I’d had the kind of week with Artwell where you either ship something or break something, and I’d managed both, occasionally to the same feature. I’d also spent it neck-deep in the debate eating every feed I follow lately: the one about AI and taste, about whether, now that the machine can make the thing, having good judgment is enough to make you a creator – or only a critic. So on Friday, fried on both counts, I took myself out for a steak.

Two tables over, a couple was having an argument, and it took me a full minute to work out what they were arguing about.

She’d said the thing was technically flawless and completely forgettable. He said she was being impossible – it was obviously high quality, look at it. She said quality wasn’t the same as good. He said that was a meaningless distinction. She said it was the only distinction that mattered. He said, and I’m quoting, “everybody’s suddenly an expert.”

It was, almost word for word, the argument I’d been reading all week. So I assumed I was eavesdropping on two people from my own corner of the world – that the debate had followed me out of my screen and into dinner. I was, I admit, a little impressed by their command of the material.

Then he delivered the closer: “If you think it’s so easy, I’d love to watch you cook it better.”

Ah. The steak. They were arguing about the steak.

Here’s the part that stayed with me, past the steak and past the secondhand discomfort of watching a date go sideways over doneness. I hadn’t misheard them. I’d heard them perfectly. The argument about the steak and the argument about AI are, line for line, the same argument. You can swap “the steak” for “the output” in every sentence either of them said and lose nothing at all. That isn’t a coincidence, and it isn’t because steak is secretly profound. It’s because both arguments are circling the same confusion, and neither side has bothered to name it. But I will try.

There are two different things we call “taste,” and they only look like one thing from a distance.

The first is the diner’s taste. You take a bite and you know. Too salty, the texture’s gone rubbery, the sauce is fighting the meat instead of helping it. You don’t need a chef’s jacket to register any of that, and the man’s “I’d love to watch you cook it better” was a bad move precisely because she never claimed she could cook it. She claimed she could tell it was gray in the middle. Those are not the same claim. She was right, and he reached for a credential to avoid admitting it. Anyone who has ever been handed “well, you’re not a chef” while pointing at an obviously overcooked steak knows the move. It’s a dodge. The diner’s verdict stands on its own.

The second is the chef’s taste, and it only shows itself when the plate is wrong and has to be fixed. The diner says “this is too salty.” The chef tastes the same bite and knows the saltiness is actually a fat problem, and the fix isn’t less salt, it’s a hit of acid to cut it. The diner can reject. The chef can correct. The distance between “this is off” and “here is the specific move that makes it right” is the entire job, and it’s the part you only get from having stood at the stove and ruined the dish a few hundred times.

Hold onto that gap, because the AI conversation lives or dies in it.

The loudest story in AI right now is a democratization story, and the optimistic version of it is genuinely true. The barrier to making things has collapsed. Anyone can generate a draft, a design, a working prototype, a passable first cut of almost anything, in the time it used to take to schedule the kickoff meeting. From there the discourse has converged, with remarkable speed, on a single comforting conclusion: since the machine now handles the making, the scarce and valuable thing left to humans is taste. Judgment is the new moat. Curators beat creators. The editor outranks the engineer. There are now four-part courses promising to train your taste in ninety days, as if taste were a certification.

And the first half of that is correct. When generation is free, discernment is what’s scarce, and the person who can tell the good output from the plausible one is holding something real. The couple at the next table had stumbled into the true half: you don’t need to be a chef to know the steak is overcooked.

But watch the move the discourse makes next, because it’s the same move the man made, just run in reverse. He took “you can’t judge it unless you can cook it” – false – and used it to win an argument he was losing. The AI conversation takes “you don’t need to be a chef to judge it” – true – and quietly extends it into “so you don’t need to be a chef to make it.” That sequel does not follow. And the reason it doesn’t follow is that when you build with AI, you are no longer the diner. You’re the one in the kitchen.

Think about what the machine actually is in this arrangement. It’s a line cook with flawless technique and no palate of its own. It will execute anything you can specify, instantly, without complaint, and it will re-plate the identical wrong dish all night because it cannot taste that the dish is wrong. At a restaurant, the diner’s “this is off” is enough, because there’s a chef back there with the second kind of taste who translates the complaint into a fix. In an AI build, there is no chef back there. There’s just the line cook and you. If all you’ve brought is the diner’s palate, you can tell the plate is wrong, but you can’t say acid-not-salt, so you send it back on a loop and get something differently wrong every time. The only thing that breaks the loop is the correction, and the correction is the chef’s.

There’s a reason the whole culture reached for “slop” to name what’s wrong with mediocre machine output – a feed-trough word the dictionaries crowned word of the year. We walked all the way into the kitchen to find the insult and didn’t notice we’d done it.

Here’s what that looks like from inside my own build, on a perfectly ordinary day.

Artwell turns long-form audio into clean, speaker-attributed transcripts and then into story. So transcription quality is foundational, and at one point I had two transcripts of the same recording sitting side by side. One of them looked obviously better. It came dressed with segments, summaries, topic tags, sentiment analysis – the whole garnish tray. The other was plainer. Word-for-word they agreed about ninety-six percent of the time, close enough that a quick look would call them equivalent and reach for the prettier plate.

The plainer one was the better transcript, and it wasn’t close. The fancy one had quietly turned “Puget Sound Energy” into “Pizza Time Energy,” and a neighborhood named SoDo into “So To.” Those aren’t formatting nitpicks. In a transcript meant to carry someone’s actual words into a published story, a confidently wrong proper noun is the gray center of the steak – the thing that looks fine under the garnish and destroys trust the moment anyone who was in the room reads it. Choosing correctly here was not “which output has more features.” It was knowing that for this product, quality means the words and the names survive intact, and the analytics layer is parsley. That’s not a verdict a diner’s palate can deliver. It’s a directorial call about what the dish is even for.

It happens at every layer. Another time, a fix to the audio-editing playback sailed through every check – tests green, types clean, lint quiet – and was still wrong, because it had bought its passing grade by making playback hesitate for a beat before it started, trading the immediate, responsive feel of the thing for technical correctness. Every box ticked. The dish still didn’t eat right. Passing the recipe is not the same as the plate being good, and no test suite has ever been able to taste the difference.

Now the uncomfortable part, the part those train-your-taste courses are selling around. The chef’s palate – the one that can direct, not just judge – is not built by curating. You do not develop it by assembling a reference library of great plates and studying them, any more than you become a winemaker by becoming a sommelier. The sommelier develops exquisite preferences by tasting. The winemaker develops judgment by making bad wine for years until it stops being bad. They are different organs, grown by different work, and only one of them can fix the batch. The judgment that lets me see past “Pizza Time Energy” came from two decades of being responsible for the finished thing – of shipping the wrong cut, watching it land badly, and learning in my hands what the fix actually was. It did not arrive on a schedule, and it did not come from watching.

