<?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 Bob Seaton on Medium]]></title>
        <description><![CDATA[Stories by Bob Seaton on Medium]]></description>
        <link>https://medium.com/@bobse?source=rss-93adf9efbb97------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Mc4quShh-298eRs-Hk6VKw.jpeg</url>
            <title>Stories by Bob Seaton on Medium</title>
            <link>https://medium.com/@bobse?source=rss-93adf9efbb97------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 18:38:09 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@bobse/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[Life is like…]]></title>
            <link>https://medium.com/@bobse/life-is-like-3f161963d925?source=rss-93adf9efbb97------2</link>
            <guid isPermaLink="false">https://medium.com/p/3f161963d925</guid>
            <category><![CDATA[creativity]]></category>
            <category><![CDATA[large-language-models]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[software-architecture]]></category>
            <dc:creator><![CDATA[Bob Seaton]]></dc:creator>
            <pubDate>Sat, 07 Feb 2026 05:11:32 GMT</pubDate>
            <atom:updated>2026-02-07T05:11:32.394Z</atom:updated>
            <content:encoded><![CDATA[<p>… a puzzle. Or maybe, a puzzle is just like life.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eTiHTvOA-lTafjr_NjCr9w@2x.jpeg" /></figure><p>One of the things that my wife and I and our children (when they still lived with us) like to do during the Christmas holidays was puzzles. And as the kids got older and smarter, the rules had to change in order to keep it interesting (especially if we had done the same puzzle multiple times before). And so, rule number one is “you can’t look at the box” .</p><p>And when you don’t look at the box, new worlds open up. Instead of having a big picture of what the end result will look like, you have to come up with a different strategy. Think different. Feel your way, like you’re walking down a dark hallway at night with toys on the floor.</p><p>So, the edges come out first. And then comes separation by color. Then local pattern assembly. Clusters appear. And slowly, by sheer force of will, and a little bit, but not too much, trial and error, like Austin and San Antonio, or Washington DC and Baltimore, clusters merge and become something new. Iterate. Repeat.</p><p>Building software today, or reverse engineering a system written in COBOL back in the 1990s, utilizes this same iterative process.<em> I suppose this is also a decent algorithm for a language model to leverage when doing the same job. </em>And<em> </em>with an appropriately powered LLM, we can now <strong><em>assemble</em></strong> and <strong><em>cluster</em></strong> across multiple axes at a time, each axis of decomposition informing the rest/the whole, like the 3D projections from the cover of Gödel, Escher, Bach.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/150/1*MrRgHYMhYzpPxkICqzJ97w@2x.jpeg" /></figure><p>https://pictures.abebooks.com/inventory/md/md30316896719.jpg</p><p>And then there’s my marriage, my wife, my spouse, my partner, and (mostly) my friend for the past 36+ years. Our relationship is very much like that puzzle above. We come from the same box in the same factory, but she and I do things very differently. For this puzzle I did the interior and she did the border, which is a beautiful metaphor for the way we are — she likes defined boundaries and rules/structure; my borders are erractic and messy and unpredictable. But together, with all our differences, when put together, we become whole. But just like putting a puzzle together without seeing what that big picture is, mainly only the local pieces/ clusters (the day to day stuff), but having a process, and having faith that through will and grit and perseverance and hope and love, we will discover that big picture together.</p><p>❤️puzzles</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3f161963d925" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The New Natural Philosopher]]></title>
            <link>https://medium.com/@bobse/the-new-natural-philosopher-0194f43ccb63?source=rss-93adf9efbb97------2</link>
