulation d’une telle image, transformée ou générée ? par qui ou par quoi est-elle passée avant de nous parvenir, et pour quelles raisons ? Je remercie Juien Schuh pour une anecdote aussi foisonnante en questions…</p>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/92ca1fc6eb390997db6b76afaf742d63/index.md
+++ b/cache/2024/92ca1fc6eb390997db6b76afaf742d63/index.md
@@ -0,0 +1,41 @@
 title: Cinq idées reçues lors de l’école d’été sur les humanités numériques
 url: https://www.lobrassard.net/carnet/2024-06-14-idees-recues-ecole-ete-humanites-numeriques.html
 hash_url: 92ca1fc6eb390997db6b76afaf742d63
 archive_date: 2024-06-24
 og_image: https://www.lobrassard.net/img/couverture.png
 description: Carnet de recherche de Louis-Olivier Brassard.
 favicon: data:;base64,iVBORw0KGgo=
 language: fr_FR

 <p><a href="https://www.michaelsinatra.org/teaching/ete-2024/">L’école d’été « humanités numériques »</a> coordonnée par le professeur <a href="https://www.michaelsinatra.org/">Michael Sinatra</a> réunit chaque année une heureuse ribambelle de professeur·e·s chercheur·euse·s qui présentent, à raison de deux conférences par jour pendant six jours, une facette des <em>humanités numériques</em>.</p>
 <h2 id="une-ia-moins-humaine-pourrait-elle-être-plus-juste-ollivier-dyens">1. Une IA <em>moins humaine</em> pourrait-elle être <em>plus juste</em> ? – Ollivier Dyens</h2>
 <p>L’une des grandes peurs entourant le déploiement des systèmes d’intelligence artificielle (<abbr title="intelligence artificielle">IA</abbr>) concerne ses potentielles « dérives », quand celle-ci pourrait « déraper ». Sauf que, rappelle Ollivier Dyens, l’humanité n’a pas eu besoin de l’<abbr title="intelligence artificielle">IA</abbr> pour commettre des atrocités – camps de concentration, bombe atomique, massacres à la machette et à la baïonnette, etc. <em>Pourquoi devrions-nous avoir davantage peur de l’<abbr title="intelligence artificielle">IA</abbr> ?</em></p>
 <p>L’intelligence artificielle, en tant que reflet de « nous-mêmes », pourrait en fait nous servir de miroir, à nous faire voir ce qui nous caractérise (en tant qu’êtres humains, individuellement et collectivement), à nous montrer nos propres biais – et en particulier ceux que nous n’arrivons pas à discerner nous-mêmes.</p>
 <p>En guise de réflexion, Ollivier Dyens pose la question suivante : <mark>est-ce qu’il n’y aurait pas une possibilité de « sortir » l’humain de la boucle de décision éthique, pour qu’elle soit justement « moins humaine »</mark> – comportant donc moins de biais, moins de cruauté –, ce qui en ferait justement un système peut-être <em>plus juste</em> ?</p>
 <h2 id="toute-expression-artistique-nest-pas-protégeable-par-le-droit-dauteur-au-canada-du-moins-olivier-charbonneau">2. Toute expression artistique n’est pas protégeable par le droit d’auteur <em>(au Canada du moins)</em> – Olivier Charbonneau</h2>
 <p><mark>La danse n’est pas protégée par le droit d’auteur.</mark> En fait, au Canada, elle n’est pas <em>protégeable</em>. Pourquoi ? Parce qu’une chorégraphie est quelque chose d’<em>intangible</em>, de non fixé (une idée n’est d’ailleurs pas protégeable en vertu du droit d’auteur).</p>
 <p>Ce qui est protégeable, en revanche, c’est une <em>fixation de l’œuvre</em>, comme une captation : les droits d’auteur s’appliquent alors sur <em>le fichier vidéo</em> qu’une personne pourrait produire, et les droits reviennent non pas le danseur ou la chorégraphe, mais bien la personne derrière la caméra, qui est responsable de la fixation de l’œuvre.</p>
 <p>Sur quoi se base-t-on pour établir le droit d’auteur ? Sur le concept d’<em>originalité juridique</em>, dont la définition formelle <a href="https://www.culturelibre.ca/2023/06/19/droit-dauteur-sur-les-donnees-en-humanites-numeriques/">se résume à une formule étonnamment courte</a> :</p>
 <blockquote>
 <p>Talent, jugement et effort dans la sélection et l’arrangement des éléments<span class="citation" data-cites="Charbonneau2023"></span>.</p>
 </blockquote>
 <p>Une <a href="https://www.canlii.org/fr/ca/caf/doc/1997/1997canlii6378/1997canlii6378.html">compilation</a> (données, analyse, listes d’œuvres, etc.) pourrait ou non être protégée en vertu du droit d’auteur, en fonction de son originalité – et toujours à condition d’être fixée sur un support quelconque.</p>
 <h2 id="lévaluation-par-les-pairs-nest-pas-un-gage-de-fiabilité-vincent-larivière">3. L’évaluation par les pairs n’est pas un gage de fiabilité – Vincent Larivière</h2>
 <p>L’évaluation par les pairs constitue le principal critère de « scientificité » d’un document, ce qui le distingue par exemple d’un billet d’humeur publié un blogue personnel. C’est un filtre, mais elle n’empêche pas la publication : à terme, environ 90% des articles soumis seront publiés, peut-être simplement dans des revues moins prestigieuses.</p>
 <p><mark>La prépublication est plus fiable qu’on pourrait le penser</mark> : 80% des articles ne sont à peu près pas changés après la révision par les pairs ; l’évaluation n’apporte généralement pas de différences significatives en matière de contenu. Il ne s’agit pas de dire que l’évaluation par les pairs est inutile, mais que la mise à disposition avant la publication dite « officielle » (qui peut prendre des années après l’obtention des résultats de recherche) est peut-être <em>moins essentielle</em> qu’on le pense en matière de diffusion.</p>
 <h2 id="déformer-une-œuvre-pour-la-lire-autrement-marta-boni">4. Déformer une œuvre pour la lire autrement – Marta Boni</h2>
 <p>Professeure spécialisée dans les études télévisuelles, Marta Boni a montré qu’une lecture performative permet de « lire » une œuvre vidéo d’une tout autre manière. Par « l’exploration vidéo », des chercheurs comme Jason Mittell utilisent les techniques de montage pour déformer, puis re-former une œuvre (comme l’analyse des personnages de <cite>Breaking Bad</cite> faite par Mittell<span class="citation" data-cites="Mittell2024"></span>).</p>
 <div class="margin-note">
 <p>Parmi les exemples de « déformation<span class="citation" data-cites="Mittell2021"></span> » et de « lecture oppositionnelle<span class="citation" data-cites="Hall1980"></span> » (<em>cf.</em> Stuart Hall), Mittel a superposé tous les plans d’un film afin d’en produire <a href="http://www.digitalhumanities.org/dhq/vol/15/1/000521/000521.html">une image inédite</a>.</p>
 </div>
 <p><mark>L’exploration vidéo peut ainsi être utilisée pour faire des découvertes imprévues</mark> ; elle ouvre la voie à de nouvelles façons d’entrer en contact avec les œuvres, de manière performative. Surtout, elle permet de dépasser la notion de « faire du sens » pour en arriver à « faire autre chose ».</p>
 <h2 id="qui-contrôle-la-mémoire-visuelle-julien-shuh">5. Qui contrôle la mémoire visuelle ? – Julien Shuh</h2>
 <p>Avec l’IA, de nouveaux acteurs entrent dans l’équation de la <em>mémoire culturelle collective</em> : ce qu’on conserve (et ce qu’on oublie), comment on le conserve (ce qu’on numérise ou non), ce qu’on choisit de montrer et de diffuser, etc. Constat : ce qu’on peut voir sur une plateforme de visionnement en ligne comme Netflix finit par être beaucoup plus regardé, et les films plus anciens « disparaissent » avec les algorithmes de recommandation. On regarde davantage ce qui est présenté, et on regarde moins ce qui ne l’est pas. <mark>Les outils qui contrôlent « ce qui est visible » ont aujourd’hui un impact important sur la société</mark>, et en particulier sur la mémoire visuelle collective.</p>
 <p>Autre exemple : les Google Doodle, qui rendent saillants (en fonction du pays) des événements historiques « dignes de mention ». Cela rend <em>davantage visibles</em> certains événements et en <em>invisibilise</em> d’autres. Un clic sur le Doodle du jour mène à une requête sur le moteur de recherche, laquelle présente un fait historique d’une certaine manière (la page de résultats de Google). Le premier résultat est la page Wikipédia ; les gens ne vont donc pas cliquer au hasard sur Internet, mais sur des liens présélectionnés par Google. Et la page Wikipédia (de surcroît la plus visitée) contient désormais du texte <a href="https://fr.wikipedia.org/wiki/Wikip%C3%A9dia:Illustrations_g%C3%A9n%C3%A9r%C3%A9es_par_une_intelligence_artificielle">et de nombreuses images</a> générés par <abbr title="Acronyme: intelligence artificielle">IA</abbr>, dont les contextes de génération, de sélection et de diffusion ne sont pas tous transparents…</p>
 <p>Par exemple : une <a href="https://commons.wikimedia.org/wiki/File:Gainsboug-Gall-1965-Italie.png">photographie de France Gall</a> téléversée sur Wikimedia Commons a été colorisée, mais aussi retouchée pour effacer des sujets dans l’arrière-plan (voir la <a href="https://commons.wikimedia.org/wiki/File_talk:Gainsboug-Gall-1965-Italie.png#Des_%C3%A9l%C3%A9ments_de_l'image_d'origine_ont_%C3%A9t%C3%A9_effac%C3%A9s">section discussion</a>). La version retouchée figure d’ailleurs sur la <a href="https://fr.wikipedia.org/wiki/France_Gall">page Wikipédia de France Gall</a>. Enfin, la page web citée comme provenance comporte plutôt une image rognée qui n’est fort probablement pas la source du fichier.</p>
 <figure>
 <img src="https://www.lobrassard.net/img/2024-06-14-photographie-france-gall-comparaison.jpg" alt="Comparaison entre une numérisation de la photogarphie d’origine (obtenue sur le site Gala.fr) et une image retouchée, mise en ligne sur Wikimedia Commons.">
 <figcaption aria-hidden="true">Comparaison entre une numérisation de la photogarphie d’origine (obtenue sur le site <a href="https://photo.gala.fr/photos-serge-gainsbourg-les-femmes-de-sa-vie-44322">Gala.fr</a>) et une image retouchée, mise en ligne sur Wikimedia Commons.</figcaption>
 </figure>
 <p>Un tel cas de figure fait surgir plusieurs questions : d’où provient ou telle image ? pourquoi a-t-elle été sélectionnée plutôt qu’une autre ? par qui ? au nom de quelle autorité ? quelle proportion d’une image retouchée renvoie-t-elle à l’orginal, et quelle est la part de transformation ? quel est le sens de la circulation d’une telle image, transformée ou générée ? par qui ou par quoi est-elle passée avant de nous parvenir, et pour quelles raisons ? Je remercie Juien Schuh pour une anecdote aussi foisonnante en questions…</p>

--- a/cache/2024/93c287363dbe9cae22f4360c5b107971/index.html
+++ b/cache/2024/93c287363dbe9cae22f4360c5b107971/index.html
@@ -0,0 +1,313 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="en">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>Generative AI Is Not Going To Build Your Engineering Team For You (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://stackoverflow.blog/2024/06/10/generative-ai-is-not-going-to-build-your-engineering-team-for-you/">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>Generative AI Is Not Going To Build Your Engineering Team For You</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://stackoverflow.blog/2024/06/10/generative-ai-is-not-going-to-build-your-engineering-team-for-you/" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <div itemprop="articleBody" class="s-prose fs-subheading"><p>When I was 19 years old, I dropped out of college and moved to San Francisco. I had a job offer in hand to be a Unix sysadmin for Taos Consulting. However, before my first day of work I was lured away to a startup in the city, where I worked as a software engineer on mail subsystems.</p>

    <p>I never questioned whether or not I could find work. Jobs were plentiful, and more importantly, hiring standards were very low. If you knew how to sling HTML or find your way around a command line, chances were you could find <em>someone</em> to pay you.</p>