Which is the quiet problem underneath the cheerful democratization story. We have declared judgment the most valuable thing in the economy at the precise moment we’re dismantling the work that produces it. The reps that used to build the chef’s palate – the grunt passes, the owned mistakes, the years on the line – are exactly the reps the machine now does for you. You can have the diner’s palate for free. The chef’s palate was never free, and it’s getting more expensive to acquire, not less, because the path to it is quietly closing.

I can’t reopen the path for everyone coming up behind me, and I’m not going to pretend a LinkedIn post does anything about it. What I can control is whether I keep doing my own reps. So here’s the thing I actually do, and it’s the reason my build process looks heavier than a solo founder’s “should.” The multi-session, multi-role orchestration I run – the planning panels, the forced checkpoints, the reviewer pass that exists only to argue with the implementer pass – is not only quality control. It’s a structure that keeps me at the stove. Left alone with a fluent line cook, the gravitational pull is to slide back into the diner’s chair: glance at the plate, looks plausible, ship it, send it back if someone complains. The orchestration is what stops the slide. It forces the second-order question every time – not “is this good,” but “what is this allowed to become, and where does it live, and what’s the one cut that keeps the product true.” It makes me direct instead of merely approve. In a world where the reps don’t happen for free anymore, I’ve built a way to keep doing them on purpose. It’s an apprenticeship for one, and I’m both the master and the kid sweeping the floor.

The couple never resolved it, for the record. He paid, stiffly. She was still right.

But the argument they were having is the one the whole industry is having, and it’s worth getting right, because the stakes aren’t really about steak or even about taste. The people who can build something good with AI are the ones who already spent the years at the stove – who carry the chef’s palate into a kitchen that now runs at impossible speed. The discourse is handing everyone else a flawless line cook and a diner’s palate and calling them chefs. And that works beautifully, right up until the plate comes back wrong and nobody at the table knows what to fix.

#Artwell #FieldNotesFromTheBuild #AIStrategy #SoloFounder

Share

Subscribe now

Artwell

For the last year, most of what I’ve written has been about how companies are struggling with narrative. Meta cloning Zuckerberg into an AI nobody can shut off. Microsoft bolting Copilot onto its products, and then bolting a narrative onto Copilot to match. Big tech thought leadership that says nothing in three thousand words. AI rollouts that stall because the engineering shipped but the story didn’t.

One pattern keeps surfacing inside all of it: companies look at consumer brand giants, pick up the wrong lessons, and apply them in places those lessons don’t fit. The case study gets framed. The headline gets repeated. The slogan version of the lesson spreads. And the underlying reality of what the company actually built gets lost in the retelling.

The Golden State Warriors are one of those examples right now.

In 2022, the team launched Golden State Entertainment – a content company that produces documentaries, signs musical artists, and stages live events. Three years in, GSE has released an album featuring E-40, Too $hort, Saweetie, and Larry June; been part of an Emmy-winning short film about Jeremy Lin; and produced documentaries about figures like Mahmoud Abdul-Rauf. It gets cited often in marketing circles as a storytelling success story – proof that companies should invest in narrative infrastructure, build community, treat content as a strategic priority.

Some of that is true. But it misses what GSE actually is.

Listen to how David Kelly, the franchise’s Chief Business Officer who runs GSE, describes the venture in his own words:

“We want to leverage the team’s brand outside of sports, telling the stories of people that shape culture and move culture.”

“We want to tell some Warriors stories... but we’re not limiting ourselves to Warriors stories.”

“This is an opportunity for musicians, entertainers, creatives, and brands to partner with the Golden State Warriors to bring their projects to life.”

That’s not a sports franchise telling its own story better. That’s a media company that happens to be owned by a basketball team, using the Warriors brand as gravitational pull to attract creators who want to be associated with it. The stories being told are mostly not Warriors stories. The artists being signed are not Warriors artists. The documentaries are not Warriors documentaries. The team’s brand is the platform; other people’s work is what runs on it.

Imagine if Adobe spun up a record label, signed Bay Area musicians, and produced a slate of indie documentaries. Nobody would frame it as a marketing lesson for the SaaS industry. We’d say: Adobe is in the media business now. We’d be curious about the strategy. But we wouldn’t pretend it teaches the average B2B software company how to tell its own story better.

That’s the move GSE represents. And it’s a move that requires something most companies don’t have: cultural gravity. A brand that creators actively want to be attached to. Decades of accumulated equity. A built-in venue, audience, and pipeline. The Warriors aren’t telling stories better than everyone else. They built a media platform that lets other people tell stories under their flag, because the flag has enough pull to make it work.

This is the part that matters for most companies looking at GSE and trying to extract a storytelling lesson: they’re looking at the wrong category. The lesson isn’t “build narrative infrastructure” or “treat content as strategy.” It’s “decide whether you have enough cultural pull to be a platform – and if you don’t, which is most companies, your job is still the harder one.” Which is telling your own story, well, about your own work, to people who need to understand it. No record label is going to save you from that.

Which brings me back to the work I’ve been writing about all year. The companies struggling with narrative right now aren’t struggling because they need a record label. They’re struggling because they treat their own story as a deliverable instead of a strategy – something produced after the engineering is done, by the team three floors down. The fix isn’t to mimic the Warriors. The fix is to do the thing the Warriors are also doing, just at a different scale: name storytelling as central to what the company is, and resource it like you mean it.

Build a media company if you want one. Just don’t confuse it with the work of telling your own story.

#Artwell #AIStrategy #SoloFounder

Share

Subscribe now

Artwell

Just like Jason Voorhees, the agile-versus-waterfall debate never dies – it keeps coming back. It comes back because it’s a useful argument, and because every new wave of developers eventually has to reach their own conclusions about how much planning is enough and how much is too much. The current resurgence has a particular flavor – it lives mostly on LinkedIn and YouTube, and it leans hard against specs. Specs are heavy. Specs are corporate. Nobody reads them. Just ship. Waterfall is dead. The dinosaurs of the old way are still scolding us from the boardroom while the rest of us iterate.

There is something true in that critique, and there is something the critique is missing. Both halves are worth taking seriously, because building a real product in 2026 means navigating both at once.