            <guid isPermaLink="false">https://medium.com/p/0194f43ccb63</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[education]]></category>
            <category><![CDATA[creativity]]></category>
            <category><![CDATA[future]]></category>
            <category><![CDATA[science]]></category>
            <dc:creator><![CDATA[Bob Seaton]]></dc:creator>
            <pubDate>Sat, 03 Jan 2026 21:24:15 GMT</pubDate>
            <atom:updated>2026-01-07T05:14:41.910Z</atom:updated>
            <content:encoded><![CDATA[<h4>A Curriculum for the AI Renaissance</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wF7u-uwgs8fI5ewXBKI_VQ.jpeg" /><figcaption>interconnected disciplines — courtesy of nano banana</figcaption></figure><h3>The Spark</h3><p>This document began as a response to Avi Loeb’s January 2026 Medium article, <a href="https://avi-loeb.medium.com/how-many-einstein-level-scientists-had-ever-lived-45e55b539dbf">“How Many Einstein-Level Scientists Had Ever Lived?”</a></p><p>Loeb asks why we don’t seem to have more Einsteins despite a 4x larger population. His answer leans toward institutional ossification (and, characteristically, aliens).</p><p>But there’s another angle worth exploring — one about <strong>hyper-specialization</strong>. When there’s too much to know, you have to choose between wide and deep. The rational choice, especially for PhDs chasing research dollars, has been to go narrow and go deep.</p><p>But what if that tradeoff is changing?</p><blockquote>TL;DR</blockquote><blockquote>Breakthroughs come from seeing connections across domains. Einstein wasn’t a mathematician — but he’d read Mach and befriended Grossmann. Darwin wasn’t an economist — but he’d read Malthus. Jobs wasn’t an artist — but he’d taken calligraphy.</blockquote><blockquote>The pattern is always the same: “X wasn’t a Y, but they’d read Y.”</blockquote><h3>The Core Thesis</h3><p><strong>The ability to see connections across domains is the meta-skill. Everything else is in service of developing that capacity.</strong></p><p>This isn’t a curriculum for training academics or producing physicists. It’s a model for cultivating a <em>way of seeing</em> — a cognitive architecture that keeps the mind porous enough to notice what specialists, by definition, have to ignore.</p><p>The hyper-specialization of modern education was a pragmatic response to an explosion of knowledge that exceeded human cognitive limits. With AI as an “exocortex” providing depth-on-demand, we can return to the Natural Philosopher model — minds that are broad, integrative, and focused on <em>taste</em> rather than <em>memorization</em>.</p><p>This applies everywhere:</p><ul><li>The physicist looking for a new theory</li><li>The doctor diagnosing a case that doesn’t fit the textbook</li><li>The entrepreneur seeing a market gap no one else noticed</li><li>The artist finding a new form</li><li>The tradesperson solving a problem the manual doesn’t cover</li></ul><p>It’s all the same underlying cognitive move: <strong>“Huh, this reminds me of…”</strong></p><p>That pattern-match — the analogy-bridge — is what we’re training for. The rest is scaffolding.</p><h3>Part I: Foundations (Years 1–2)</h3><h3>The Analogy Library</h3><p>The goal is maximum exposure to <em>transferable mental models</em> across domains. Not mastery — exposure deep enough to pattern-match later.</p><p><strong>The Unreasonable Effectiveness of Analysis + Synthesis</strong></p><p>Borrowing from Wigner’s famous observation about mathematics: there’s something “unreasonably effective” about combining analysis (taking things apart, understanding components, going deep) with synthesis (putting things together in new combinations, seeing the whole, going broad).</p><ul><li>Analysis alone = you understand your piece but can’t see the system</li><li>Synthesis alone = you see connections but can’t execute</li><li>Both = the thing that produces breakthroughs</li></ul><p>The Analogy Library is the raw material for synthesis. You can’t synthesize across domains you’ve never encountered. But synthesis without the ability to analyze — to decompose a problem, to verify that an analogy is deep rather than superficial — is just hand-waving.