    <p>Was I some kind of genius, born with my hands on a computer keyboard? Assuredly not. I was homeschooled in the backwoods of Idaho. I didn’t touch a computer until I was sixteen and in college. I escaped to university on a classical performance piano scholarship, which I later traded in for a peripatetic series of nontechnical majors: classical Latin and Greek, musical theory, philosophy. Everything I knew about computers I learned on the job, doing sysadmin work for the university and CS departments.</p>

 <p>In retrospect, I was so lucky to enter the industry when I did. It makes me blanch to think of what would have happened if I had come along a few years later. Every one of the ladders my friends and I took into the industry has long since vanished.</p><p>To some extent, this is just what happens as an industry matures. The early days of any field are something of a Wild West, where the stakes are low, regulation nonexistent, and standards nascent. If you look at the early history of other industries—medicine, cinema, radio—the similarities are striking.</p>

 <p>There is a magical moment with any young technology where the boundaries between roles are porous and opportunity can be seized by anyone who is motivated, curious, and willing to work their asses off.</p>

 <p>It never lasts. It can’t; it shouldn’t. The amount of prerequisite knowledge and experience you must have before you can enter the industry swells precipitously. The stakes rise, the magnitude of the mission increases, the cost of mistakes soars. We develop certifications, trainings, standards, legal rites. We wrangle over whether or not <a href="https://www.hillelwayne.com/post/are-we-really-engineers/">software engineers are really engineers</a>.</p><p>Nowadays, you wouldn’t want a teenaged dropout like me to roll out of junior year and onto your pager rotation. The prerequisite knowledge you need to enter the industry has grown, the pace is faster, and the stakes are much higher, so you can no longer learn literally <em>everything</em> on the job, as I once did.</p>

 <p>However, it’s not like you can learn everything you need to know at college either. A CS degree typically prepares you better for a life of computing research than life as a workaday software engineer. A more practical path into the industry may be a good coding bootcamp, with its emphasis on problem solving and learning a modern toolkit. In either case, you don’t so much learn “how to do the job” as you do “learn enough of the basics to understand and use the tools you need to use to learn the job.”</p>

 <p>Software is an apprenticeship industry. You can’t learn to be a software engineer by reading books. You can only learn by doing…and doing, and doing, and doing some more. No matter what your education consists of, most learning happens on the job—period. And it never ends! Learning and teaching are lifelong practices; they have to be, the industry changes so fast.</p>

 <p>It takes a solid seven-plus years to forge a competent software engineer. (Or as most job ladders would call it, a “senior software engineer”.) That’s many years of writing, reviewing, and deploying code every day, on a team alongside more experienced engineers. That’s just how long it seems to take.</p><p>Here is where I often get some very indignant pushback to my timelines, e.g.:</p>

 <p>“Seven years?! Pfft, it took me two years!”</p><p>“I was promoted to Senior Software Engineer in less than five years!”</p>

 <p>Good for you. True, there is nothing magic about seven years. But it takes time and experience to mature into an experienced engineer, the kind who can anchor a team. More than that, it takes <em>practice</em>.</p>

 <p>I think we have come to use “Senior Software Engineer” as shorthand for engineers who can ship code and be a net positive in terms of productivity, and I think that’s a huge mistake. It implies that less senior engineers must be a net negative in terms of productivity, which is untrue. And it elides the real nature of the work of software engineering, of which writing code is only a small part.</p>

 <p>To me, being a senior engineer is not primarily a function of your ability to write code. It has far more to do with your ability to understand, maintain, explain, and manage a large body of software in production over time, as well as the ability to translate business needs into technical implementation. So much of the work is around crafting and curating these large, complex sociotechnical systems, and code is just one representation of these systems.</p>

 <p>What does it mean to be a senior engineer? It means you have learned how to <em>learn</em>, first and foremost, and how to teach; how to hold these models in your head and reason about them, and how to maintain, extend, and operate these systems over time. It means you have good judgment, and instincts you can trust.</p>

 <p>Which brings us to the matter of AI.</p><p>It is really, really tough to get your first role as an engineer. I didn’t realize how hard it was until I watched my little sister (new grad, terrific grades, some hands on experience, fiendishly hard worker) struggle for nearly <em>two years</em> to land a real job in her field. That was a few years ago; anecdotally, it seems to have gotten even harder since then.</p>

 <p>This past year, I have read a steady drip of articles about <a href="https://www.nytimes.com/2024/04/10/business/investment-banking-jobs-artificial-intelligence.html">entry-level jobs in various industries</a> being replaced by AI. Some of which absolutely have merit. Any job that consists of drudgery such as converting a document from one format to another, reading and summarizing a bunch of text, or replacing one set of icons with another, seems pretty obviously vulnerable. This doesn’t feel all that revolutionary to me, it’s just extending the existing boom in automation to cover textual material as well as mathy stuff.</p>

 <p>Recently, however, a number of execs and so-called “thought leaders” in tech seem to have genuinely convinced themselves that generative AI is on the verge of replacing all the work done by junior engineers. I have read so many articles about how junior engineering work is being automated out of existence, or that the need for junior engineers is shriveling up. It has officially driven me bonkers.</p>

 <p>All of this bespeaks a deep misunderstanding about what engineers actually <em>do</em>. By not hiring and training up junior engineers, we are cannibalizing our own future. We need to <em>stop doing that</em>.</p><p>People act like writing code is the hard part of software. It is not. It never has been, it never will be. <strong>Writing code is the easiest part of software engineering</strong>, and it’s getting easier by the day. The hard parts are what you do with that code—operating it, understanding it, extending it, and governing it over its entire lifecycle.</p>

 <p>A junior engineer begins by learning how to write and debug lines, functions, and snippets of code. As you practice and progress towards being a senior engineer, you learn to compose systems out of software, and guide systems through waves of change and transformation.</p>

 <p>Sociotechnical systems consist of software, tools, and people; understanding them requires familiarity with the interplay between software, users, production, infrastructure, and continuous changes over time. These systems are fantastically complex and subject to chaos, nondeterminism and emergent behaviors. If anyone claims to understand the system they are developing and operating, the system is either exceptionally small or (more likely) they don’t know enough to know what they don’t know. Code is easy, in other words, but <em>systems are hard</em>.</p>

 <p>The present wave of generative AI tools has done a lot to help us generate lots of code, very fast. The easy parts are becoming even easier, at a truly remarkable pace. But it has not done a thing to aid in the work of managing, understanding, or operating that code. If anything, it has only made the hard jobs harder.</p><p>But that code first has to be understood, adapted and integrated cleanly into the system.</p><p>If you read a lot of breathless think pieces, you may have a mental image of software engineers merrily crafting prompts for ChatGPT, or using Copilot to generate reams of code, then committing whatever emerges to GitHub and walking away. That does not resemble our reality.</p>

 <p>The right way to think about tools like Copilot is more like a really fancy autocomplete or copy-paste function, or maybe like the unholy love child of Stack Overflow search results plus Google’s “I feel lucky”. You roll the dice, every time.</p>

 <p>These tools are at their best when there’s already a parallel in the file, and you want to just copy-paste the thing with slight modifications. Or when you’re writing tests and you have a giant block of fairly repetitive YAML, and it repeats the pattern while inserting the right column and field names, like an automatic template.</p>

 <p>However, <strong>you cannot trust generated code</strong>. I can’t emphasize this enough. AI-generated code always looks quite plausible, but even when it kind of “works”, it’s rarely congruent with your wants and needs. It will happily generate code that doesn’t parse or compile. It will make up variables, method names, function calls; it will hallucinate fields that don’t exist. Generated code will not follow your coding practices or conventions. It is not going to refactor or come up with intelligent abstractions for you. The more important, difficult or meaningful a piece of code is, the less likely you are to generate a usable artifact using AI.</p>

 <p>You may save time by not having to type the code in from scratch, but you will need to step through the output line by line, revising as you go, before you can commit your code, let alone ship it to production. In many cases this will take as much or more time as it would take to simply write the code—especially these days, now that autocomplete has gotten so clever and sophisticated. It can be a LOT of work to bring AI-generated code into compliance and coherence with the rest of your codebase. It isn’t always worth the effort, quite frankly.</p>

 <p>Generating code that can compile, execute, and pass a test suite isn’t especially hard; the hard part is crafting a code base that many people, teams, and successive generations of teams can navigate, mutate, and reason about for years to come.</p><p>So that’s the TLDR: you can generate a lot of code, really fast, but you can’t trust what comes out. At all. However, there are some use cases where generative AI consistently shines.</p>

 <p>For example, it’s often easier to ask chatGPT to generate example code using unfamiliar APIs than by reading the API docs—the corpus was trained on repositories where the APIs are being used for real life workloads, after all.</p>

 <p>Generative AI is also pretty good at producing code that is annoying or tedious to write, yet tightly scoped and easy to explain. The more predictable a scenario is, the better these tools are at writing the code for you. If what you need is effectively copy-paste with a template—any time you could generate the code you want using sed/awk or vi macros—generative AI is quite good at this.</p>

 <p>It’s also very good at writing little functions for you to do things in unfamiliar languages or scenarios. If you have a snippet of Python code and you want the same thing in Java, but you don’t know Java, generative AI has got your back.</p>

 <p>Again, remember, the odds are 50/50 that the result is completely made up. You always have to assume the results are incorrect until you can verify it by hand. But these tools can absolutely accelerate your work in countless ways.</p><p>One of the engineers I work with, Kent Quirk, describes generative AI as “an excitable junior engineer who types really fast”. I love that quote—it leaves an indelible mental image.</p>

 <p>Generative AI is like a junior engineer in that you can’t roll their code off into production. You are responsible for it—legally, ethically, and practically. You still have to take the time to understand it, test it, instrument it, retrofit it stylistically and thematically to fit the rest of your code base, and ensure your teammates can understand and maintain it as well.</p>

 <p>The analogy is a decent one, actually, but <em>only</em> if your code is disposable and self-contained, i.e. not meant to be integrated into a larger body of work, or to survive and be read or modified by others.</p>

 <p>And hey—there are corners of the industry like this, where most of the code is write-only, throwaway code. There are agencies that spin out dozens of disposable apps per year, each written for a particular launch or marketing event and then left to wither on the vine. <em>But that is not most software.</em> Disposable code is rare; code that needs to work over the long term is the norm. Even when we think a piece of code will be disposable, we are often (urf) wrong.</p><p>In that particular sense—generating code that you know is untrustworthy—GenAI is a bit like a junior engineer. But in every other way, the analogy fails. Because adding a person who writes code to your team is nothing like autogenerating code. That code could have come from anywhere—Stack Overflow, Copilot, whatever. You don’t know, and it doesn’t really matter. There’s no feedback loop, no person on the other end trying iteratively to learn and improve, and no impact to your team vibes or culture.</p>

 <p>To state the supremely obvious: giving code review feedback to a junior engineer is not like editing generated code. Your effort is worth more when it is invested into someone else’s apprenticeship. It’s an opportunity to pass on the lessons you’ve learned in your own career. Even just the act of framing your feedback to explain and convey your message forces you to think through the problem in a more rigorous way, and has a way of helping you understand the material more deeply.</p>

 <p>And adding a junior engineer to your team will immediately change team dynamics. It creates an environment where asking questions is normalized and encouraged, where teaching as well as learning is a constant. We’ll talk more about team dynamics in a moment.</p>

 <p>The time you invest into helping a junior engineer level up can pay off remarkably quickly. Time flies. ☺️ When it comes to hiring, we tend to valorize senior engineers almost as much as we underestimate junior engineers. Neither stereotype is helpful.</p><p>People seem to think that once you hire a senior engineer, you can drop them onto a team and they will be immediately productive, while hiring a junior engineer will be a tax on team performance forever. Neither are true. Honestly, <em>most</em> of the work that <em>most</em> teams have to do is not that difficult, once it’s been broken down into its constituent parts. There’s plenty of room for lower level engineers to execute and flourish.</p>

 <p>The grossly simplified perspective of your accountant goes something like this. “Why should we pay $100k for a junior engineer to slow things down, when we could pay $200k for a senior engineer to speed things up?” It makes no sense!</p>

 <p>But you know and I know—every engineer who is paying attention should know—<strong>that’s not how engineering works</strong>. This is an apprenticeship industry, and productivity is defined by the output and carrying capacity of each team, not each person.</p>