The part that’s true is straightforward. The dead corporate version of a spec – the thousand-page document nobody reads, the approval chains, the false certainty, the planning that calcifies into fiction – really is a waste. That’s not the same thing as saying specs are a waste. It’s saying that one particular implementation of a spec is a waste, and confusing the implementation with the concept is the kind of category error that lets people argue against object-oriented programming because they once saw a bad Java codebase. The bad version of anything is always bad. The interesting question is whether the good version still has work to do.

The bigger problem with the current debate is that it’s being conducted in 2026 with the same binary it had in 2014, which in AI years is roughly the Cretaceous. The debate certainly is not extinct, but the framing of it is. Both sides of the argument still assume that a spec is a document a human writes for another human to read. Even when the spec gets handed to an AI rather than a human, the assumption underneath stays the same – the spec is a static artifact passed across a boundary, used as a one-time briefing rather than as a living substrate the work persists within. From that premise, the entire shape of the debate makes sense. Specs are heavy because humans are slow readers. Specs feel like waste because the people they’re written for never finish them. Defending specs makes sense from inside that frame. So does dismissing them. The two sides are both correct, and they’re both arguing about the wrong thing.

Before getting to what I actually think the debate is missing, there’s a useful principle from the older argument worth keeping, because it dissolves the binary on its own. Some decisions in front of you are expensive to reverse. Some are cheap. Architecture is expensive. A button label is cheap. You don’t iterate your way into a data model; you iterate from one. The cost of being wrong on the foundation is high, and the cost of thinking carefully about it up front is low. That phase rewards real specs. The wiring phase, where you’re shaping what the user sees and how a flow feels and where the copy lands, is different. Those things only reveal themselves once people start using them, and heavy specs there would actively hurt the work. The principle is simple: match the process weight to the reversibility of the decision. Spec the expensive, iterate the cheap.

Most serious builders already work some version of that, whether they call it that or not. The agile-versus-waterfall debate falls apart the moment you accept that different decisions in the same project deserve different amounts of upfront thinking. But the principle still isn’t the deepest thing the debate is missing. The deepest thing is that the audience for the spec has changed, and nobody on either side of the public argument seems to have noticed.

I’m building Artwell, an AI narrative and voice platform, as a solo founder. The foundational specification I worked through with multiple AI models, across many days of conversation, came out around 80 pages. The user-experience specification that followed is roughly 20 pages. I have never read either of them cover to cover, and I never will. Neither would anyone else. That isn’t the failure of those specs. That is the point of those specs.

When you build a real product with AI in the loop – through model versions that change, across context windows that overflow and crash, across sessions you start fresh tomorrow because the current one is exhausted, across multiple models running in different roles – you need somewhere for the work to live that isn’t just your head and isn’t just one fragile session. The spec is that somewhere. A new session, a new model, a new tool reads the document in seconds and comes up to speed on weeks of accumulated decisions, definitions, constraints, edge cases, and intentions. Without that document, every restart is a regression and every model swap is a tax. With it, the work persists across all of it.

And the conversation that produces the spec is itself iterative, which is the part that breaks the frame entirely. The 80-page foundational spec wasn’t a waterfall artifact handed down from a planning phase. It was the consolidated output of multi-day back-and-forths across multiple sessions, multiple models, and multiple roles – a planner session proposing, a reviewer session pushing back, an implementer session surfacing a hole the planner missed. Features I would never have discovered through pure vibe-coding came out of the planning conversation itself, because that conversation had the time and the surface area to let edge cases and competitor analysis and full workflows breathe. The output happens to be a long document, but the document is the consolidated record of an extremely agile process. The thing the public debate calls “spec writing” and the thing it calls “iteration” are happening in the same place. They just produce different artifacts at different points.

That changes the cost calculation entirely. A 50-page spec that humans have to read end-to-end before they can start working is corporate theater, and the critics are right to dismiss it. A 50-page spec that an AI ingests in seconds and that survives the next model swap is infrastructure, and dismissing it is a category error. Same document. Completely different function. Whether the work of writing it is wasted depends entirely on who’s reading it and what they’re using it for.

And the spec itself doesn’t stop evolving once implementation begins. We treat it as the working substrate of the project – not just as a planning artifact but as the implementation document itself, broken into phases, lanes, and slices that get worked through over the following weeks and months. As that work surfaces things the planning didn’t anticipate, the spec gets updated. Sometimes we revise the original document directly. Sometimes we preserve the original intact for the historical record and write addendums that capture founder-approved course corrections. Either way, the spec stays current with the actual state of the project, which means new sessions and new models always get the truth of where we are – not a snapshot of where we thought we’d be six weeks ago. The flexibility to change direction during implementation doesn’t break the foundation. It’s built on top of it, which is the entire reason the foundation was worth specifying in the first place.

This isn’t new thinking outside of software. Many years in media production taught me that the work you put into pre-production is what makes production possible. An indie filmmaker can shoot from the hip with no script, no shot list, no production schedule, and pull off something charming – the kind of thing that wins a festival short slot and becomes a calling card. The same filmmaker cannot pull off a feature that way. They certainly cannot pull off a series. The amount of structure required scales with the ambition of the thing being built. The script that nobody reads cover-to-cover on set is doing real work as a shared reference. The shot list the DP scribbles all over is doing real work as a daily handoff. The shooting schedule that gets revised every morning is doing real work as a coordination layer. None of it is waterfall, and none of it is wasted. It’s the substrate that makes the days on set possible.

Specs work the same way. A weekend script doesn’t need one. A product with customers, retention policies, billing, trademarks, compliance posture, and a brand audience whose trust breaks if anything breaks does. What changed isn’t whether the substrate matters. What changed is who reads it.

There’s a related point that follows from this, and it matters because it’s where the public debate quietly does the most damage. When people complain that AI claimed it finished a task it didn’t actually finish, or that the AI skipped a section of the spec, or that the AI ignored a constraint, the default response is to blame the AI. The more honest diagnosis is that the AI is doing exactly what its scaffolding allows it to do. A single prompt-and-pray session with no checkpoint, no completion verification, no review panel, and no orchestration layer is going to produce inconsistent results regardless of which model is on the other end. That isn’t an AI failure. That’s a builder choosing not to construct the orchestration that the work requires.

The builders who are shipping serious products with AI assistance are running multi-session, multi-model, multi-role workflows with forced checkpoints, completion verification, and review passes. The builders who are shipping weekend tools and clever demos can get away with one session and a good prompt. Both are doing AI-assisted development. They are not doing the same job. The advice that works for the second does not port to the first, and treating the two as equivalent is the original sin underneath most of the loudest takes on agile, specs, and how to build with AI today.

