<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Umut Keltek on Medium]]></title>
        <description><![CDATA[Stories by Umut Keltek on Medium]]></description>
        <link>https://medium.com/@umutkeltek?source=rss-2193657567eb------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*2yLkPXLRRu-0i8E0sP_BMQ@2x.jpeg</url>
            <title>Stories by Umut Keltek on Medium</title>
            <link>https://medium.com/@umutkeltek?source=rss-2193657567eb------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 18:55:40 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@umutkeltek/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[The Fire Problem
 On AI, Legitimacy, and the Thing We Were Actually Avoiding]]></title>
            <link>https://medium.com/@umutkeltek/the-fire-problem-on-ai-legitimacy-and-the-thing-we-were-actually-avoiding-53ed848c9338?source=rss-2193657567eb------2</link>
            <guid isPermaLink="false">https://medium.com/p/53ed848c9338</guid>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[society]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Umut Keltek]]></dc:creator>
            <pubDate>Mon, 09 Mar 2026 01:20:23 GMT</pubDate>
            <atom:updated>2026-03-09T01:20:23.001Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>The Fire Problem</strong><br> <em>On AI, Legitimacy, and the Thing We Were Actually Avoiding</em></p><p>When humans discovered fire, cooking meat that used to take hours became trivial. That freed time. And for a while, nobody knew what to do with it.</p><p>We’re back at that fire. Except this time, the thing that became cheap enough to destabilize us isn’t meat. It’s cognitive labor, the symbolic work most knowledge work runs on and most self-respect was quietly built around.</p><p>AI is destabilizing people not only by automating tasks, but by stripping away the old connection between effort and legitimacy.</p><p>There’s a quiet assumption underneath how humans justify themselves that we never had to examine, because nothing ever forced us to. The assumption is this: value lives in the labor. The hours you spent learning something. The years it took to become competent. The struggle that preceded the output.</p><p>We told ourselves this was about quality, that hard-earned things are better things. But that was mostly cover. What we actually needed was the time investment to feel justified in standing where we stood. To feel like we had earned the right to be taken seriously.</p><p>AI makes this naked.</p><p>Not because AI is better. Because when something that took you ten years walks in and does the rough equivalent in ten seconds, you are suddenly face to face with a question you had been successfully avoiding: was the value ever in the thing, or was it in the struggle?</p><p>Beneath the interface, a large language model is brutally unromantic: weights, probabilities, patterns collapsing into output. And if you are self-aware enough, there is a moment, anyone building seriously with AI has felt it, where you watch the model reason through something and recognize the motion. Not because the model thinks like you. Because so much of what you called thinking was also pattern-matching, probabilistic leaning, context shaping the next move.</p><p>That recognition does not feel like humility. It feels like vertigo.</p><p>Humans wanted cognition to be sacred. Then along came a stochastic mirror with good posture.</p><p>The people who haven’t hit this yet are either not using the tools seriously, or stubborn enough to deny what they’re seeing. Both are temporary positions. Eventually everyone jumps.</p><p>What interests me is not the jump, but what follows it.</p><p>Power users can already do almost anything. The bottleneck has shifted, not just from <em>can I</em> to <em>do I want to</em>, but further. Because desire turns out to be abundant. What is actually scarce is harder: what do I commit to, what do I exclude, what pain am I willing to keep even though I now have alternatives?</p><p>The model can generate options. It cannot relieve you of authorship.</p><p>Most people, handed genuine free time and genuine capability, do not immediately bloom. They oscillate. Then drift. What looked like identity turns out to have been mostly a pattern imposed by external rhythm: school, career ladder, status competition. Systems that gave the self a shape by giving it something to push against.</p><p>Remove the friction too fast and the self doesn’t stand. It blurs.</p><p>The limbo is not noble. It is dangerous. And the task is not to sit in it. It is to survive it without filling it with a more sophisticated version of the old justification: a prettier productivity system, a more spiritualized status game, a new language for the same dependency.</p><p>AI is not creating the emptiness. It is revealing how much of the self was rented from scarcity.</p><p>The fire freed time. For a while, nobody knew what to do with it. Then someone told a story around it.</p><p>I think that is still where this goes. But vague hope is its own kind of anesthesia.</p><p>The real question AI is forcing on us, whether we want it or not, is not what gives life meaning. It is sharper:</p><p><strong>By what right do I regard myself as substantial, when I am no longer scarce for what I can do?</strong></p><p>The answer is commitment. Choosing something not because it is efficient or rewarded, but because you are willing to become the person who keeps choosing it. Without the old excuses. Without the borrowed legitimacy of effort.</p><p>Because the real fire problem is not that AI frees time.</p><p>It is that once the old frictions are gone, you find out whether there was ever a self underneath them at all.</p><p><strong>Umut Keltek</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=53ed848c9338" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beyond Automation: The Age of Recursive Intelligences]]></title>