 <p>There are lots of ways a person can contribute to the overall velocity of a team, just like there are lots of ways a person can sap the energy out of a team or add friction and drag to everyone around them. These do not always correlate with the person’s level (at least not in the direction people tend to assume), and writing code is only one way.</p>

 <p>Furthermore, every engineer you hire requires ramp time and investment before they can contribute. Hiring and training new engineers is a costly endeavor, no matter what level they are. It will take any senior engineer time to build up their mental model of the system, familiarize themselves with the tools and technology, and ramp up to speed. How long? It depends on how clean and organized the codebase is, past experience with your tools and technologies, how good you are at onboarding new engineers, and more, but likely around 6-9 months. They probably won’t reach cruising altitude for about a year.</p>

 <p>Yes, the ramp will be longer for a junior engineer, and yes, it will require more investment from the team. But it’s not indefinite. Your junior engineer should be a net positive within roughly the same time frame, six months to a year, and they develop far more rapidly than more senior contributors. (Don’t forget, their contributions may vastly exceed the code they personally write.)</p><p>In terms of writing and shipping features, some of the most productive engineers I’ve ever known have been intermediate engineers. Not yet bogged down with all the meetings and curating and mentoring and advising and architecture, their calendars not yet pockmarked with interruptions, they can <strong>just build stuff</strong>. You see them put their headphones on first thing in the morning, write code all day, and cruise out the door in the evening having made incredible progress.</p>

 <p>Intermediate engineers sit in this lovely, temporary state where they have gotten good enough at programming to be very productive, but they are still learning how to build and care for systems. All they do is write code, reams and reams of code.</p>

 <p>And they’re energized…engaged. They’re having fun! They aren’t bored with writing a web form or a login page for the 1000th time. Everything is new, interesting, and exciting, which typically means they will <em>do a better job</em>, especially under the light direction of someone more experienced. Having intermediate engineers on a team is amazing. The only way you get them is by hiring junior engineers.</p>

 <p>Having junior and intermediate engineers on a team is a shockingly good inoculation against overengineering and premature complexity. They don’t yet know enough about a problem to imagine all the infinite edge cases that need to be solved for. They help keep things simple, which is one of the hardest things to do.</p><p>If you ask, nearly everybody will wholeheartedly agree that hiring junior engineers is a good thing…and someone else should do it. This is because the long-term arguments for hiring junior engineers are compelling and fairly well understood.</p>

 <ol><li>We need more senior engineers as an industry</li><li><em>Somebody</em> has to train them</li><li>Junior engineers are cheaper</li><li>They may add some much-needed diversity</li><li>They are often very loyal to companies who invest in training them, and will stick around for years instead of job hopping</li><li>Did we already mention that somebody needs to do it?</li></ol>

 <p>But long-term thinking is not a thing that companies, or capitalism in general, are typically great at. Framed this way, it makes it sound like you hire junior engineers as a selfless act of public service, at great cost to yourself. Companies are much more likely to want to externalize costs like those, which is how we got to where we are now.</p><p>However, there are at least as many arguments to be made for hiring junior engineers in the short term—selfish, hard-nosed, profitable reasons for why it benefits the team and the company to do so. You just have to shift your perspective slightly, from individuals to teams, to bring them into focus.</p>

 <p>Let’s start here: hiring engineers is not a process of “picking the best person for the job”. <strong>Hiring engineers is about composing teams</strong>. The smallest unit of software ownership is not the individual, it’s the team. Only teams can own, build, and maintain a corpus of software. It is inherently a collaborative, cooperative activity.</p>

 <p>If hiring engineers was about picking the “best people”, it would make sense to hire the most senior, experienced individual you can get for the money you have, because we are using “senior” and “experienced” as a proxy for “productivity”. (Questionable, but let’s not nitpick.) But the productivity of each individual is not what we should be optimizing for. The productivity of the <em>team</em> is all that matters.</p>

 <p>And the best teams are <em>always</em> the ones with a diversity of strengths, perspectives, and levels of expertise. A monoculture can be spectacularly successful in the short term—it may even outperform a diverse team. But they do not scale well, and they do not adapt to unfamiliar challenges gracefully. The longer you wait to diversify, the harder it will be.</p>

 <p>We need to hire junior engineers, and not just once, but consistently. We need to keep feeding the funnel from the bottom up. Junior engineers only stay junior for a couple years, and intermediate engineers turn into senior engineers. Super-senior engineers are not actually the best people to mentor junior engineers; the most effective mentor is usually someone just one level ahead, who vividly remembers what it was like in your shoes.</p><p>A healthy team is an ecosystem. You wouldn’t staff a product engineering team with six DB experts and one mobile developer. Nor should you staff it with six staff+ engineers and one junior developer. A good team is composed of a range of skills and levels.</p>

 <p>Have you ever been on a team packed exclusively with staff or principal engineers? It is <em>not fun</em>. That is not a high-functioning team. There is only so much high-level architecture and planning work to go around, there are only so many big decisions that need to be made. These engineers spend most of their time doing work that feels boring and repetitive, so they tend to over-engineer solutions and/or cut corners—sometimes at the same time. They compete for the “fun” stuff and find reasons to pick technical fights with each other. They chronically under-document and under-invest in the work that makes systems simple and tractable.</p>

 <p>Teams that only have intermediate engineers (or beginners, or seniors, or whatever) will have different pathologies, but similar problems with contention and blind spots. The work itself has a wide range in complexity and difficulty—from simple, tightly scoped functions to tough, high-stakes architecture decisions. It makes sense for the people doing the work to occupy a similar range.</p>

 <p>The best teams are ones where no one is bored, because every single person is working on something that challenges them and pushes their boundaries. The only way you can get this is by having a range of skill levels on the team.</p><p>The bottleneck we face now is not our ability to train up new junior engineers and give them skills. Nor is it about juniors learning to hustle harder; I see a lot of <a href="https://chioualexander.medium.com/how-to-effectively-apply-to-jobs-as-a-junior-engineer-37a90312d4fd">solid</a>, <a href="https://blog.pragmaticengineer.com/advice-for-junior-software-engineers/">well-meaning advice</a> on this topic, but it’s not going to solve the problem. <strong>The bottleneck is giving them their first jobs</strong>. The bottleneck consists of companies who see them as a cost to externalize, not an investment in their—the <em>company’s</em>—future.</p>

 <p>After their first job, an engineer can usually find work. But getting that first job, from what I can see, is<em> murder</em>. It is all but impossible—if you didn’t graduate from a top college, and you aren’t entering the feeder system of Big Tech, then it’s a roll of the dice, a question of luck or who has the best connections. It was rough <em>before</em> the chimera of “Generative AI can replace junior engineers” rose up from the swamp. And now…oof.</p>

 <p>Where would <em>you</em> be, if you hadn’t gotten into tech when you did?</p>

 <p>I know where I would be, and it is <em>not here</em>.</p>

 <p>The internet loves to make fun of Boomers, the generation that famously coasted to college, home ownership, and retirement, then pulled the ladder up after them while mocking younger people as snowflakes. “Ok, Boomer” may be here to stay, but can we try to keep “Ok, Staff Engineer” from becoming a thing?</p><p>Lots of people seem to think we don’t need junior engineers, but nobody is arguing that we need fewer senior engineers, or will need fewer senior engineers in the foreseeable future.</p>

 <p>I think it’s safe to assume that anything deterministic and automatable will eventually be automated. Software engineering is no different—we are ground zero! Of course we’re always looking for ways to automate and improve efficiency, as we should be.</p>

 <p>But large software systems are unpredictable and nondeterministic, with emergent behaviors. The mere existence of users injects chaos into the system. Components can be automated, but complexity can only be managed.</p>

 <p>Even if systems could be fully automated and managed by AI, the fact that we cannot understand how AI makes decisions is a huge, possibly insurmountable problem. Running your business on a system that humans can’t debug or understand seems like a risk so existential that no security, legal or finance team would ever sign off on it. Maybe some version of this future will come to pass, but it’s hard to see it from here. I would not bet my career or my company on it happening.</p>

 <p>In the meantime, we still need more senior engineers. The only way to grow them is by fixing the funnel.</p><p>No. You need to be able to set them up for success. Some factors that disqualify you from hiring junior engineers:</p>

 <ul><li>You have less than two years of runway</li><li>Your team is constantly in firefighting mode, or you have no slack in your system</li><li>You have no experienced managers, or you have bad managers, or no managers at all</li><li>You have no product roadmap</li><li>Nobody on your team has any interest in being their mentor or point person</li></ul>

 <p>The only thing worse than never hiring any junior engineers is hiring them into an awful experience where they can’t learn anything. (I wouldn’t set the bar quite as high as Cindy does <a href="https://copyconstruct.medium.com/tactical-challenges-in-hiring-junior-engineers-29e31634a9bd">in this article</a>; while I understand where she’s coming from, it is <em>so much easier</em> to land your second job than your first job that I think most junior engineers would frankly choose a crappy first job over none at all.)</p>

 <p>Being a fully distributed company isn’t a complete dealbreaker, but it does make things even harder. I would counsel junior engineers to seek out office jobs if at all possible. You learn so much faster when you can soak up casual conversations and technical chatter, and you lose that working from home. If you are a remote employer, know that you will need to work harder to compensate for this. I suggest connecting with others who have done this successfully (they exist!) for advice.</p>

 <p>I also advise companies not to start by hiring a single junior engineer. If you’re going to hire one, hire two or three. Give them a cohort of peers, so it’s a little less intimidating and isolating.</p><p>I have come to believe that the only way this will ever change is if engineers and engineering managers across our industry take up this fight and make it personal.</p>

 <p>Most of the places I know that do have a program for hiring and training entry level engineers, have it only because an engineer decided to fight for it. Engineers—sometimes engineering managers—were the ones who made the case and pushed for resources, then designed the program, interviewed and hired the junior engineers, and set them up with mentors. This is not an exotic project, it is well within the capabilities of most motivated, experienced engineers (and good for <em>your</em> career as well).</p>

 <p>Finance isn’t going to lobby for this. Execs aren’t likely to step in. The more a person’s role inclines them to treat engineers like fungible resources, the less likely they are to understand why this matters.</p>

 <p>AI is not coming to solve all our problems and write all our code for us—and <em>even if it was</em>, <em>it wouldn’t matter</em>. Writing code is but a sliver of what professional software engineers do, and arguably the easiest part. Only we have the context and the credibility to drive the changes we <em>know </em>form the bedrock for great teams and engineering excellence..</p>

 <p>Great teams are how great engineers get made. Nobody knows this better than engineers and EMs. It’s time for us to make the case, and make it happen.</p></div>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/93c287363dbe9cae22f4360c5b107971/index.md
+++ b/cache/2024/93c287363dbe9cae22f4360c5b107971/index.md
@@ -0,0 +1,146 @@
 title: Generative AI Is Not Going To Build Your Engineering Team For You
 url: https://stackoverflow.blog/2024/06/10/generative-ai-is-not-going-to-build-your-engineering-team-for-you/
 hash_url: 93c287363dbe9cae22f4360c5b107971
 archive_date: 2024-06-24
 og_image: https://cdn.stackoverflow.co/images/jo7n4k8s/production/dab8e8438e824245858d2442f2fa4569fb34e027-12000x6300.jpg?w=1200&fm=png&auto=format
 description: 
 favicon: https://stackoverflow.blog/favicon.ico
 language: en_US

 <div itemprop="articleBody" class="s-prose fs-subheading"><p>When I was 19 years old, I dropped out of college and moved to San Francisco. I had a job offer in hand to be a Unix sysadmin for Taos Consulting. However, before my first day of work I was lured away to a startup in the city, where I worked as a software engineer on mail subsystems.</p>

    <p>I never questioned whether or not I could find work. Jobs were plentiful, and more importantly, hiring standards were very low. If you knew how to sling HTML or find your way around a command line, chances were you could find <em>someone</em> to pay you.</p>

    <p>Was I some kind of genius, born with my hands on a computer keyboard? Assuredly not. I was homeschooled in the backwoods of Idaho. I didn’t touch a computer until I was sixteen and in college. I escaped to university on a classical performance piano scholarship, which I later traded in for a peripatetic series of nontechnical majors: classical Latin and Greek, musical theory, philosophy. Everything I knew about computers I learned on the job, doing sysadmin work for the university and CS departments.</p>