So the real debate isn’t agile versus waterfall, and it isn’t even spec versus no-spec. It’s whether you’re building something that requires a real orchestration architecture and a persistent record of its own thinking, or whether you’re building something that doesn’t. It’s whether the documents you’re producing are written for human readers who’ll never finish them, or for the AI collaborators who actually carry the work across the sessions you can’t.

I don’t read my own specs. They aren’t for me. They’re for the next session, the next model, the next implementer, the next reviewer, the next time the context window blows up and tomorrow’s version of the project has to be reconstituted from yesterday’s decisions. The specs are how the project remembers itself. They are the foundation that makes everything else I do possible, even though they look, on the surface, like exactly the kind of artifact the current debate has decided to dismiss.

That isn’t waterfall coming back. It isn’t the old discipline in new clothes. It’s a new discipline that the public debate hasn’t built the vocabulary for yet – and one that the people still arguing about specs versus no-specs are going to need, whether they decide to learn it or not.

#Artwell #FieldNotesFromTheBuild #AIStrategy #SoloFounder

Share

Subscribe now

Artwell

Nine months ago, when I began building Artwell in earnest, Deepgram was the right call for speech-to-text. Best quality at the time. Best pricing. Solid intelligence layer.

I also flagged, in the first week, that Artwell should not be permanently coupled to a single transcription provider. That was not some brilliant revelation. Every senior architect would have seen the same thing.

But there is a difference between seeing an architectural need and building it immediately. Early in the build, the more important job was learning the shape of the product: the workflow, the transcript model, the editing layer, the downstream story outputs, the places where quality actually mattered. Premature abstraction is its own kind of debt. So we kept moving.

This week I went back and re-tested the speech-to-text market against Artwell’s real needs – long-form audio, diarization, speaker handling, intelligence layers, downstream editing, cost at scale. The landscape had not moved by inches. It had moved by miles.

ElevenLabs’ latest Scribe v2 has overtaken on quality and pricing. AssemblyAI is right behind it, with a comparable intelligence layer – sentiment, topics, summaries, speaker identification inferred from transcript context. For the configuration Artwell needs, the incumbent provider now costs more than double what the leaders charge, with measurably higher word error rates (WER) and a thinner intelligence layer. This is no small thing. This changes margins, pricing flexibility, and what a product can afford to offer users.

With Artwell’s launch on the horizon, this became the moment to build the abstraction layer.

Artwell now has a transcription layer that can route across providers, support fallbacks, separate development from production, and absorb the next provider change without touching the product itself. The intelligence layer is next, same pattern – OpenAI and Anthropic today, whatever proves best tomorrow.

The lesson though is not “build abstraction layers.” The lesson is that AI infrastructure decisions have a shorter shelf life than most teams are conditioned for. The right provider today may not be the right provider one year from now. It might not even be the right provider one month from now. That’s simply the speed of the market we are all building in.

“Which provider is best” is the question every team asks. The better question to ask is what it will cost you to switch when that answer changes.

Architecture buys you the right to change your mind.

And in AI, that optionality may be the difference between building a moat and building a wrapper.

𝘛𝘩𝘢𝘵 𝘪𝘴 𝘸𝘩𝘢𝘵 𝘈𝘐 𝘴𝘵𝘳𝘢𝘵𝘦𝘨𝘺 𝘭𝘰𝘰𝘬𝘴 𝘭𝘪𝘬𝘦 𝘧𝘳𝘰𝘮 𝘛𝘩𝘦 𝘉𝘶𝘪𝘭𝘥.

#Artwell #FieldNotesFromTheBuild #AIStrategy #SoloFounder

Share

Subscribe now

Artwell

Two stories get cited every time someone recites the lean MVP gospel: WhatsApp and Craigslist. Both are real. Neither is actually about the product.

WhatsApp launched in 2009 with a Russian contractor hired off RentACoder, and the first version crashed so often that founder Jan Koum almost gave up and went job hunting. What saved it wasn’t iteration on user feedback – it was Apple later shipping push notifications, and Koum knowing exactly what to do with them. The founders weren’t kids racing to ship. They were two former Yahoo engineers in their forties with twenty years of combined experience, building in a market where “free international messaging” was a structural advantage against expensive SMS. By the time Facebook bought it for $19 billion in 2014, the company still had only 56 employees. The product was thin. The founders weren’t.

Craigslist is even less of a “ship fast” story. Craig Newmark was a 40-something engineer with seventeen years at IBM when he started emailing event listings to friends in 1995. The site didn’t become a company until 1999. He treated it as a side project for four years while his day job was contract programming. The competition was newspaper classifieds. There was no saturation, no race, no MVP gospel – just an experienced engineer letting a network effect build quietly because he understood, structurally, what he was actually building.

Strip the retrofitted “vanilla MVP, iterate fast” narrative away and the same pattern surfaces in both stories. Experienced operators. Structural market reads. Tools so primitive that “minimum” wasn’t a strategy, it was the only option. The product wasn’t the differentiator. The founder was. We just didn’t notice because the tools obscured it.

In 2026, the tools have commoditized, and the founder is the signal.

Recently I spent four days writing and rewriting a 20-page UX architecture document for Artwell – workflow definitions, workspace logic, interaction rules, edge cases, language, and intent. Then I handed it to Claude’s new design tool. Within an hour, I had beautiful, meaningful, production-quality interface templates. But the real story isn’t that one hour – it’s those four days. The tool turned judgment into pixels at speed. Without the judgment, it would have produced beautifully designed nothing.

That’s the shift the MVP discourse hasn’t caught up to yet. When tooling was primitive, the product was the bottleneck, so the product was the variable. You shipped vanilla because you couldn’t ship anything else, and you iterated because the market was unsaturated enough to give you the runway. In 2026, none of that is true. Tools are commoditized. Markets are saturated. Any serious founder can ship something competent in a fraction of the time it used to take. So what’s left to differentiate?

The founder. That’s it. That’s the only variable that doesn’t commoditize.

Which is why some of the loudest MVP advice online feels increasingly incomplete. It isn’t just vibe coding anymore. It’s vibe architecting, vibe marketing, vibe founding. AI is being used to compensate for judgment that hasn’t been built yet, and the output reflects exactly that. Beautifully designed nothing, shipped fast, drowned in a saturated market full of other beautifully designed nothings. The amplifier did its job. The signal underneath was the problem.

The MVP framework wasn’t wrong; it was a product of its constraints, and those constraints have changed. The new minimum bar isn’t a feature set. It’s a founder. A minimum viable founder – someone with enough judgment, taste, market read, and conviction to know what to build before they touch a tool, and to know why it matters before they ask anyone else to care.