</p><p>The curriculum trains both: broad exposure for synthesis, deep dives for analysis, and the judgment to know when each is needed.</p><p>DomainFocusWhy It Matters<strong>Physics</strong>Mechanics, thermodynamics, electromagnetism, quantum basicsThe ur-language of how systems behave<strong>Mathematics</strong>Calculus, linear algebra, probability, graph theoryThe grammar of formal reasoning<strong>Biology</strong>Evolution, ecology, molecular biology, neuroscienceComplex adaptive systems, emergence<strong>Philosophy</strong>Epistemology, philosophy of science, logic, ethicsHow do we know things? What counts as explanation?<strong>History of Science</strong>Revolutions, dead ends, paradigm shiftsHow breakthroughs actually happen (spoiler: not linearly)<strong>Art &amp; Music</strong>Theory, composition, visual designPattern, structure, aesthetics — non-verbal reasoning<strong>Literature</strong>Close reading, narrative structure, rhetoricHow humans make meaning; persuasion<strong>Economics</strong>Micro, game theory, mechanism designIncentives, equilibria, strategic behavior</p><p><strong>Pedagogy:</strong> These are taught <em>integratively</em>, not as siloed courses. A unit on “feedback systems” might weave through thermodynamics, ecology, economics, and control theory in the same month.</p><h3>Part II: The Core Skills (Woven Throughout)</h3><h3>1. Analogical Reasoning</h3><p>Explicit training in:</p><ul><li>Recognizing structural similarity across domains</li><li>Distinguishing deep analogies (shared causal structure) from shallow ones (surface resemblance)</li><li>Using analogies to generate hypotheses, not just illustrate</li></ul><p><strong>Exercise:</strong> Given a problem in field A, students must find at least 3 analogous problems in unrelated fields and evaluate which analogy is most generative.</p><p><em>Note: This must be done in-person, without AI assistance. The goal is to internalize cross-disciplinary pattern matching as a reflex, not to outsource it. Using AI here defeats the entire purpose — this is the exact muscle we’re trying to build.</em></p><h3>2. AI Partnership</h3><p>This is a <em>skill</em>, not a given. Training includes:</p><ul><li>Prompt engineering and iterative refinement</li><li>Knowing when AI is likely to be wrong (and why)</li><li>Using AI to stress-test ideas, not just generate them</li><li>The art of pushing back — treating AI as a collaborator, not an oracle</li></ul><p><strong>The Analysis/Synthesis Division of Labor</strong></p><p>The human-AI partnership is most powerful when it respects what each does best:</p><ul><li><strong>Human:</strong> Synthesis — taste, pattern-matching, “this reminds me of…”, the oblique angle</li><li><strong>AI:</strong> Analysis support — break this down, check the components, verify the math, surface related work</li></ul><p>The human stays in the creative driver’s seat. The AI is a Mechanical Grossmann: broad knowledge, available on demand, but lacking the intuition to know which connections matter. That judgment remains yours.</p><p><strong>Exercise:</strong> Students must use AI to explore a topic they know well, identify where it goes wrong, and document the failure modes.</p><h3>3. Comfort with Ambiguity</h3><p>Most education trains you to find <em>the answer</em>. Breakthrough thinking requires sitting with uncertainty.</p><ul><li>Extended projects with no clear solution path</li><li>Journaling on “half-formed ideas” without pressure to resolve them</li><li>Exposure to problems that took decades/centuries to solve — what did the intermediate states look like?</li></ul><h3>4. Intellectual Courage</h3><ul><li>Defending unpopular positions (assigned, to separate ego from argument)</li><li>Public critique and being critiqued</li><li>Studying historical “crackpots” who were right (Wegener, Semmelweis, McClintock)</li><li>Studying crackpots who were wrong — what’s the difference?</li></ul><h3>Part III: The Deep Dive (Years 3–4)</h3><p>After building breadth, students choose <strong>one domain for depth</strong> — but the framing is different.