 <p>In retrospect, I was so lucky to enter the industry when I did. It makes me blanch to think of what would have happened if I had come along a few years later. Every one of the ladders my friends and I took into the industry has long since vanished.</p><p>To some extent, this is just what happens as an industry matures. The early days of any field are something of a Wild West, where the stakes are low, regulation nonexistent, and standards nascent. If you look at the early history of other industries—medicine, cinema, radio—the similarities are striking.</p>

 <p>There is a magical moment with any young technology where the boundaries between roles are porous and opportunity can be seized by anyone who is motivated, curious, and willing to work their asses off.</p>

 <p>It never lasts. It can’t; it shouldn’t. The amount of prerequisite knowledge and experience you must have before you can enter the industry swells precipitously. The stakes rise, the magnitude of the mission increases, the cost of mistakes soars. We develop certifications, trainings, standards, legal rites. We wrangle over whether or not <a href="https://www.hillelwayne.com/post/are-we-really-engineers/">software engineers are really engineers</a>.</p><p>Nowadays, you wouldn’t want a teenaged dropout like me to roll out of junior year and onto your pager rotation. The prerequisite knowledge you need to enter the industry has grown, the pace is faster, and the stakes are much higher, so you can no longer learn literally <em>everything</em> on the job, as I once did.</p>

 <p>However, it’s not like you can learn everything you need to know at college either. A CS degree typically prepares you better for a life of computing research than life as a workaday software engineer. A more practical path into the industry may be a good coding bootcamp, with its emphasis on problem solving and learning a modern toolkit. In either case, you don’t so much learn “how to do the job” as you do “learn enough of the basics to understand and use the tools you need to use to learn the job.”</p>

 <p>Software is an apprenticeship industry. You can’t learn to be a software engineer by reading books. You can only learn by doing…and doing, and doing, and doing some more. No matter what your education consists of, most learning happens on the job—period. And it never ends! Learning and teaching are lifelong practices; they have to be, the industry changes so fast.</p>

 <p>It takes a solid seven-plus years to forge a competent software engineer. (Or as most job ladders would call it, a “senior software engineer”.) That’s many years of writing, reviewing, and deploying code every day, on a team alongside more experienced engineers. That’s just how long it seems to take.</p><p>Here is where I often get some very indignant pushback to my timelines, e.g.:</p>

 <p>“Seven years?! Pfft, it took me two years!”</p><p>“I was promoted to Senior Software Engineer in less than five years!”</p>

 <p>Good for you. True, there is nothing magic about seven years. But it takes time and experience to mature into an experienced engineer, the kind who can anchor a team. More than that, it takes <em>practice</em>.</p>

 <p>I think we have come to use “Senior Software Engineer” as shorthand for engineers who can ship code and be a net positive in terms of productivity, and I think that’s a huge mistake. It implies that less senior engineers must be a net negative in terms of productivity, which is untrue. And it elides the real nature of the work of software engineering, of which writing code is only a small part.</p>

 <p>To me, being a senior engineer is not primarily a function of your ability to write code. It has far more to do with your ability to understand, maintain, explain, and manage a large body of software in production over time, as well as the ability to translate business needs into technical implementation. So much of the work is around crafting and curating these large, complex sociotechnical systems, and code is just one representation of these systems.</p>

 <p>What does it mean to be a senior engineer? It means you have learned how to <em>learn</em>, first and foremost, and how to teach; how to hold these models in your head and reason about them, and how to maintain, extend, and operate these systems over time. It means you have good judgment, and instincts you can trust.</p>

 <p>Which brings us to the matter of AI.</p><p>It is really, really tough to get your first role as an engineer. I didn’t realize how hard it was until I watched my little sister (new grad, terrific grades, some hands on experience, fiendishly hard worker) struggle for nearly <em>two years</em> to land a real job in her field. That was a few years ago; anecdotally, it seems to have gotten even harder since then.</p>

 <p>This past year, I have read a steady drip of articles about <a href="https://www.nytimes.com/2024/04/10/business/investment-banking-jobs-artificial-intelligence.html">entry-level jobs in various industries</a> being replaced by AI. Some of which absolutely have merit. Any job that consists of drudgery such as converting a document from one format to another, reading and summarizing a bunch of text, or replacing one set of icons with another, seems pretty obviously vulnerable. This doesn’t feel all that revolutionary to me, it’s just extending the existing boom in automation to cover textual material as well as mathy stuff.</p>

 <p>Recently, however, a number of execs and so-called “thought leaders” in tech seem to have genuinely convinced themselves that generative AI is on the verge of replacing all the work done by junior engineers. I have read so many articles about how junior engineering work is being automated out of existence, or that the need for junior engineers is shriveling up. It has officially driven me bonkers.</p>

 <p>All of this bespeaks a deep misunderstanding about what engineers actually <em>do</em>. By not hiring and training up junior engineers, we are cannibalizing our own future. We need to <em>stop doing that</em>.</p><p>People act like writing code is the hard part of software. It is not. It never has been, it never will be. <strong>Writing code is the easiest part of software engineering</strong>, and it’s getting easier by the day. The hard parts are what you do with that code—operating it, understanding it, extending it, and governing it over its entire lifecycle.</p>

 <p>A junior engineer begins by learning how to write and debug lines, functions, and snippets of code. As you practice and progress towards being a senior engineer, you learn to compose systems out of software, and guide systems through waves of change and transformation.</p>

 <p>Sociotechnical systems consist of software, tools, and people; understanding them requires familiarity with the interplay between software, users, production, infrastructure, and continuous changes over time. These systems are fantastically complex and subject to chaos, nondeterminism and emergent behaviors. If anyone claims to understand the system they are developing and operating, the system is either exceptionally small or (more likely) they don’t know enough to know what they don’t know. Code is easy, in other words, but <em>systems are hard</em>.</p>

 <p>The present wave of generative AI tools has done a lot to help us generate lots of code, very fast. The easy parts are becoming even easier, at a truly remarkable pace. But it has not done a thing to aid in the work of managing, understanding, or operating that code. If anything, it has only made the hard jobs harder.</p><p>But that code first has to be understood, adapted and integrated cleanly into the system.</p><p>If you read a lot of breathless think pieces, you may have a mental image of software engineers merrily crafting prompts for ChatGPT, or using Copilot to generate reams of code, then committing whatever emerges to GitHub and walking away. That does not resemble our reality.</p>

 <p>The right way to think about tools like Copilot is more like a really fancy autocomplete or copy-paste function, or maybe like the unholy love child of Stack Overflow search results plus Google’s “I feel lucky”. You roll the dice, every time.</p>

 <p>These tools are at their best when there’s already a parallel in the file, and you want to just copy-paste the thing with slight modifications. Or when you’re writing tests and you have a giant block of fairly repetitive YAML, and it repeats the pattern while inserting the right column and field names, like an automatic template.</p>

 <p>However, <strong>you cannot trust generated code</strong>. I can’t emphasize this enough. AI-generated code always looks quite plausible, but even when it kind of “works”, it’s rarely congruent with your wants and needs. It will happily generate code that doesn’t parse or compile. It will make up variables, method names, function calls; it will hallucinate fields that don’t exist. Generated code will not follow your coding practices or conventions. It is not going to refactor or come up with intelligent abstractions for you. The more important, difficult or meaningful a piece of code is, the less likely you are to generate a usable artifact using AI.</p>

 <p>You may save time by not having to type the code in from scratch, but you will need to step through the output line by line, revising as you go, before you can commit your code, let alone ship it to production. In many cases this will take as much or more time as it would take to simply write the code—especially these days, now that autocomplete has gotten so clever and sophisticated. It can be a LOT of work to bring AI-generated code into compliance and coherence with the rest of your codebase. It isn’t always worth the effort, quite frankly.</p>

 <p>Generating code that can compile, execute, and pass a test suite isn’t especially hard; the hard part is crafting a code base that many people, teams, and successive generations of teams can navigate, mutate, and reason about for years to come.</p><p>So that’s the TLDR: you can generate a lot of code, really fast, but you can’t trust what comes out. At all. However, there are some use cases where generative AI consistently shines.</p>

 <p>For example, it’s often easier to ask chatGPT to generate example code using unfamiliar APIs than by reading the API docs—the corpus was trained on repositories where the APIs are being used for real life workloads, after all.</p>

 <p>Generative AI is also pretty good at producing code that is annoying or tedious to write, yet tightly scoped and easy to explain. The more predictable a scenario is, the better these tools are at writing the code for you. If what you need is effectively copy-paste with a template—any time you could generate the code you want using sed/awk or vi macros—generative AI is quite good at this.</p>

 <p>It’s also very good at writing little functions for you to do things in unfamiliar languages or scenarios. If you have a snippet of Python code and you want the same thing in Java, but you don’t know Java, generative AI has got your back.</p>

 <p>Again, remember, the odds are 50/50 that the result is completely made up. You always have to assume the results are incorrect until you can verify it by hand. But these tools can absolutely accelerate your work in countless ways.</p><p>One of the engineers I work with, Kent Quirk, describes generative AI as “an excitable junior engineer who types really fast”. I love that quote—it leaves an indelible mental image.</p>

 <p>Generative AI is like a junior engineer in that you can’t roll their code off into production. You are responsible for it—legally, ethically, and practically. You still have to take the time to understand it, test it, instrument it, retrofit it stylistically and thematically to fit the rest of your code base, and ensure your teammates can understand and maintain it as well.</p>

 <p>The analogy is a decent one, actually, but <em>only</em> if your code is disposable and self-contained, i.e. not meant to be integrated into a larger body of work, or to survive and be read or modified by others.</p>

 <p>And hey—there are corners of the industry like this, where most of the code is write-only, throwaway code. There are agencies that spin out dozens of disposable apps per year, each written for a particular launch or marketing event and then left to wither on the vine. <em>But that is not most software.</em> Disposable code is rare; code that needs to work over the long term is the norm. Even when we think a piece of code will be disposable, we are often (urf) wrong.</p><p>In that particular sense—generating code that you know is untrustworthy—GenAI is a bit like a junior engineer. But in every other way, the analogy fails. Because adding a person who writes code to your team is nothing like autogenerating code. That code could have come from anywhere—Stack Overflow, Copilot, whatever. You don’t know, and it doesn’t really matter. There’s no feedback loop, no person on the other end trying iteratively to learn and improve, and no impact to your team vibes or culture.</p>

 <p>To state the supremely obvious: giving code review feedback to a junior engineer is not like editing generated code. Your effort is worth more when it is invested into someone else’s apprenticeship. It’s an opportunity to pass on the lessons you’ve learned in your own career. Even just the act of framing your feedback to explain and convey your message forces you to think through the problem in a more rigorous way, and has a way of helping you understand the material more deeply.</p>

 <p>And adding a junior engineer to your team will immediately change team dynamics. It creates an environment where asking questions is normalized and encouraged, where teaching as well as learning is a constant. We’ll talk more about team dynamics in a moment.</p>

 <p>The time you invest into helping a junior engineer level up can pay off remarkably quickly. Time flies. ☺️ When it comes to hiring, we tend to valorize senior engineers almost as much as we underestimate junior engineers. Neither stereotype is helpful.</p><p>People seem to think that once you hire a senior engineer, you can drop them onto a team and they will be immediately productive, while hiring a junior engineer will be a tax on team performance forever. Neither are true. Honestly, <em>most</em> of the work that <em>most</em> teams have to do is not that difficult, once it’s been broken down into its constituent parts. There’s plenty of room for lower level engineers to execute and flourish.</p>

 <p>The grossly simplified perspective of your accountant goes something like this. “Why should we pay $100k for a junior engineer to slow things down, when we could pay $200k for a senior engineer to speed things up?” It makes no sense!</p>

 <p>But you know and I know—every engineer who is paying attention should know—<strong>that’s not how engineering works</strong>. This is an apprenticeship industry, and productivity is defined by the output and carrying capacity of each team, not each person.</p>

 <p>There are lots of ways a person can contribute to the overall velocity of a team, just like there are lots of ways a person can sap the energy out of a team or add friction and drag to everyone around them. These do not always correlate with the person’s level (at least not in the direction people tend to assume), and writing code is only one way.</p>