If you’re building right now, that’s the work. Not the stack. Not the speed. Not the iteration cadence. Invest in your read of the market. Invest in your editorial taste. Invest in your ability to write a comprehensive UX document that tells a tool exactly what to make and why. The product will follow. Tools are too good now for it not to.

The founder is the differentiator. The tools didn’t change that. They just made it impossible to miss.

#Artwell #AIStrategy #SoloFounder

Share

Subscribe now

Artwell

A few months ago I wrote that American pride starts with remembering who we are – and that a country which stops doing hard things eventually forgets it ever could.

Today, Artemis II came home.

Four astronauts, a lunar flyby, and a splashdown that looked a lot like the photos my parents watched live on television more than fifty years ago. Except this wasn't a memory. This was hours ago.

I have been thinking all evening about what it felt like to watch it. Not the engineering, though the engineering is staggering. Not the politics, because this moment belongs to something larger than that. What I felt was the thing I wrote about and was not sure I would feel again in my lifetime: the quiet, uncomplicated pride of watching my country do something unambiguously worth being proud of.

Pride is not denial. It is commitment. It is the willingness to believe that building impossible things together is still who we are, and then going and proving it.

Today we proved it. Welcome home, Artemis II. And thank you for reminding a lot of us how that's supposed to feel.

#Artwell #AIStrategy #SoloFounder #ArtemisII

Share

Subscribe now

Artwell

Breakthrough: Scientists have successfully replicated Schrödinger's quantum superposition using only venture capital and high confidence.

#SchrödiconValley #QuantumUnicorn #TechSatire

Share

Subscribe now

Artwell

When I was a kid I had a black-and-white photo of Bruce Lee pinned to my bedroom mirror. I was fat. I got picked on. And every morning I woke up and looked at a man who the world underestimated until they couldn’t anymore.

I didn’t care about the kicks. I cared about the philosophy.

Bruce Lee studied every martial art he could find, and then he threw them all away. Not because they were worthless. Because they had become more important than the fight itself. The forms were beautiful. The katas were precise. And they would get you killed, because they trained you to follow a pattern instead of responding to what was actually in front of you.

But here’s what most people leave out of that story. He didn’t walk away from nothing. He mastered Wing Chun under Ip Man. He studied Western boxing, fencing, judo, wrestling, grappling, taekwondo. He didn’t reject discipline – he completed it. Jeet Kune Do wasn’t the absence of form. It was what came after mastery of form. The formless form. And that distinction matters, because anyone can tear up a playbook. Only someone who has mastered the playbook earns the right to replace it with something better.

He replaced all of it with one principle. He inscribed it on the Jeet Kune Do symbol in Chinese characters:

Using no way as way. Having no limitation as limitation.

The whole world remembers “be like water.” Almost nobody remembers what he actually meant. He wasn’t talking about style. Everybody gets that wrong. He was talking about form. Rigid form. He was warning against it. The kind of form that looks powerful in a controlled environment and shatters the second reality shows up unannounced.

I’ve been thinking about this a lot lately. Because the AI industry is building nothing but form.

The Industry Is Performing Kata

Every major company deploying AI right now is performing kata.

Choreographed sequences. Executed the same way every time. There’s a transformation roadmap. A change management framework. An all-hands where a senior leader says the words “more productive.” A polished internal memo. An external campaign.

Beautiful choreography.

Adoption goes up 13%. Trust goes down 18%. 77% of employees say AI increased their workload. 56% of companies report zero financial return. 59% of hiring managers admit they dress layoff announcements in AI language because it plays better than telling the truth.

I Watched This Happen From the Inside

I spent almost twenty years inside Microsoft. I watched this happen in real time. Not once. Over and over. The playbook kept running even after reality had moved on. The structure was still standing, the choreography was still being performed, but the truth had left the building and nobody noticed because the form was still intact.

Bruce Lee saw this exact pattern in every traditional martial art that had stopped questioning itself:

“Set patterns, incapable of adaptability, of pliability, only offer a better cage. Truth is outside of all patterns.”

– Bruce Lee

The cage is the playbook. The cage is the framework. The cage is the bolted-on narrative that tells employees AI will empower them while their colleagues disappear.

The Imitation Problem

There’s a quote from the Tao of Jeet Kune Do that I forgot for thirty years and remembered tonight.

“We have more faith in what we imitate than in what we originate. We cannot derive a sense of absolute certitude from anything which has its roots in us.”

– Bruce Lee, Tao of Jeet Kune Do

Read that again in the context of how companies are deploying AI.

Every enterprise is imitating every other enterprise. The same playbooks. The same transformation language. The same announcements. The same Copilot rollout strategy. Nobody is originating anything because originating requires trusting your own judgment about what your people actually need. And that’s terrifying when the whole industry is moving at once and nobody wants to be the one who did it differently.

So they imitate. They bolt on what worked somewhere else. And they wonder why it doesn’t work here.

“If you follow the classical pattern, you are understanding the routine, the tradition, the shadow – you are not understanding yourself.”

– Bruce Lee

That’s the AI industry’s problem in one sentence. Companies are following the classical pattern. They understand the routine. They do not understand themselves. They do not understand their own people, their own culture, their own customers, the actual lived reality of the humans who are supposed to benefit from all of this.

The Audience Has Already Left

Half of Americans say AI in daily life makes them more concerned than excited. 65% of workers barely use it. The people building AI narratives live in a bubble where everyone uses AI every day, and they assume the rest of the world is right there with them.

It isn’t.

Coca-Cola’s AI holiday ads got rejected as soulless. McDonald’s pulled their AI Christmas commercial after audiences called it slop. iHeartMedia launched a “Guaranteed Human” campaign because 9 in 10 consumers want their media made by actual people. Merriam-Webster named “slop” their word of the year.

The forms keep performing. The audience has already left.

What I’d Say If I Could Be Honest

If I could sit across from every CEO pushing AI transformation and say one honest thing, here’s what I’d say.

Your version of AI is crushing something deeply human. It’s crushing endeavor. It’s crushing dignity in work. People feel it even when they can’t articulate it, and your internal communications aren’t fooling anyone.

But here’s where most critics get it wrong, and I want to be precise.

AI is not the problem.

I use AI every waking hour. It replaced Google for me. It replaced how-to videos. It replaced a hundred small friction points. It is one of the most powerful tools I have ever used. I am building an entire company on it.

The problem is how AI has been framed, marketed, and introduced to the world. We have done a terrible job helping people understand where AI helps, where it harms, and where the line should be. And we have done an even worse job listening to what people actually need before deciding what to build and how to sell it.