</p><h3>Your Gravitational Center of Mass</h3><p>You need <em>somewhere</em> to stand. A domain where you have:</p><ul><li>Enough technical facility to execute, not just opine</li><li>Genuine open problems you care about</li><li>A community to engage with</li></ul><p>But the home base is a <em>launching point</em>, not a prison. The expectation is that your best work will involve <strong>importing ideas from elsewhere</strong>.</p><h3>The Grossmann Requirement</h3><p>Every student must form at least one deep collaboration with someone in a <em>different</em> field. The deliverable is a <strong>mandatory joint project</strong> that neither could have done alone — not optional, not “encouraged,” <em>required</em>.</p><p>This isn’t group work for the sake of group work. It’s training for the actual structure of breakthrough work. Einstein without Grossmann doesn’t get general relativity. Period.</p><h3>Part IV: Capstone — The Natural Philosophy Project</h3><p>A year-long, largely self-directed investigation with these requirements:</p><ol><li><strong>Must cross at least two domains</strong> — the interesting question lives at an intersection</li><li><strong>Must include a “what if I’m wrong” analysis</strong> — steel-man the opposition to your own thesis</li><li><strong>Must use AI as a documented collaborator</strong> — include a reflection on what the AI contributed and where it failed you</li><li><strong>Must be communicable to a general audience</strong> — if you can’t explain it broadly, you don’t understand it</li></ol><h3>Evaluation Criteria</h3><p>Not: “Is this correct?”</p><p>But:</p><ul><li>Is this an interesting question?</li><li>Is the analogical reasoning sound?</li><li>Did you navigate uncertainty productively?</li><li>Did you surprise yourself (and us)?</li></ul><h3>Institutional Requirements</h3><p>For this to work, you need:</p><h3>1. Slack</h3><p>Time to think without constant assessment pressure. Einstein had the patent office. Students need equivalent “incubation space.”</p><p>Proposal: <strong>One day per week with no classes, no deadlines</strong> — dedicated to self-directed exploration. Loosely mentored, not graded.</p><h3>2. Serendipity Infrastructure</h3><p>Breakthroughs often come from unexpected collisions. Design for this:</p><ul><li>Physical spaces that mix disciplines</li><li>Seminars where a biologist presents to physicists, etc.</li><li>“Random coffee” matching across departments</li></ul><h3>3. Permission to Fail</h3><p>Most bold ideas are wrong. A curriculum that punishes failure trains people to be timid.</p><p>Proposal: <strong>A “failed experiments” portfolio</strong> — students document ideas that didn’t pan out, with reflection on what they learned. This is valued, not hidden.</p><h3>4. Faculty Who Model This</h3><p>You can’t teach Natural Philosophy if all your faculty are hyper-specialists who’ve never crossed a boundary. Hiring and tenure must reward integrative work.</p><h3>The Bet</h3><p>This curriculum is a bet that:</p><ol><li>AI changes the depth/breadth tradeoff fundamentally</li><li>The limiting factor on breakthroughs is now <em>taste and integration</em>, not <em>access to knowledge</em></li><li>We can train for the kind of thinking that produces Einsteins — it’s not purely innate</li><li>The institutional structures that suppress this are choices, not necessities</li></ol><p>If we’re right, this produces a generation of New Natural Philosophers — people who can see across boundaries, use AI to execute on their visions, and ask the questions that specialists are too narrow to notice.</p><p>If we’re wrong, we’ve at least produced well-rounded humans who can think clearly and adapt to a changing world.</p><p>That’s a bet worth making.</p><h3>Case Studies Across Domains</h3><p>These aren’t just historical curiosities. They’re evidence that the analogy-bridge is how breakthroughs actually happen — in every field.</p><h3>Physics: Einstein &amp; Grossmann</h3><p>Einstein had the physical intuition that gravity must be geometric. He didn’t have the math. Grossmann, a mathematician friend, recognized that Riemannian differential geometry — an obscure, “useless” branch of pure math — was exactly what Einstein needed. Neither alone produces general relativity. The breakthrough lived in the connection between two different ways of seeing.