 <p>Furthermore, every engineer you hire requires ramp time and investment before they can contribute. Hiring and training new engineers is a costly endeavor, no matter what level they are. It will take any senior engineer time to build up their mental model of the system, familiarize themselves with the tools and technology, and ramp up to speed. How long? It depends on how clean and organized the codebase is, past experience with your tools and technologies, how good you are at onboarding new engineers, and more, but likely around 6-9 months. They probably won’t reach cruising altitude for about a year.</p>

 <p>Yes, the ramp will be longer for a junior engineer, and yes, it will require more investment from the team. But it’s not indefinite. Your junior engineer should be a net positive within roughly the same time frame, six months to a year, and they develop far more rapidly than more senior contributors. (Don’t forget, their contributions may vastly exceed the code they personally write.)</p><p>In terms of writing and shipping features, some of the most productive engineers I’ve ever known have been intermediate engineers. Not yet bogged down with all the meetings and curating and mentoring and advising and architecture, their calendars not yet pockmarked with interruptions, they can <strong>just build stuff</strong>. You see them put their headphones on first thing in the morning, write code all day, and cruise out the door in the evening having made incredible progress.</p>

 <p>Intermediate engineers sit in this lovely, temporary state where they have gotten good enough at programming to be very productive, but they are still learning how to build and care for systems. All they do is write code, reams and reams of code.</p>

 <p>And they’re energized…engaged. They’re having fun! They aren’t bored with writing a web form or a login page for the 1000th time. Everything is new, interesting, and exciting, which typically means they will <em>do a better job</em>, especially under the light direction of someone more experienced. Having intermediate engineers on a team is amazing. The only way you get them is by hiring junior engineers.</p>

 <p>Having junior and intermediate engineers on a team is a shockingly good inoculation against overengineering and premature complexity. They don’t yet know enough about a problem to imagine all the infinite edge cases that need to be solved for. They help keep things simple, which is one of the hardest things to do.</p><p>If you ask, nearly everybody will wholeheartedly agree that hiring junior engineers is a good thing…and someone else should do it. This is because the long-term arguments for hiring junior engineers are compelling and fairly well understood.</p>

 <ol><li>We need more senior engineers as an industry</li><li><em>Somebody</em> has to train them</li><li>Junior engineers are cheaper</li><li>They may add some much-needed diversity</li><li>They are often very loyal to companies who invest in training them, and will stick around for years instead of job hopping</li><li>Did we already mention that somebody needs to do it?</li></ol>

 <p>But long-term thinking is not a thing that companies, or capitalism in general, are typically great at. Framed this way, it makes it sound like you hire junior engineers as a selfless act of public service, at great cost to yourself. Companies are much more likely to want to externalize costs like those, which is how we got to where we are now.</p><p>However, there are at least as many arguments to be made for hiring junior engineers in the short term—selfish, hard-nosed, profitable reasons for why it benefits the team and the company to do so. You just have to shift your perspective slightly, from individuals to teams, to bring them into focus.</p>

 <p>Let’s start here: hiring engineers is not a process of “picking the best person for the job”. <strong>Hiring engineers is about composing teams</strong>. The smallest unit of software ownership is not the individual, it’s the team. Only teams can own, build, and maintain a corpus of software. It is inherently a collaborative, cooperative activity.</p>

 <p>If hiring engineers was about picking the “best people”, it would make sense to hire the most senior, experienced individual you can get for the money you have, because we are using “senior” and “experienced” as a proxy for “productivity”. (Questionable, but let’s not nitpick.) But the productivity of each individual is not what we should be optimizing for. The productivity of the <em>team</em> is all that matters.</p>

 <p>And the best teams are <em>always</em> the ones with a diversity of strengths, perspectives, and levels of expertise. A monoculture can be spectacularly successful in the short term—it may even outperform a diverse team. But they do not scale well, and they do not adapt to unfamiliar challenges gracefully. The longer you wait to diversify, the harder it will be.</p>

 <p>We need to hire junior engineers, and not just once, but consistently. We need to keep feeding the funnel from the bottom up. Junior engineers only stay junior for a couple years, and intermediate engineers turn into senior engineers. Super-senior engineers are not actually the best people to mentor junior engineers; the most effective mentor is usually someone just one level ahead, who vividly remembers what it was like in your shoes.</p><p>A healthy team is an ecosystem. You wouldn’t staff a product engineering team with six DB experts and one mobile developer. Nor should you staff it with six staff+ engineers and one junior developer. A good team is composed of a range of skills and levels.</p>

 <p>Have you ever been on a team packed exclusively with staff or principal engineers? It is <em>not fun</em>. That is not a high-functioning team. There is only so much high-level architecture and planning work to go around, there are only so many big decisions that need to be made. These engineers spend most of their time doing work that feels boring and repetitive, so they tend to over-engineer solutions and/or cut corners—sometimes at the same time. They compete for the “fun” stuff and find reasons to pick technical fights with each other. They chronically under-document and under-invest in the work that makes systems simple and tractable.</p>

 <p>Teams that only have intermediate engineers (or beginners, or seniors, or whatever) will have different pathologies, but similar problems with contention and blind spots. The work itself has a wide range in complexity and difficulty—from simple, tightly scoped functions to tough, high-stakes architecture decisions. It makes sense for the people doing the work to occupy a similar range.</p>

 <p>The best teams are ones where no one is bored, because every single person is working on something that challenges them and pushes their boundaries. The only way you can get this is by having a range of skill levels on the team.</p><p>The bottleneck we face now is not our ability to train up new junior engineers and give them skills. Nor is it about juniors learning to hustle harder; I see a lot of <a href="https://chioualexander.medium.com/how-to-effectively-apply-to-jobs-as-a-junior-engineer-37a90312d4fd">solid</a>, <a href="https://blog.pragmaticengineer.com/advice-for-junior-software-engineers/">well-meaning advice</a> on this topic, but it’s not going to solve the problem. <strong>The bottleneck is giving them their first jobs</strong>. The bottleneck consists of companies who see them as a cost to externalize, not an investment in their—the <em>company’s</em>—future.</p>

 <p>After their first job, an engineer can usually find work. But getting that first job, from what I can see, is<em> murder</em>. It is all but impossible—if you didn’t graduate from a top college, and you aren’t entering the feeder system of Big Tech, then it’s a roll of the dice, a question of luck or who has the best connections. It was rough <em>before</em> the chimera of “Generative AI can replace junior engineers” rose up from the swamp. And now…oof.</p>

 <p>Where would <em>you</em> be, if you hadn’t gotten into tech when you did?</p>

 <p>I know where I would be, and it is <em>not here</em>.</p>

 <p>The internet loves to make fun of Boomers, the generation that famously coasted to college, home ownership, and retirement, then pulled the ladder up after them while mocking younger people as snowflakes. “Ok, Boomer” may be here to stay, but can we try to keep “Ok, Staff Engineer” from becoming a thing?</p><p>Lots of people seem to think we don’t need junior engineers, but nobody is arguing that we need fewer senior engineers, or will need fewer senior engineers in the foreseeable future.</p>

 <p>I think it’s safe to assume that anything deterministic and automatable will eventually be automated. Software engineering is no different—we are ground zero! Of course we’re always looking for ways to automate and improve efficiency, as we should be.</p>

 <p>But large software systems are unpredictable and nondeterministic, with emergent behaviors. The mere existence of users injects chaos into the system. Components can be automated, but complexity can only be managed.</p>

 <p>Even if systems could be fully automated and managed by AI, the fact that we cannot understand how AI makes decisions is a huge, possibly insurmountable problem. Running your business on a system that humans can’t debug or understand seems like a risk so existential that no security, legal or finance team would ever sign off on it. Maybe some version of this future will come to pass, but it’s hard to see it from here. I would not bet my career or my company on it happening.</p>

 <p>In the meantime, we still need more senior engineers. The only way to grow them is by fixing the funnel.</p><p>No. You need to be able to set them up for success. Some factors that disqualify you from hiring junior engineers:</p>

 <ul><li>You have less than two years of runway</li><li>Your team is constantly in firefighting mode, or you have no slack in your system</li><li>You have no experienced managers, or you have bad managers, or no managers at all</li><li>You have no product roadmap</li><li>Nobody on your team has any interest in being their mentor or point person</li></ul>

 <p>The only thing worse than never hiring any junior engineers is hiring them into an awful experience where they can’t learn anything. (I wouldn’t set the bar quite as high as Cindy does <a href="https://copyconstruct.medium.com/tactical-challenges-in-hiring-junior-engineers-29e31634a9bd">in this article</a>; while I understand where she’s coming from, it is <em>so much easier</em> to land your second job than your first job that I think most junior engineers would frankly choose a crappy first job over none at all.)</p>

 <p>Being a fully distributed company isn’t a complete dealbreaker, but it does make things even harder. I would counsel junior engineers to seek out office jobs if at all possible. You learn so much faster when you can soak up casual conversations and technical chatter, and you lose that working from home. If you are a remote employer, know that you will need to work harder to compensate for this. I suggest connecting with others who have done this successfully (they exist!) for advice.</p>

 <p>I also advise companies not to start by hiring a single junior engineer. If you’re going to hire one, hire two or three. Give them a cohort of peers, so it’s a little less intimidating and isolating.</p><p>I have come to believe that the only way this will ever change is if engineers and engineering managers across our industry take up this fight and make it personal.</p>

 <p>Most of the places I know that do have a program for hiring and training entry level engineers, have it only because an engineer decided to fight for it. Engineers—sometimes engineering managers—were the ones who made the case and pushed for resources, then designed the program, interviewed and hired the junior engineers, and set them up with mentors. This is not an exotic project, it is well within the capabilities of most motivated, experienced engineers (and good for <em>your</em> career as well).</p>

 <p>Finance isn’t going to lobby for this. Execs aren’t likely to step in. The more a person’s role inclines them to treat engineers like fungible resources, the less likely they are to understand why this matters.</p>

 <p>AI is not coming to solve all our problems and write all our code for us—and <em>even if it was</em>, <em>it wouldn’t matter</em>. Writing code is but a sliver of what professional software engineers do, and arguably the easiest part. Only we have the context and the credibility to drive the changes we <em>know </em>form the bedrock for great teams and engineering excellence..</p>