Not better AI. Better honesty about AI.

Hack Away at the Inessentials

Bruce Lee didn’t build Jeet Kune Do by learning more techniques. He said it himself:

“It is not a daily increase, but a daily decrease. Hack away at the inessentials.”

– Bruce Lee

The AI industry needs to hack away at the inessentials. The theater. The hype. The fake wisdom. The obsession with scale for the sake of scale. The idea that making humans smaller somehow counts as progress.

The version of AI nobody is building toward is the one that quietly makes people stronger. Smarter.Clearer. More capable. Not more dependent. Not more passive. Not more replaceable.

Something that serves the human without swallowing the human. Powerful, but disciplined. Something that helps you move better, not disappear faster.

A formless form. Not chaos. Readiness. The ability to respond to what’s real without being locked into a pattern designed for a different fight.

The Mirror

When I was thirteen I pinned Bruce Lee to my mirror because I wanted to become something better than what I was. I was a fat kid who got picked on, and he represented the possibility that the person everyone underestimates can be the one who sees most clearly.

So I did something about it. I started training in Shotokan karate. I became a Varsity wrestler. Not because I wanted to fight anybody. Because I wanted to stop being afraid. And it worked. Not overnight, not dramatically, but in the way real things work – slowly, through repetition, through showing up when I didn’t feel like it. I lost the weight. I found confidence I didn’t know I had. I learned that discipline wasn’t something you performed. It was something you practiced until it became part of you.

That kid went on to spend twenty years at Microsoft and build a content program from nothing into something that generated $59 million in revenue. I mastered the playbook. And then I earned the right to throw it away and build something of my own. And I forgot, for a very long time, where all of it started.

I’m fifty-three now. I forgot that photo. I forgot the philosophy. I let decades of corporate life and monthly bills and other people’s priorities bury something that used to matter to me more than anything.

I’m not forgetting it again.

If I could pin something to my mirror today it would say this:

You built today. You earned tomorrow.

That’s not a Bruce Lee quote. That’s mine. But I think he’d understand it. Because it’s about the same thing he spent his whole life teaching – that truth is earned through honest work, not inherited through rigid form. And that the person willing to throw away the playbook and start from what’s real will always see further than the ones still performing kata in an empty room.

Be like water.

Bruce Bracken is the founder of Artwell.ai and former Head of Podcasts & Digital Experiences at Microsoft, where he built a content ecosystem that generated $59M in attributed Azure revenue. He writes about the gap between what AI companies say and what humans actually experience.

Share

Subscribe now

Artwell

In December, I published an article making the case that Microsoft’s real AI problem was integration, not capability. Copilot wasn’t failing because the intelligence wasn’t good enough. It was failing because AI showed up in their products as something added rather than something assumed. A sidebar. A button. A helper. Never the organizing principle of the experience itself. I called it Artificial Integration.

Three months later, Microsoft announced Copilot Cowork, integrating Anthropic’s Claude Cowork technology into Microsoft 365 Copilot. The move made strategic sense. But it also reinforced the same pattern: even when a company brings in genuinely powerful AI capability, the result still feels bolted on. The platform absorbs the innovation without reorganizing around it.

But this article isn’t about products. It’s about something companies are getting even more wrong, and almost nobody is talking about it.

The Same Mistake, One Level Up

Every major technology company is currently telling two stories about AI. Internally, they’re telling employees that AI will make them more productive and more capable. Externally, they’re telling consumers that AI makes everything better. Both stories share the same structural flaw: they’re bolted on.

Companies are bolting AI narratives onto their organizations the same way Microsoft bolted Copilot onto Office. The narrative shows up as a memo, a talking point, an all-hands presentation, an ad campaign. Always disconnected from where people actually live. The story exists in a completely different psychological space than the lived experience. And people can feel the gap immediately.

This is an architecture problem, not a communications problem. And it’s the same one I wrote about in December, just one level up from the product.

The Internal Narrative Is Broken

Companies are throwing AI narratives at the wall like spaghetti, hoping something sticks. A productivity story here. A transformation memo there. An empowerment talking point in the all-hands. None of it connects to what employees actually experience when they sit down at their desk, and the data makes the disconnect impossible to ignore.

ManpowerGroup’s 2026 Global Talent Barometer found that while regular AI usage jumped 13% in 2025, confidence in the technology plummeted 18%. Adoption going up while trust goes down is a signal that the story around the technology is failing, not the technology itself.

Great Place to Work found a staggering gap between how executives and frontline employees perceive AI communication. Executives overwhelmingly believe their messaging is landing. Frontline workers largely disagree. It’s one of the widest perception divides in recent workplace research.

Upwork found that 96% of C-suite leaders expect AI to boost productivity, while 77% of employees say the tools have actually increased their workload. Leadership is telling a productivity story. Employees are living the opposite of it.

PwC’s 29th Global CEO Survey found that 56% of companies have seen zero financial return from AI investments. Zero. Only 12% reported both revenue gains and cost reductions. Hundreds of billions spent, and most companies have nothing to show for it.

And then there’s the layoff story. A December 2025 Resume.org survey found that 59% of hiring managers admit they emphasize AI in layoff announcements because it “plays better with stakeholders” than admitting financial constraints. Oxford Internet Institute researchers have documented this directly, noting that companies are scapegoating AI to disguise conventional cost-cutting.

None of this is what a real narrative strategy looks like. Strategists don’t throw spaghetti. They research their audience, they listen, they build plans around what people actually need to hear and experience. What the data shows is that most companies don’t have anyone doing that work. They have messaging without strategy, communications without architecture, and stories that no one on the receiving end believes.

The 2026 Edelman Trust Barometer confirms it. When asked who they trust to tell the truth about AI, workers chose their peers over CEOs by a factor of two. Two-thirds of AI distrusters said adoption feels forced, not voluntary. The people sending the bolt-on narratives are the people employees trust least.

The Consumer Narrative Is Broken Too

The internal story is failing. The external one might be worse.

Pew Research Center found that half of U.S. adults say AI in daily life makes them more concerned than excited, up from 37% in 2021. Only 10% are more excited than concerned. And 65% of American workers still say they don’t use AI much or at all. The people building AI narratives are talking to the 35%. The other 65% are watching from the outside, increasingly anxious about something they barely interact with.

The people building these narratives are geographically and professionally concentrated in exactly the places you’d expect. AI users skew urban, skew educated, skew toward banking, finance, and technology. The teams creating AI advertising campaigns and consumer messaging live inside a bubble where everyone uses AI every day. They assume the rest of the world is right there with them. It isn’t.