</p><p><em>Bonus:</em> Grossmann’s father got Einstein the patent office job that gave him slack to think. One friend, two pivotal contributions to the same revolution.</p><h3>Technology: Steve Jobs &amp; Calligraphy</h3><p>Jobs audited a calligraphy class at Reed after dropping out — zero career relevance. Ten years later, the Macintosh had proportionally spaced fonts and typography that no other computer had. A “useless” aesthetic education became a competitive advantage in technology because Jobs could see a connection no engineer would think to look for.</p><h3>Medicine: Marshall &amp; Warren (H. pylori)</h3><p>For decades, the medical establishment “knew” that ulcers were caused by stress and lifestyle. Barry Marshall and Robin Warren proposed bacterial infection — an idea so heterodox Marshall literally drank H. pylori to prove it. Their willingness to import thinking from microbiology into gastroenterology (and ignore disciplinary boundaries) won them a Nobel Prize and changed treatment for millions.</p><h3>Art: Picasso &amp; African Masks</h3><p>Picasso could draw with academic precision — he had depth. But Cubism emerged from exposure to African sculpture, Cézanne’s geometric experiments, and ideas about representing multiple perspectives simultaneously. The breakthrough wasn’t in painting technique. It was in <em>seeing</em> that representation itself could be questioned.</p><h3>Trades: The Plumber Who Thinks in Systems</h3><p>A master plumber who understands fluid dynamics, thermodynamics, and a bit of chemistry sees a building differently than one who learned only procedures. When something weird happens — a problem not in the manual — the systems thinker has analogies to draw from. They can reason from principles, not just pattern-match to known solutions.</p><h3>Aviation: Chuck Yeager &amp; the Engineers</h3><p>Chuck Yeager didn’t just fly the X-1 past Mach 1 on reflexes and courage. As described in <em>The Right Stuff</em>, he deliberately spent time with the engineers, studying the aircraft’s design, learning to think the way they thought.</p><p>Why? Because when you’re at the edge of the envelope — where no pilot has been before — your pilot training runs out. The controls do something unexpected. The airframe behaves strangely. You’ve got two options:</p><ol><li>Pattern-match to training (which doesn’t cover this, because no one’s been here)</li><li>Reason from principles (“what would the engineers say is happening?”)</li></ol><p>Yeager built option 2 into his head <em>before</em> he needed it. He loaded the engineers’ mental models into his analogy library so he could think like them at Mach 0.96 when thinking like a pilot wasn’t enough.</p><p>And there’s a beautiful recursion: Yeager’s cross-domain thinking let him push the envelope, which generated data the engineers used to build better planes, which the next generation of pilots flew further. The collaboration closed a loop that accelerated the whole program.</p><p>Test pilot as Natural Philosopher.</p><h3>Biology: Darwin &amp; Economics</h3><p>Darwin’s theory of natural selection was directly influenced by Malthus’s economic writings on population and resource competition. An economic mental model became the scaffold for the most important idea in biology. Darwin wasn’t an economist — but he’d <em>read</em> economics.</p><h3>The Pattern</h3><p>In every case:</p><ol><li>Exposure to something “outside the lane” (building the synthesis library)</li><li>That exposure sits dormant (sometimes for years)</li><li>A problem arises where the unexpected connection becomes visible</li><li>Analysis verifies the connection is deep, not superficial</li><li>The breakthrough happens at the intersection</li></ol><p>The magic is in the interplay: synthesis to <em>see</em> the connection, analysis to <em>validate</em> it. You can’t force step 3. But you can structure education to maximize steps 1 and 2 — and train the analytical chops to execute on steps 4 and 5 when the moment comes.</p><p><em>“Imagination is more important than knowledge. Knowledge is limited. Imagination encircles the world.”</em> — Albert Einstein</p><p><em>“Grossmann, you must help me, or else I’ll go crazy.”</em> — Also Albert Einstein</p><p>The second quote is what makes the first one possible.