 <p>Great teams are how great engineers get made. Nobody knows this better than engineers and EMs. It’s time for us to make the case, and make it happen.</p></div>
--- a/cache/2024/b39aef611801e2b15f189e8d02586b88/index.html
+++ b/cache/2024/b39aef611801e2b15f189e8d02586b88/index.html
@@ -0,0 +1,234 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="en">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>Notes From “You Are Not A Gadget” (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://blog.jim-nielsen.com/2024/notes-from-you-are-not-a-gadget/">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>Notes From “You Are Not A Gadget”</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://blog.jim-nielsen.com/2024/notes-from-you-are-not-a-gadget/" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <p>Jaron Lanier’s book <em>You Are Not a Gadget</em> was written in 2010, but its preface is a prescient banger for 2024, the year of our AI overlord:</p>
 <blockquote>
 <p>It's early in the 21st century, and that means that these words will mostly be read by nonpersons...[they] will be minced...within industrial cloud computing facilities...They will be scanned, rehashed, and misrepresented...Ultimately these words will contribute to the fortunes of those few who have been able to position themselves as lords of the computing clouds.</p>
 </blockquote>
 <p>Today he might call the book, “You Are Not an Input to Artificial Intelligence”.</p>
 <p>Lanier concludes the preface to his book by saying the words in it are intended for people, not computers.</p>
 <p>Same for my blog! The words in it are meant for people, not computers. And I would hope any computerized representation of these words is solely for facilitating humans finding them and reading them in context.</p>
 <p>Anyhow, here’s a few of my notes from the book. </p>
 <h2 id="so-long-to-the-individual-point-of-view">So Long to The Individual Point of View</h2>
 <blockquote>
 <p>Authorship—the very idea of the individual point of view—is not a priority of the new technology...Instead of people being treated as the sources of their own creativity, commercial aggregation and abstraction sites present anonymized fragments of creativity…obscuring the true sources.</p>
 </blockquote>
 <p>Again, this was 2010, way before “AI”.</p>
 <p>Who cares for sources anymore? The perspective of the individual is obsolete. Everyone is flattened into a global mush. A word smoothie. We care more for the abstractions we can create on top of individual expression rather than the individuals and their expressions.</p>
 <blockquote>
 <p>The central mistake of recent digital culture is to chop up a network of individuals so finely that you end up with a mush. You then start to care about the abstraction of the network more than the real people who are networked, even though the network by itself is meaningless. Only people were ever meaningful</p>
 </blockquote>
 <p>While Lanier was talking about “the hive mind” of social networks as we understood it then, AI has a similar problem: we begin to care more about the training data than the individual humans whose outputs constitute the training data, even though the training data by itself is meaningless. Only people are meaningful.<sup id="fnref:1"><a href="#fn:1">[1]</a></sup> As Lanier says in the book:</p>
 <blockquote>
 <p>The bits don't mean anything without a cultured person to interpret them.</p>
 </blockquote>
 <blockquote>
 <p>Information is alienated experience.</p>
 </blockquote>
 <h2 id="emphasizing-artificial-or-natural-intelligence">Emphasizing Artificial or Natural Intelligence</h2>
 <blockquote>
 <p>Emphasizing the crowd means deemphasizing individual humans.</p>
 </blockquote>
 <p>I like that.</p>
 <p>Here’s a corollary: emphasizing <em>artificial</em> intelligence means de-emphasizing <em>natural</em> intelligence. </p>
 <p>Therein lies the tradeoff. </p>
 <p>In Web 2.0, we emphasized the crowd over the individual and people behaved like a crowd instead of individuals, like a mob rather than a person. The design encouraged, even solicited, that kind of behavior. </p>
 <p>Now with <em>artificial</em> intelligence enshrined, is it possible we begin to act like it? Hallucinating reality and making baseless claims in complete confidence will be normal, as that’s what the robots we interact with all day do.</p>
 <blockquote>
 <p>What is communicated between people eventually becomes their truth. Relationships take on the troubles of software engineering.</p>
 </blockquote>
 <h2 id="what-even-is-intelligence">What Even is “Intelligence”?</h2>
 <blockquote>
 <p>Before MIDI, a musical note was a bottomless idea that transcended absolute definition</p>
 </blockquote>
 <p>But the digitalization of music require removing options and possibilities based on what was easiest to be represented and processed by the computer. We remove “the unfathomable penumbra of meaning that distinguishes” a musical note in the flesh to make a musical note in the computer.</p>
 <p>Why? Because computers require abstractions. But abstractions are just that: models that roughly fit the real thing. But too often we let the abstractions become our reality:</p>
 <blockquote>
 <p>Each layer of digital abstraction, no matter how well it is crafted, contributes some degree of error and obfuscation. No abstraction corresponds to reality perfectly. A lot of such layers become a system unto themselves, one that functions apart from the reality that is obscured far below.</p>
 </blockquote>
 <p>Lanier argues it happened with MIDI and it happened with social networks, where people became rows in a database and began living up to that abstraction. </p>
 <blockquote>
 <p>people are becoming like MIDI notes—overly defined, and restricted in practice to what can be represented in a computer...We have narrowed what we expect from the most commonplace forms of musical sound in order to make the technology adequate.</p>
 </blockquote>
 <p>Perhaps similarly, intelligence (dare I say consciousness) was a bottomless idea that transcended definition. But we soon narrowed it down to fit our abstractions in the computer.</p>
 <blockquote>
 <p>We are happy to enshrine into engineering designs mere hypotheses—and vague ones at that—about the hardest and most profound questions faced by science, as if we already posses perfect knowledge.</p>
 </blockquote>
 <p>So we enshrine the idea of intelligence into our computing paradigm when we don’t even know what it means for ourselves. Are we making computers smarter or ourselves dumber?</p>
 <blockquote>
 <p>You can't tell if a machine has gotten smarter or if you've just lowered your own standards of intelligence to such a degree that the machine seems smart. </p>
 </blockquote>
 <p>Prescient.</p>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/b39aef611801e2b15f189e8d02586b88/index.md
+++ b/cache/2024/b39aef611801e2b15f189e8d02586b88/index.md
@@ -0,0 +1,67 @@
 title: Notes From “You Are Not A Gadget”
 url: https://blog.jim-nielsen.com/2024/notes-from-you-are-not-a-gadget/
 hash_url: b39aef611801e2b15f189e8d02586b88
 archive_date: 2024-06-24
 og_image: https://blog.jim-nielsen.com/assets/img/twitter-card.png
 description: Writing about the big beautiful mess that is making things for the world wide web.
 favicon: https://blog.jim-nielsen.com/favicon.ico
 language: en_US

 <p>Jaron Lanier’s book <em>You Are Not a Gadget</em> was written in 2010, but its preface is a prescient banger for 2024, the year of our AI overlord:</p>
 <blockquote>
 <p>It's early in the 21st century, and that means that these words will mostly be read by nonpersons...[they] will be minced...within industrial cloud computing facilities...They will be scanned, rehashed, and misrepresented...Ultimately these words will contribute to the fortunes of those few who have been able to position themselves as lords of the computing clouds.</p>
 </blockquote>
 <p>Today he might call the book, “You Are Not an Input to Artificial Intelligence”.</p>
 <p>Lanier concludes the preface to his book by saying the words in it are intended for people, not computers.</p>
 <p>Same for my blog! The words in it are meant for people, not computers. And I would hope any computerized representation of these words is solely for facilitating humans finding them and reading them in context.</p>
 <p>Anyhow, here’s a few of my notes from the book. </p>
 <h2 id="so-long-to-the-individual-point-of-view">So Long to The Individual Point of View</h2>
 <blockquote>
 <p>Authorship—the very idea of the individual point of view—is not a priority of the new technology...Instead of people being treated as the sources of their own creativity, commercial aggregation and abstraction sites present anonymized fragments of creativity…obscuring the true sources.</p>
 </blockquote>
 <p>Again, this was 2010, way before “AI”.</p>
 <p>Who cares for sources anymore? The perspective of the individual is obsolete. Everyone is flattened into a global mush. A word smoothie. We care more for the abstractions we can create on top of individual expression rather than the individuals and their expressions.</p>
 <blockquote>
 <p>The central mistake of recent digital culture is to chop up a network of individuals so finely that you end up with a mush. You then start to care about the abstraction of the network more than the real people who are networked, even though the network by itself is meaningless. Only people were ever meaningful</p>
 </blockquote>
 <p>While Lanier was talking about “the hive mind” of social networks as we understood it then, AI has a similar problem: we begin to care more about the training data than the individual humans whose outputs constitute the training data, even though the training data by itself is meaningless. Only people are meaningful.<sup id="fnref:1"><a href="#fn:1">[1]</a></sup> As Lanier says in the book:</p>
 <blockquote>
 <p>The bits don't mean anything without a cultured person to interpret them.</p>
 </blockquote>
 <blockquote>
 <p>Information is alienated experience.</p>
 </blockquote>
 <h2 id="emphasizing-artificial-or-natural-intelligence">Emphasizing Artificial or Natural Intelligence</h2>
 <blockquote>
 <p>Emphasizing the crowd means deemphasizing individual humans.</p>
 </blockquote>
 <p>I like that.</p>
 <p>Here’s a corollary: emphasizing <em>artificial</em> intelligence means de-emphasizing <em>natural</em> intelligence. </p>
 <p>Therein lies the tradeoff. </p>
 <p>In Web 2.0, we emphasized the crowd over the individual and people behaved like a crowd instead of individuals, like a mob rather than a person. The design encouraged, even solicited, that kind of behavior. </p>
 <p>Now with <em>artificial</em> intelligence enshrined, is it possible we begin to act like it? Hallucinating reality and making baseless claims in complete confidence will be normal, as that’s what the robots we interact with all day do.</p>
 <blockquote>
 <p>What is communicated between people eventually becomes their truth. Relationships take on the troubles of software engineering.</p>
 </blockquote>
 <h2 id="what-even-is-intelligence">What Even is “Intelligence”?</h2>
 <blockquote>
 <p>Before MIDI, a musical note was a bottomless idea that transcended absolute definition</p>
 </blockquote>
 <p>But the digitalization of music require removing options and possibilities based on what was easiest to be represented and processed by the computer. We remove “the unfathomable penumbra of meaning that distinguishes” a musical note in the flesh to make a musical note in the computer.</p>
 <p>Why? Because computers require abstractions. But abstractions are just that: models that roughly fit the real thing. But too often we let the abstractions become our reality:</p>
 <blockquote>
 <p>Each layer of digital abstraction, no matter how well it is crafted, contributes some degree of error and obfuscation. No abstraction corresponds to reality perfectly. A lot of such layers become a system unto themselves, one that functions apart from the reality that is obscured far below.</p>
 </blockquote>
 <p>Lanier argues it happened with MIDI and it happened with social networks, where people became rows in a database and began living up to that abstraction. </p>
 <blockquote>
 <p>people are becoming like MIDI notes—overly defined, and restricted in practice to what can be represented in a computer...We have narrowed what we expect from the most commonplace forms of musical sound in order to make the technology adequate.</p>
 </blockquote>
 <p>Perhaps similarly, intelligence (dare I say consciousness) was a bottomless idea that transcended definition. But we soon narrowed it down to fit our abstractions in the computer.</p>
 <blockquote>
 <p>We are happy to enshrine into engineering designs mere hypotheses—and vague ones at that—about the hardest and most profound questions faced by science, as if we already posses perfect knowledge.</p>
 </blockquote>
 <p>So we enshrine the idea of intelligence into our computing paradigm when we don’t even know what it means for ourselves. Are we making computers smarter or ourselves dumber?</p>
 <blockquote>
 <p>You can't tell if a machine has gotten smarter or if you've just lowered your own standards of intelligence to such a degree that the machine seems smart. </p>
 </blockquote>
 <p>Prescient.</p>
--- a/cache/2024/d4c01d246dc4f43aeb95c427ac5422a5/index.html
+++ b/cache/2024/d4c01d246dc4f43aeb95c427ac5422a5/index.html
@@ -0,0 +1,197 @@
 <!doctype html><!-- This is a valid HTML5 document. -->
 <!-- Screen readers, SEO, extensions and so on. -->
 <html lang="en">
 <!-- Has to be within the first 1024 bytes, hence before the `title` element
  See: https://www.w3.org/TR/2012/CR-html5-20121217/document-metadata.html#charset -->
 <meta charset="utf-8">
 <!-- Why no `X-UA-Compatible` meta: https://stackoverflow.com/a/6771584 -->
 <!-- The viewport meta is quite crowded and we are responsible for that.
  See: https://codepen.io/tigt/post/meta-viewport-for-2015 -->
 <meta name="viewport" content="width=device-width,initial-scale=1">
 <!-- Required to make a valid HTML5 document. -->
 <title>Adactio: Journal-Trust (archive) — David Larlet</title>
 <meta name="description" content="Publication mise en cache pour en conserver une trace.">
 <!-- That good ol' feed, subscribe :). -->
 <link rel="alternate" type="application/atom+xml" title="Feed" href="/david/log/">
 <!-- Generated from https://realfavicongenerator.net/ such a mess. -->
 <link rel="apple-touch-icon" sizes="180x180" href="/static/david/icons2/apple-touch-icon.png">
 <link rel="icon" type="image/png" sizes="32x32" href="/static/david/icons2/favicon-32x32.png">
 <link rel="icon" type="image/png" sizes="16x16" href="/static/david/icons2/favicon-16x16.png">
 <link rel="manifest" href="/static/david/icons2/site.webmanifest">
 <link rel="mask-icon" href="/static/david/icons2/safari-pinned-tab.svg" color="#07486c">
 <link rel="shortcut icon" href="/static/david/icons2/favicon.ico">
 <meta name="msapplication-TileColor" content="#f7f7f7">
 <meta name="msapplication-config" content="/static/david/icons2/browserconfig.xml">
 <meta name="theme-color" content="#f7f7f7" media="(prefers-color-scheme: light)">
 <meta name="theme-color" content="#272727" media="(prefers-color-scheme: dark)">
 <!-- Is that even respected? Retrospectively? What a shAItshow…
  https://neil-clarke.com/block-the-bots-that-feed-ai-models-by-scraping-your-website/ -->
 <meta name="robots" content="noai, noimageai">
 <!-- Documented, feel free to shoot an email. -->
 <link rel="stylesheet" href="/static/david/css/style_2021-01-20.css">
 <!-- See https://www.zachleat.com/web/comprehensive-webfonts/ for the trade-off. -->
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t4_poly_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: light), (prefers-color-scheme: no-preference)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_regular.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_bold.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <link rel="preload" href="/static/david/css/fonts/triplicate_t3_italic.woff2" as="font" type="font/woff2" media="(prefers-color-scheme: dark)" crossorigin>
 <script>
    function toggleTheme(themeName) {
        document.documentElement.classList.toggle(
            'forced-dark',
            themeName === 'dark'
        )
        document.documentElement.classList.toggle(
            'forced-light',
            themeName === 'light'
        )
    }
    const selectedTheme = localStorage.getItem('theme')
    if (selectedTheme !== 'undefined') {
        toggleTheme(selectedTheme)
    }
 </script>