This is why we keep seeing the same failures repeat. Coca-Cola’s AI-generated holiday campaigns, two years running, were rejected as soulless. McDonald’s Netherlands pulled its AI Christmas ad after audiences called it “AI slop” that “ruined my Christmas spirit.” Researchers call this the “authenticity premium,” a measurable trust penalty that activates the moment consumers sense a machine is behind the message. A study from the Nuremberg Institute for Market Decisions found that simply labeling an ad as AI-generated makes people perceive it as less natural and less useful.

Here’s the part that should make every marketer uncomfortable. The McDonald’s AI ad required seven weeks of production with a team of AI specialists doing extensive refinement on each shot. They didn’t save time. They didn’t save money. They just lost the human element. And observers have pointed out that AI companies are now caught telling two contradictory stories at once: telling the public that AI content requires just as much human effort as traditional production, while telling investors it eliminates the need for human labor.

Compare this to what great technology storytelling actually looks like. Apple’s “Misunderstood” ad from 2013. The teenager who appears antisocial at Christmas, glued to his phone, until he reveals he’s been making a beautiful family video the whole time. The ad didn’t mention a single product feature. The technology was invisible. The story was about a family, and the product was just the tool that made the moment possible. That ad destroyed people emotionally because it started with the human experience, not the technology. The iPhone was the least important thing in the frame. The fact that one of the most recognized examples of native technology storytelling is over a decade old tells you something about how rare this actually is.

The Backlash Is Already Here

The market is responding to the broken narrative in ways that should concern anyone building an AI strategy. iHeartMedia launched a “Guaranteed Human” campaign built on research showing that 9 in 10 consumers want their media created by real people, even consumers who use AI tools themselves. iHeartMedia CEO Bob Pittman put it bluntly: consumers aren’t just looking for convenience. They’re searching for meaning.

In Hollywood, the credits of the Apple TV hit “Pluribus” from Vince Gilligan now read: “This show was made by humans.” Across New York City, subway ads for the AI wearable “Friend” have been vandalized with messages like “AI is not your friend” and “talk to a neighbor.” An artist created a browser extension called Slop Evader that filters search results to include only content from before ChatGPT’s release.

Merriam-Webster chose “slop” as their 2025 word of the year.

Gartner found that 53% of consumers distrust AI-powered search results, and 61% want the ability to toggle AI on or off. People don’t want AI removed from the world. They want control over when and how it enters their experience. That distinction matters enormously, and bolt-on narrative ignores it entirely.

The Pattern

Three months ago, I argued that Microsoft’s AI was smart enough but not native enough. That diagnosis applies far beyond products now.

Companies are treating AI adoption as a deployment problem. Deploy the technology, then communicate about it. Ship the product, then explain it. The story always arrives after the fact, as an afterthought, disconnected from where people actually experience the change.

The companies that win the next decade will be the ones that understand something the current generation of corporate leadership mostly doesn’t. Adoption is a narrative design problem, and the narrative has to be built into the experience from the beginning, not applied afterward like wallpaper over drywall.

Most organizations don’t have anyone thinking about this at a systems level. The communications function in most companies was designed to announce decisions, not to shape them. Change management still runs on playbooks written before generative AI existed. And marketing departments are producing AI-themed campaigns from inside a bubble that looks nothing like the world their customers actually inhabit. The role that should be doing this work hasn’t been built yet in most organizations.

The gap between what companies say about AI and what people actually experience keeps widening. And almost nobody in corporate leadership recognizes that the story is the adoption mechanism. Not decoration around it. Not a follow-up to the deployment plan. The mechanism itself.

I diagnosed the product version of this problem in December. The narrative version is next.

Bruce Bracken is the founder of Artwell.ai and former Head of Podcasts & Digital Experiences at Microsoft, where he built a content ecosystem that generated $59M in attributed Azure revenue. He writes about the gap between what AI companies say and what humans actually experience.

Share

Subscribe now

Artwell

The best writing rooms don’t happen in writing rooms. They happen at a bar at 11pm when five opinionated writers are three drinks in, arguing over the same draft. One catches the plot hole. One catches the bad dialogue. One catches the joke that’s technically funny but kills the pacing. Nobody gets the whole picture alone. The final script is better than any single person at the table could have produced.

That’s also the best description I’ve found for how to actually use AI well – except that’s not how most people use it. Most people are sitting at that bar alone, scribbling “make something good” on a cocktail napkin, sliding it to the bartender, and leaving a one-star review when they don’t get the drink they wanted.

I spend more time than I should on Reddit watching people declare AI models are broken. “GPT deleted my entire codebase.” “Claude refused to do what I asked.” “Gemini hallucinated a lawsuit that doesn’t exist.” And every time, buried about four comments deep, someone asks the question nobody wants to answer: “What was your prompt?”

*crickets*

There’s a phrase people love to say when they want to sound pragmatic about AI: garbage in, garbage out. It’s old. It’s unsexy. It’s also the closest thing we have to a law of physics in this space. The part that keeps surprising me is how many “AI power users” repeat that phrase while continuing to feed models prompts that look like they were typed with one thumb at a red light.

And I’m not talking about casual use. I’m talking about high-stakes work – code review gates, implementation checkpoints, security-sensitive pathways, decisions that either prevent damage… or politely approve it. Every vague prompt that produces bad code becomes tech debt you won’t discover until it bites you in production six months from now. Every contradictory instruction that forces a rerun – and then a re-rerun – burns tokens, burns hours, and burns the subscription you’re paying for whether the output is useful or not.

Over the last two weeks, I took on a full foundation refactor of Artwell, the Story Engine platform I’m building. Fourteen phases, dozens of implementation slices, fourteen days straight – the kind of rebuild where you either get the architecture right or you get to do it again in six months. Because the stakes were high and the surface area was enormous, I decided to run three frontier models as a review panel across the entire process, from the initial planning review through every meaningful checkpoint along the way. Not a vibe check. Not cherry-picked screenshots. A repeated, evidence-backed panel as the work moved from plan to implementation to remediation. Dozens of verdicts. Reruns when blockers were disputed. Retrospective tracking when reality proved who was right and who was merely confident.

What fell out of that process wasn’t just a ranking. It was a sharper view of what most “model evaluations” actually are: prompt evaluations in disguise. When the instructions were coherent, the panel looked brilliant. When the instructions contradicted themselves, even the best models began behaving like perfectly obedient interns following a bad spec – faithfully, catastrophically, and without apology.