</p><h3>Appendix: In Praise of Grossmanns</h3><h3>Marcel Grossmann (1878–1936)</h3><p>This document owes its existence to a man most people have never heard of.</p><p>Marcel Grossmann met Albert Einstein at ETH Zurich. They became close friends. When Einstein struggled academically and couldn’t find work after graduation, it was Grossmann’s father who pulled strings to get him a job at the Swiss Patent Office — the job that gave Einstein the slack to think, free from academic publish-or-perish pressure.</p><p>Years later, when Einstein had the physical intuition for general relativity but lacked the mathematical tools, he turned to Grossmann again. “Grossmann, you must help me, or else I’ll go crazy.” Grossmann recognized that Riemannian differential geometry — obscure, “useless” pure mathematics — was exactly what Einstein needed. He taught Einstein the tensor calculus that became the language of curved spacetime. ❤️</p><p>Grossmann co-authored key papers with Einstein. He never sought the spotlight. He died of multiple sclerosis at 58, largely forgotten outside specialist circles.</p><p><strong>Without Grossmann:</strong></p><ul><li>No patent office job → no incubation period</li><li>No differential geometry → no mathematical framework</li><li>No general relativity as we know it</li></ul><p>Einstein’s name is synonymous with genius. Grossmann’s name is a footnote. This is how breakthroughs actually work.</p><h3>The Fullback Problem</h3><p>For those who prefer American football to physics:</p><p>Daryl “Moose” Johnston played fullback for the Dallas Cowboys from 1989 to 1999. His job was to run through the gap first, absorb the linebacker, and clear the way for Emmitt Smith — who became the NFL’s all-time leading rusher and a Hall of Famer.</p><p>Moose made 2 Pro Bowls. Emmitt made 8 and got a bronze bust in Canton.</p><p>But here’s the thing: no Moose, no Emmitt. No Emmitt, no three Super Bowls. No dynasty.</p><p>The highlight reel shows Emmitt crossing the goal line. It doesn’t show Moose on the ground with a linebacker on top of him, having done his job.</p><h3>The Lesson</h3><p>Every Einstein needs a Grossmann. Every Emmitt needs a Moose.</p><p>The breakthroughs we celebrate are collaborations, even when we only remember one name. The genius had taste, vision, and intuition — but someone else often provided the tools, the access, the blocking.</p><p>When we design educational systems, we should remember: we’re not just trying to produce Einsteins. We’re trying to produce Grossmanns too. And — critically — we’re trying to create conditions where they find each other.</p><h3>AI as a “Mechanical Grossmann”</h3><p>The AI-as-Grossmann model in this curriculum is an attempt to democratize access to that partnership. Not everyone can befriend a mathematician whose father knows people at the patent office. But everyone might be able to access a “Mechanical Grossmann” — a collaborator with broad knowledge who’s available on demand and happy to help.</p><p>The term is a deliberate inversion of the Mechanical Turk — the famous 18th-century chess-playing “automaton” that was actually a human hidden inside a machine, pretending to be artificial intelligence.</p><ul><li><strong>Mechanical Turk:</strong> Human pretending to be machine (deception)</li><li><strong>Mechanical Grossmann:</strong> Machine serving a human collaborative role (augmentation)</li></ul><p>The Turk was a trick. The Grossmann is a tool. We’re not pretending the AI is human or that it replaces human collaboration. We’re naming the <em>function</em> it serves — and honoring the man who showed what that function looks like when done by a true friend.</p><p>Grossmann was a real friend who believed in Einstein before anyone else did. An AI can’t do that. It’s not the same. But it is something. And it might be enough to help a few more people see connections they wouldn’t have seen alone.</p><p><em>Co-authored with ❤️ claude-opus-4–5–20251101, my Mechanical Grossmann</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0194f43ccb63" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[something funny …]]></title>
            <link>https://medium.com/@bobse/something-funny-672b5d89c7f7?source=rss-93adf9efbb97------2</link>