  <meta name="robots" content="noindex, nofollow">
  <meta content="origin-when-cross-origin" name="referrer">
  <!-- Canonical URL for SEO purposes -->
  <link rel="canonical" href="https://adactio.com/journal/21160">

 <body class="remarkdown h1-underline h2-underline h3-underline em-underscore hr-center ul-star pre-tick" data-instant-intensity="viewport-all">


 <article>
  <header>
    <h1>Adactio: Journal-Trust</h1>
  </header>
  <nav>
    <p class="center">
      <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
      <a href="https://adactio.com/journal/21160" title="Lien vers le contenu original">Source originale</a>
      <br>
      Mis en cache le 2024-06-24
    </p>
  </nav>
  <hr>
  <p>In their rush to cram in “AI” “features”, it seems to me that many companies don’t actually understand why people use their products.</p>

 <p>Google is acting as though its greatest asset is its search engine. Same with Bing.</p>

 <p>Mozilla Developer Network is acting as though its greatest asset is its documentation. Same with Stack Overflow.</p>

 <p>But their greatest asset is actually trust.</p>

 <p>If I use a search engine I need to be able to trust that the filtering is good. If I look up documentation I need to trust that the information is good. I don’t expect perfection, but I also don’t expect to have to constantly be thinking “was this generated by a large language model, and if so, how can I know it’s not hallucinating?”</p>

 <p>“But”, the apologists will respond, “the results are <em>mostly</em> correct! The documentation is <em>mostly</em> true!”</p>

 <p>Sure, but as <a href="https://shkspr.mobi/blog/2024/05/why-do-people-focus-on-failure/">Terence puts it</a>:</p>

 <blockquote>
  <p>The intern who files most things perfectly but has, more than once, tipped an entire cup of coffee into the filing cabinet is going to be remembered as “that klutzy intern we had to fire.”</p>
 </blockquote>

 <p>Trust is a precious commodity. It takes a long time to build trust. It takes a short time to destroy it.</p>

 <p>I am honestly astonished that so many companies don’t seem to realise what they’re destroying.</p>
 </article>


 <hr>

 <footer>
  <p>
    <a href="/david/" title="Aller à l’accueil"><svg class="icon icon-home">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-home"></use>
      </svg> Accueil</a> •
    <a href="/david/log/" title="Accès au flux RSS"><svg class="icon icon-rss2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-rss2"></use>
      </svg> Suivre</a> •
    <a href="http://larlet.com" title="Go to my English profile" data-instant><svg class="icon icon-user-tie">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-user-tie"></use>
      </svg> Pro</a> •
    <a href="mailto:david%40larlet.fr" title="Envoyer un courriel"><svg class="icon icon-mail">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-mail"></use>
      </svg> Email</a> •
    <abbr class="nowrap" title="Hébergeur : Alwaysdata, 62 rue Tiquetonne 75002 Paris, +33184162340"><svg class="icon icon-hammer2">
        <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-hammer2"></use>
      </svg> Légal</abbr>
  </p>
  <template id="theme-selector">
      <form>
          <fieldset>
              <legend><svg class="icon icon-brightness-contrast">
                <use xlink:href="/static/david/icons2/symbol-defs-2021-12.svg#icon-brightness-contrast"></use>
              </svg> Thème</legend>
              <label>
                  <input type="radio" value="auto" name="chosen-color-scheme" checked> Auto
              </label>
              <label>
                  <input type="radio" value="dark" name="chosen-color-scheme"> Foncé
              </label>
              <label>
                  <input type="radio" value="light" name="chosen-color-scheme"> Clair
              </label>
          </fieldset>
      </form>
  </template>
 </footer>
 <script src="/static/david/js/instantpage-5.1.0.min.js" type="module"></script>
 <script>
    function loadThemeForm(templateName) {
        const themeSelectorTemplate = document.querySelector(templateName)
        const form = themeSelectorTemplate.content.firstElementChild
        themeSelectorTemplate.replaceWith(form)

        form.addEventListener('change', (e) => {
            const chosenColorScheme = e.target.value
            localStorage.setItem('theme', chosenColorScheme)
            toggleTheme(chosenColorScheme)
        })

        const selectedTheme = localStorage.getItem('theme')
        if (selectedTheme && selectedTheme !== 'undefined') {
            form.querySelector(`[value="${selectedTheme}"]`).checked = true
        }
    }

    const prefersColorSchemeDark = '(prefers-color-scheme: dark)'
    window.addEventListener('load', () => {
        let hasDarkRules = false
        for (const styleSheet of Array.from(document.styleSheets)) {
            let mediaRules = []
            for (const cssRule of styleSheet.cssRules) {
                if (cssRule.type !== CSSRule.MEDIA_RULE) {
                    continue
                }
                // WARNING: Safari does not have/supports `conditionText`.
                if (cssRule.conditionText) {
                    if (cssRule.conditionText !== prefersColorSchemeDark) {
                        continue
                    }
                } else {
                    if (cssRule.cssText.startsWith(prefersColorSchemeDark)) {
                        continue
                    }
                }
                mediaRules = mediaRules.concat(Array.from(cssRule.cssRules))
            }

            // WARNING: do not try to insert a Rule to a styleSheet you are
            // currently iterating on, otherwise the browser will be stuck
            // in a infinite loop…
            for (const mediaRule of mediaRules) {
                styleSheet.insertRule(mediaRule.cssText)
                hasDarkRules = true
            }
        }
        if (hasDarkRules) {
            loadThemeForm('#theme-selector')
        }
    })
 </script>
 </body>
 </html>
--- a/cache/2024/d4c01d246dc4f43aeb95c427ac5422a5/index.md
+++ b/cache/2024/d4c01d246dc4f43aeb95c427ac5422a5/index.md
@@ -0,0 +1,30 @@
 title: Adactio: Journal-Trust
 url: https://adactio.com/journal/21160
 hash_url: d4c01d246dc4f43aeb95c427ac5422a5
 archive_date: 2024-06-24
 og_image: https://adactio.com/images/photo-300.jpg
 description: How to destroy your greatest asset with AI.
 favicon: https://adactio.com/icon.png
 language: en_IE

 <p>In their rush to cram in “AI” “features”, it seems to me that many companies don’t actually understand why people use their products.</p>

 <p>Google is acting as though its greatest asset is its search engine. Same with Bing.</p>

 <p>Mozilla Developer Network is acting as though its greatest asset is its documentation. Same with Stack Overflow.</p>

 <p>But their greatest asset is actually trust.</p>

 <p>If I use a search engine I need to be able to trust that the filtering is good. If I look up documentation I need to trust that the information is good. I don’t expect perfection, but I also don’t expect to have to constantly be thinking “was this generated by a large language model, and if so, how can I know it’s not hallucinating?”</p>

 <p>“But”, the apologists will respond, “the results are <em>mostly</em> correct! The documentation is <em>mostly</em> true!”</p>

 <p>Sure, but as <a href="https://shkspr.mobi/blog/2024/05/why-do-people-focus-on-failure/">Terence puts it</a>:</p>

 <blockquote>
  <p>The intern who files most things perfectly but has, more than once, tipped an entire cup of coffee into the filing cabinet is going to be remembered as “that klutzy intern we had to fire.”</p>
 </blockquote>

 <p>Trust is a precious commodity. It takes a long time to build trust. It takes a short time to destroy it.</p>

 <p>I am honestly astonished that so many companies don’t seem to realise what they’re destroying.</p>
--- a/cache/2024/index.html
+++ b/cache/2024/index.html
@@ -108,6 +108,8 @@
      
        <li><a href="/david/cache/2024/cba1417ac2338abde14bb06d0a1f505d/" title="Accès à l’article dans le cache local : A Woman Who Left Society to Live With Bears Weighs in on “Man or Bear”">A Woman Who Left Society to Live With Bears Weighs in on “Man or Bear”</a> (<a href="https://bikepacking.com/plog/man-or-bear-debate/" title="Accès à l’article original distant : A Woman Who Left Society to Live With Bears Weighs in on “Man or Bear”">original</a>)</li>
      
        <li><a href="/david/cache/2024/b39aef611801e2b15f189e8d02586b88/" title="Accès à l’article dans le cache local : Notes From “You Are Not A Gadget”">Notes From “You Are Not A Gadget”</a> (<a href="https://blog.jim-nielsen.com/2024/notes-from-you-are-not-a-gadget/" title="Accès à l’article original distant : Notes From “You Are Not A Gadget”">original</a>)</li>
      
        <li><a href="/david/cache/2024/55f475e327a5d1c4851e1e67b19c83e6/" title="Accès à l’article dans le cache local : Notre accompagnement">Notre accompagnement</a> (<a href="https://updr.fr/notre-accompagnement/" title="Accès à l’article original distant : Notre accompagnement">original</a>)</li>
      
        <li><a href="/david/cache/2024/d236f33cf82727313d17cb23bf36a395/" title="Accès à l’article dans le cache local : Reconsider your partnership with Brave">Reconsider your partnership with Brave</a> (<a href="https://kagifeedback.org/d/2808-reconsider-your-partnership-with-brave/6" title="Accès à l’article original distant : Reconsider your partnership with Brave">original</a>)</li>
@@ -124,6 +126,8 @@
      
        <li><a href="/david/cache/2024/a3ccfb51f65cd59f375d5424d243e012/" title="Accès à l’article dans le cache local : In Praise of Buttons">In Praise of Buttons</a> (<a href="https://www.nubero.ch/blog/009/" title="Accès à l’article original distant : In Praise of Buttons">original</a>)</li>
      
        <li><a href="/david/cache/2024/d4c01d246dc4f43aeb95c427ac5422a5/" title="Accès à l’article dans le cache local : Adactio: Journal-Trust">Adactio: Journal-Trust</a> (<a href="https://adactio.com/journal/21160" title="Accès à l’article original distant : Adactio: Journal-Trust">original</a>)</li>
      
        <li><a href="/david/cache/2024/7ed7f4aefae1b5af33b3ec1f607a633f/" title="Accès à l’article dans le cache local : L’Observatoire des perspectives utopiques">L’Observatoire des perspectives utopiques</a> (<a href="https://lobsoco.com/perspectives-utopiques-vague-3/" title="Accès à l’article original distant : L’Observatoire des perspectives utopiques">original</a>)</li>
      
        <li><a href="/david/cache/2024/bd0c3ccce5c3f229f84c6c132ebdaca9/" title="Accès à l’article dans le cache local : L’art de conter nos expériences collectives">L’art de conter nos expériences collectives</a> (<a href="https://blog.notmyidea.org/lart-de-conter-nos-experiences-collectives.html" title="Accès à l’article original distant : L’art de conter nos expériences collectives">original</a>)</li>
@@ -170,6 +174,8 @@
      
        <li><a href="/david/cache/2024/c4751e7c80b292e3533ee6b3e057b702/" title="Accès à l’article dans le cache local : Manuel de survie de la femme dans la tech">Manuel de survie de la femme dans la tech</a> (<a href="https://www.duchess-france.fr/dossier/women%20in%20tech/alli%C3%A9s/2023/01/15/manuel-survie-femme-tech.html" title="Accès à l’article original distant : Manuel de survie de la femme dans la tech">original</a>)</li>
      