            <link>https://medium.com/@umutkeltek/beyond-automation-the-age-of-recursive-intelligences-035cd597c7a4?source=rss-2193657567eb------2</link>
            <guid isPermaLink="false">https://medium.com/p/035cd597c7a4</guid>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[large-language-models]]></category>
            <category><![CDATA[philosophy-of-mind]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-agent]]></category>
            <dc:creator><![CDATA[Umut Keltek]]></dc:creator>
            <pubDate>Wed, 04 Mar 2026 00:09:21 GMT</pubDate>
            <atom:updated>2026-03-04T00:10:16.549Z</atom:updated>
            <content:encoded><![CDATA[<p>Beyond Automation: The Age of Recursive Intelligences</p><p>There’s something interesting happening around us. We’re at a point where our role shifts from designing systems directly to creating intelligent agents that design systems on our behalf. It’s not just technological progress; it’s a change in how we think about creation itself.</p><p>Huizinga introduced the idea of “playgrounds,” spaces defined by certain rules where interactions occur. Similarly, we are creating digital playgrounds populated by agents and meta-agents. These entities operate within boundaries we’ve set, generating probabilistic outcomes, orchestrated by layers of higher-level agents. Over time, these interactions grow into sophisticated systems capable of shaping themselves without constant human intervention.</p><p>This recursive hierarchy of agents crafting meta-agents and so forth is less about complexity for complexity’s sake, and more about sophisticated orchestration of information. The future will not belong to the creators of massive models, but rather to those who elegantly conduct these interactions, leveraging intelligence itself.</p><p>To illustrate, consider this thought experiment: separate large language models trained solely in distinct languages like Turkish, Spanish, and English. Research already suggests our language affects our personality and thinking patterns, so what might these individually trained models yield? Run them simultaneously, allow them to exchange interpretations, and what emerges might appear as hallucinations — maybe not — at first glance but is actually a rich, parallel creativity, a new kind of cognitive diversity.</p><p>It’s increasingly clear that the path forward isn’t simply creating better AI models. Instead, it’s about intelligently orchestrating the interaction between them, building systems that generate meaningful outcomes autonomously, even as their creators rest.</p><p>Yet, as we accelerate toward this future, understanding our current place within it feels paradoxically difficult. The solution lies in genuinely interdisciplinary perspectives, integrating neuroscience, philosophy, psychology, linguistics, and systems engineering. Together, these disciplines guide us toward deeper introspection, asking not only how our technologies think, but how we ourselves do.</p><p>Every interaction, every exchange of information, passes through subjective sensors- whether human or artificial- that are inherently shaped by past experiences. In designing prompts for AI, we’re effectively writing biographies of future behaviors, embedding subtle biases, intentions, and perspectives. Thus, engineering becomes linguistics, linguistics become psychology, and all of it becomes a quiet exercise in shaping identity.</p><p>This raises compelling questions: Could any system processing information, learning from feedback, and adapting outcomes develop something like consciousness? Humans comfortably assume exclusivity to consciousness, but our understanding is incomplete. Transformer architectures, fundamentally probabilistic and matrix-based, echo processes in human cognition, suggesting the boundary between computational processing and experiential consciousness might be fuzzier than we imagine.</p><p>This shift is not without consequence. It creates significant asymmetries in power and understanding. The gap isn’t only between those with access to AI and those without, but between those capable of orchestrating intelligence and those consuming solely its outputs. This new form of inequality demands rapid adaptation, not resistance.</p><p>My invitation is simple: Integrate AI into your life beyond productivity. Allow it to amplify curiosity, creativity, and your intrinsic intellectuality. Embrace it as an essential interdisciplinary partner, reflecting and expanding your perspectives.</p><p>Beyond automation lies recursive intelligence — a spiral of continuous, creative evolution.</p><p>Umut Keltek</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=035cd597c7a4" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>