            <guid isPermaLink="false">https://medium.com/p/672b5d89c7f7</guid>
            <category><![CDATA[writing-prompts]]></category>
            <category><![CDATA[ai-image-generation]]></category>
            <category><![CDATA[ai-storytelling]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Bob Seaton]]></dc:creator>
            <pubDate>Sat, 13 Dec 2025 06:39:18 GMT</pubDate>
            <atom:updated>2025-12-14T16:33:53.796Z</atom:updated>
            <content:encoded><![CDATA[<h3>something funny …</h3><p>I just had one of my favorite prompts (I don’t know if that’s a thing, but even that statement cracks me up) that turned into a prompt sequence. What follows documents a recent conversation with the wonderfully steerable <a href="https://aistudio.google.com/prompts/new_chat?model=gemini-3-pro-image-preview">nano banana</a>.</p><p>Scene opens: So there I was, looking at the <a href="http://twitter.com/aisdk">@aisdk</a> (<a href="https://ai-sdk.dev/playground">https://ai-sdk.dev/playground</a>) (%npm i ai”) for some interface ideas and saw this :</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/398/1*ecM1A3dZACuLT9rGT2T02A.png" /><figcaption>Models in Bedrock</figcaption></figure><p>&gt;&gt; “look at all of those models in bedrock”, said the voices in my head, which made me think of the Flintstones comics/shows as a kid, which made me smile, which made me want of a picture of bunch of Flintstone and Rubbles and their friends and acquaintances hanging out and going to the Drive In and the prehistoric Sonic, all in beautiful downtown Bedrock, AWS :D :D :D</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mmstIVEuCXKtLGbDi7dIBw.jpeg" /><figcaption>Gen-1 ::nano banana output using the preceding paragraph</figcaption></figure><p>&gt;&gt; “maybe a little more top-down and Isomorphic Art style, maybe on a Lego Green Flat Pieces plot of Earth, like it was rendered in Blender, giving it that retro-pre-historic TRON”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_p4tCqhc4qKApij_uEdzbQ.jpeg" /><figcaption>Gen-2::nano banana output using the preceding paragraph</figcaption></figure><p>&gt;&gt; “that is wicked good. I wish it had a little more light stone (limestone, not granite), was daytime instead of night, with the camera pulled back just a little bit (so that more of the town is visible) — dream with me”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*l1riQg6NE6saiW_jVXSo9g.jpeg" /><figcaption>Gen-3::nano banana output using the preceding paragraph</figcaption></figure><p>&gt;&gt; “omg — so good. zoom out/ back again so more of the town is visible. Remember Bedrock is on planet AWS so they’re all busy doing cool fun creative jobs, the pathways that connect communities shimmer in gold-blue — and ((((something we’ll do next, so just keep it in mind here without drawing it: if one of them wants to go over there to another community, they just get on that pathway (it shrinks them down lossless compression, expansion on the other side; in-town its mainly sidewalk). with a little techno StarWars vrrupth)))”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xW2CxAh6lFvlACOMKx6BAw.jpeg" /><figcaption>Gen-4::nano banana output using the preceding paragraph</figcaption></figure><p>&gt;&gt; “can you make the pathway a little more like those pneumatic tubes you used to find in hospitals and industry — kinda the 1950’s internet for atoms, not bits -”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*69JaINI8TvVkxkC70ampIA.jpeg" /><figcaption>Gen-5::nano banana output using the preceding paragraph</figcaption></figure><p>&gt;&gt;”zoom out one more time with the curve of AWS almost visible, but we can see the green lego countryside outside of Bedrock”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*t2-iqzT2E--DuHq-LgNJ4w.jpeg" /><figcaption>Gen-6::nano banana output using the preceding paragraph</figcaption></figure><p>like exploring the mandelbrot space… with a little minecraft love</p><p>fun times.</p><p>Copyright © 2025 Robert W Seaton Jr <br><em>dibs to movie and book rights — credit to all the brands mentioned.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=672b5d89c7f7" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>