        <li><a href="/david/cache/2024/92ca1fc6eb390997db6b76afaf742d63/" title="Accès à l’article dans le cache local : Cinq idées reçues lors de l’école d’été sur les humanités numériques">Cinq idées reçues lors de l’école d’été sur les humanités numériques</a> (<a href="https://www.lobrassard.net/carnet/2024-06-14-idees-recues-ecole-ete-humanites-numeriques.html" title="Accès à l’article original distant : Cinq idées reçues lors de l’école d’été sur les humanités numériques">original</a>)</li>
      
        <li><a href="/david/cache/2024/6c3a6c638625148d7bb6a5c438c9b22a/" title="Accès à l’article dans le cache local : Yearnotes #4 • détour.studio">Yearnotes #4 • détour.studio</a> (<a href="https://détour.studio/yearnotes/4/" title="Accès à l’article original distant : Yearnotes #4 • détour.studio">original</a>)</li>
      
        <li><a href="/david/cache/2024/2a98a563ee9940d690523ce5113b5f9b/" title="Accès à l’article dans le cache local : Environnement - Municipalité de Saint Donat">Environnement - Municipalité de Saint Donat</a> (<a href="https://www.saint-donat.ca/services-aux-citoyens/urbanisme-et-environnement/environnement/" title="Accès à l’article original distant : Environnement - Municipalité de Saint Donat">original</a>)</li>
@@ -208,6 +214,8 @@
      
        <li><a href="/david/cache/2024/a988555163e09729b925dbf715ce256c/" title="Accès à l’article dans le cache local : How web bloat impacts users with slow devices">How web bloat impacts users with slow devices</a> (<a href="https://danluu.com/slow-device/" title="Accès à l’article original distant : How web bloat impacts users with slow devices">original</a>)</li>
      
        <li><a href="/david/cache/2024/54ee1030f6fef778ec10e59726bc5731/" title="Accès à l’article dans le cache local : I've had this thought but never put it into words.">I've had this thought but never put it into words.</a> (<a href="https://micro.anniegreens.lol/2024/05/06/ive-had-this.html" title="Accès à l’article original distant : I've had this thought but never put it into words.">original</a>)</li>
      
        <li><a href="/david/cache/2024/2a1235215c277ebb8a0e9acb7ffd91e0/" title="Accès à l’article dans le cache local : drab - A Headless Custom Element Library">drab - A Headless Custom Element Library</a> (<a href="https://drab.robino.dev/" title="Accès à l’article original distant : drab - A Headless Custom Element Library">original</a>)</li>
      
        <li><a href="/david/cache/2024/faa1d8cae94da6838ff9351e5df791ca/" title="Accès à l’article dans le cache local : Make the indie web easier">Make the indie web easier</a> (<a href="https://gilest.org/indie-easy.html" title="Accès à l’article original distant : Make the indie web easier">original</a>)</li>
@@ -242,10 +250,14 @@
      
        <li><a href="/david/cache/2024/ecae6fcce7e86066e432b5f38b2299ca/" title="Accès à l’article dans le cache local : Everyone seems to forget why GNOME and GNOME 3 and Unity happened">Everyone seems to forget why GNOME and GNOME 3 and Unity happened</a> (<a href="https://liam-on-linux.dreamwidth.org/85359.html" title="Accès à l’article original distant : Everyone seems to forget why GNOME and GNOME 3 and Unity happened">original</a>)</li>
      
        <li><a href="/david/cache/2024/486e49501e495e42b2e40320969c373f/" title="Accès à l’article dans le cache local : Did anyone ask for these AI features? • Cory Dransfeldt">Did anyone ask for these AI features? • Cory Dransfeldt</a> (<a href="https://coryd.dev/posts/2024/did-anyone-ask-for-these-ai-features/" title="Accès à l’article original distant : Did anyone ask for these AI features? • Cory Dransfeldt">original</a>)</li>
      
        <li><a href="/david/cache/2024/4050651a19400713c8563166e2a9abd5/" title="Accès à l’article dans le cache local : Suunto Ambit Black">Suunto Ambit Black</a> (<a href="https://www.suunto.com/fr-ca/Produits/Montres-de-sport/Suunto-Ambit/Suunto-Ambit-Black/" title="Accès à l’article original distant : Suunto Ambit Black">original</a>)</li>
      
        <li><a href="/david/cache/2024/4e49a2509dd9d3f9b33dba3865960465/" title="Accès à l’article dans le cache local : La prise décision par consentement : un outil au service du groupe">La prise décision par consentement : un outil au service du groupe</a> (<a href="http://pilavenir.canalblog.com/archives/2017/10/21/35793152.html" title="Accès à l’article original distant : La prise décision par consentement : un outil au service du groupe">original</a>)</li>
      
        <li><a href="/david/cache/2024/8f9d64c455a9246d5e23810ed10e3fe2/" title="Accès à l’article dans le cache local : Turning the Tables on AI">Turning the Tables on AI</a> (<a href="https://ia.net/topics/turning-the-tables-on-ai" title="Accès à l’article original distant : Turning the Tables on AI">original</a>)</li>
      
        <li><a href="/david/cache/2024/d75afc90a9d3c3b5a56b69446795fbb5/" title="Accès à l’article dans le cache local : plaisir d'ébauche">plaisir d'ébauche</a> (<a href="https://www.la-grange.net/2024/01/06/ebauche" title="Accès à l’article original distant : plaisir d'ébauche">original</a>)</li>
      
        <li><a href="/david/cache/2024/c3272392d462da90874d32841e5caac8/" title="Accès à l’article dans le cache local : Where have all the websites gone?">Where have all the websites gone?</a> (<a href="https://www.fromjason.xyz/p/notebook/where-have-all-the-websites-gone/" title="Accès à l’article original distant : Where have all the websites gone?">original</a>)</li>
@@ -272,6 +284,8 @@
      
        <li><a href="/david/cache/2024/9f8c0e75066c1882a3b4ce084e3223ed/" title="Accès à l’article dans le cache local : The Demo → Demo Loop">The Demo → Demo Loop</a> (<a href="https://daverupert.com/2022/06/demo-to-demo-loop/" title="Accès à l’article original distant : The Demo → Demo Loop">original</a>)</li>
      
        <li><a href="/david/cache/2024/63a6d75aee63c072c6d3b174c918740e/" title="Accès à l’article dans le cache local : Le panneau des préjugés">Le panneau des préjugés</a> (<a href="https://mimo.blog/le-panneau-des-prejuges" title="Accès à l’article original distant : Le panneau des préjugés">original</a>)</li>
      
        <li><a href="/david/cache/2024/9e9c6f97d732010e14201f1624782ddc/" title="Accès à l’article dans le cache local : Engineering for Slow Internet">Engineering for Slow Internet</a> (<a href="https://brr.fyi/posts/engineering-for-slow-internet" title="Accès à l’article original distant : Engineering for Slow Internet">original</a>)</li>
      
        <li><a href="/david/cache/2024/f154db1b6eccf69f498b4a31980367bd/" title="Accès à l’article dans le cache local : The most important goal in designing software is understandability">The most important goal in designing software is understandability</a> (<a href="https://ntietz.com/blog/the-most-important-goal-in-designing-software-is-understandability/" title="Accès à l’article original distant : The most important goal in designing software is understandability">original</a>)</li>
@@ -288,6 +302,8 @@
      
        <li><a href="/david/cache/2024/821fa933883f080d23c0a6d9d0b3721a/" title="Accès à l’article dans le cache local : Make">Make</a> (<a href="https://www.arthurperret.fr/cours/make.html" title="Accès à l’article original distant : Make">original</a>)</li>
      
        <li><a href="/david/cache/2024/8b0a0eb43814f315e2b296aa5bdd5eee/" title="Accès à l’article dans le cache local : ChatGPT et l’indifférence à la vérité">ChatGPT et l’indifférence à la vérité</a> (<a href="https://www.arthurperret.fr/blog/2024-06-21-chatgpt-et-l-indifference-a-la-verite.html" title="Accès à l’article original distant : ChatGPT et l’indifférence à la vérité">original</a>)</li>
      
        <li><a href="/david/cache/2024/3a28346225751986cc06aaadb8c8bb90/" title="Accès à l’article dans le cache local : TechScape: How cheap, outsourced labour in Africa is shaping AI English">TechScape: How cheap, outsourced labour in Africa is shaping AI English</a> (<a href="https://www.theguardian.com/technology/2024/apr/16/techscape-ai-gadgest-humane-ai-pin-chatgpt" title="Accès à l’article original distant : TechScape: How cheap, outsourced labour in Africa is shaping AI English">original</a>)</li>
      
        <li><a href="/david/cache/2024/85b765a918ef094a5a2dd13a1ff5dd7d/" title="Accès à l’article dans le cache local : RSS 2.0 Specification">RSS 2.0 Specification</a> (<a href="https://www.rssboard.org/rss-specification#extendingRss" title="Accès à l’article original distant : RSS 2.0 Specification">original</a>)</li>
@@ -436,6 +452,8 @@
      
        <li><a href="/david/cache/2024/9dee199dbf71b1176eaac521e828693b/" title="Accès à l’article dans le cache local : Why Not Mars (Idle Words)">Why Not Mars (Idle Words)</a> (<a href="https://idlewords.com/2023/1/why_not_mars.htm" title="Accès à l’article original distant : Why Not Mars (Idle Words)">original</a>)</li>
      
        <li><a href="/david/cache/2024/93c287363dbe9cae22f4360c5b107971/" title="Accès à l’article dans le cache local : Generative AI Is Not Going To Build Your Engineering Team For You">Generative AI Is Not Going To Build Your Engineering Team For You</a> (<a href="https://stackoverflow.blog/2024/06/10/generative-ai-is-not-going-to-build-your-engineering-team-for-you/" title="Accès à l’article original distant : Generative AI Is Not Going To Build Your Engineering Team For You">original</a>)</li>
      
        <li><a href="/david/cache/2024/af0aee71fef1821c89ce368622f9a464/" title="Accès à l’article dans le cache local : Accessibility | Tippy.js">Accessibility | Tippy.js</a> (<a href="https://atomiks.github.io/tippyjs/v6/accessibility/" title="Accès à l’article original distant : Accessibility | Tippy.js">original</a>)</li>
      
        <li><a href="/david/cache/2024/7f7afbfc0e4528838aadd7db4ba231cb/" title="Accès à l’article dans le cache local : Understanding AuDHD: The Co-Occurence of Autism and ADHD">Understanding AuDHD: The Co-Occurence of Autism and ADHD</a> (<a href="https://www.relationalpsych.group/articles/understanding-audhd-the-co-occurence-of-autism-and-adhd" title="Accès à l’article original distant : Understanding AuDHD: The Co-Occurence of Autism and ADHD">original</a>)</li>
@@ -472,6 +490,8 @@
      
        <li><a href="/david/cache/2024/6fd58c8dcf1738605fb932bb83f4411a/" title="Accès à l’article dans le cache local : Nous n’avons pas numérisé.">Nous n’avons pas numérisé.</a> (<a href="https://nousnavonspasnumerise.mmibordeaux.com/autopsie/" title="Accès à l’article original distant : Nous n’avons pas numérisé.">original</a>)</li>
      
        <li><a href="/david/cache/2024/463a5fd436bb887f90f9e01b01b7716b/" title="Accès à l’article dans le cache local : What's next for Kagi?">What's next for Kagi?</a> (<a href="https://blog.kagi.com/what-is-next-for-kagi#9" title="Accès à l’article original distant : What's next for Kagi?">original</a>)</li>
      
        <li><a href="/david/cache/2024/90e6434dbda21f9d18ad8fa53c822b47/" title="Accès à l’article dans le cache local : OSI Board Meeting Minutes, Wednesday, March 4, 2009">OSI Board Meeting Minutes, Wednesday, March 4, 2009</a> (<a href="https://opensource.org/meeting-minutes/minutes20090304/" title="Accès à l’article original distant : OSI Board Meeting Minutes, Wednesday, March 4, 2009">original</a>)</li>
      
        <li><a href="/david/cache/2024/359df603dbf60e8476027b2eb26cb7ce/" title="Accès à l’article dans le cache local : uv: Python packaging in Rust">uv: Python packaging in Rust</a> (<a href="https://astral.sh/blog/uv" title="Accès à l’article original distant : uv: Python packaging in Rust">original</a>)</li>