One of the cleanest examples was also one of the most banal: a prompt pack that required mandatory reads while simultaneously enforcing contamination rules that prohibited reading them. The model didn’t hallucinate. It didn’t get confused. It did exactly what it was told to do – and then hard-stopped on a procedural blocker because the human who wrote the instructions created an impossible situation.

That should feel familiar to anyone who’s ever handed someone two conflicting priorities and then acted surprised when they froze. Tell your direct report “move fast” and “don’t break anything” without defining the boundary, and watch what happens. The only difference with AI is that the failure happens in language first – which makes it easier to dismiss as “model weirdness” instead of what it actually is: governance debt.

The Other Kind of Failure

There’s also a second prompt failure mode that’s almost the opposite of contradiction, and it’s sneakier because it produces beautiful green dashboards.

Prompts that are too strict. Too narrow. Too “paint-by-numbers.”

When the guardrails are a millimeter apart, models will stay perfectly between them, report everything is green, and never once mention the cliff right next to the track. If your prompt only asks, “Does the existing checklist pass,” the model can honestly answer “yes” while completely missing that the checklist itself failed to ask the right questions. That’s not the model being dumb – that’s the model being obedient.

I’ve seen this happen in a particularly irritating way: you can almost watch the model notice something odd during its reasoning – a strange invariant, a suspicious edge path, a “wait, why is that allowed?” moment – and then omit it from the final report because the prompt never explicitly gave permission to include “extra” observations. The model isn’t hiding it. It’s following your instructions like a watchdog you accidentally trained to wag at intruders.

And the missed defect doesn’t vanish. It ships. It hides in your codebase, passing every check you built, until some future Tuesday when it surfaces as a production incident and you’re suddenly paying emergency rates to fix something a better prompt would have caught for free.

So What’s the Right Balance?

Prompt quality isn’t just “be comprehensive.” It’s “aim for the right balance.”

Some reviews should be surgical. You want a model to prove a specific invariant and treat anything else as noise. That’s the right move when you’re validating a security promise, or locking a contract, or verifying a remediation. In those moments, freedom is the enemy and precision is the product.

Other reviews benefit from breathing room – permission to look outside the planned lane, to ask “what changed during implementation,” to flag weirdness even if it’s not on the checklist, and to challenge assumptions the planning stage might have missed. That’s how you catch the things that slip through, not because anyone was careless, but because real systems evolve in ways planning documents can’t perfectly predict.

And this is where the models come in – because each one failed in ways that illuminate a different kind of prompt problem.

The Review Panel

Across roughly thirty tracked review cycles – planning reviews, checkpoint gates, reruns, the whole messy parade – the overall effectiveness ranking came out as:

1. GPT-5.3-codex-xhigh

2. GPT-5.2-xhigh

3. Claude Opus-4.6

If you stop at the ranking, you’ll miss what matters, because the story is really about failure patterns and role-fit.

GPT-5.3 was the most reliable sentinel – fast, cross-stack realistic, and consistently first to catch the quiet defects that make a system appear to work until it doesn’t. Its weakness was the cleanest illustration of the prompt contradiction problem: give it conflicting instructions and it will hard-stop every time. You wrote an unsatisfiable contract and then blamed the interpreter.

GPT-5.2 is the deep defect hunter – the model that obsesses over control flow, schema validation, and security leakage at the level where things look fine from a distance and fail spectacularly up close. It produced some of the highest-value unique catches in the entire dataset, finding blocker-class bugs the other two models cleared without hesitation. I’ll be honest: I spent weeks hoping the data would let me drop this model from the panel just to speed things up. Its response times could run three to five times longer than the others. The data didn’t cooperate. The catches were too important.

Opus-4.6 is the strongest architecture and process synthesizer in the group – the one you want reading the whole packet and telling you whether the system makes sense as a system, not just whether a diff compiles. Where it fell short was first-pass blocker severity. It can be, in a very human way, too reasonable. It gives the benefit of the doubt. That’s a great trait in a colleague. It’s a risky trait in a gatekeeper. In several cycles, Claude said “Ready” while the other models caught real, later-validated bugs. The prompt gave it permission to pass, and it took it.

Every model had moments of clarity and brilliance. Every model had moments of complete failure. And the failures didn’t follow a simple pattern – the one that caught a critical security gap in one cycle missed a durability blocker in another.

Why the Answer Isn’t “Pick the Best One”

This is why the best approach wasn’t mono-model. It was a panel with roles.

The strongest configuration was GPT-5.3 and GPT-5.2 as the primary technical gate pair – speed and durability coverage on one side, adversarial depth on the other – with Claude as the coherence and verification seat that forces the work to make sense and confirms closure is real. Not just different prompts. Different brains. These are fundamentally different LLMs trained on different data, and part of the value is that they genuinely see different things – the way three writers at that bar will catch three different problems in the same draft because they read with different instincts.

Think of it as separation of concerns applied to judgment. In a writing room, the person who spots plot holes isn’t always the person who fixes pacing – and you don’t force them to be. Detection and validation are different cognitive tasks. You wouldn’t ask the same engineer to both find the bugs and certify the release. You shouldn’t ask the same model to do both, either.

And the prompt design for each seat was different. The sentinel seat got tight, specific instructions: prove this invariant, flag this class of defect, stop on this condition. The depth seat got wider scope: look for what the checklist missed, challenge assumptions, interrogate edge paths. The synthesis seat got a holistic mandate: does this system make sense end-to-end, and is closure real?

When I matched the prompt to the role and the role to the model’s strengths, the panel was remarkably effective. When I used the same prompt for everyone, I got noise from the fast reviewer, missed defects from the synthesizer, and procedural gridlock from the deep hunter. Same models. Different outcomes. The variable was the prompt.

The Part People Keep Skipping

And that brings me back to the part people keep skipping because it feels annoyingly “process-y”:

Writing a good prompt is not typing. It’s engineering.

It requires you to know what you actually want, resolve contradictions before you outsource reasoning, and decide whether you’re trying to validate a known checklist or invite the model to challenge the checklist itself. It forces you to define what counts as a blocker and what counts as follow-up. It exposes gaps in your own thinking because the model will walk exactly where you pointed – and if you pointed into fog, it will give you fog-shaped certainty right back.

The next year of AI won’t be won by the teams with the fanciest model access. Everyone has access now. The advantage is shifting to teams who build systems around models – workflows that treat prompting as an engineering artifact, panels that separate detection from validation, and prompt packs that don’t contain landmines or blinders.

In other words, teams who stop treating prompts like a text field and start treating them like a spec.

I started this process thinking I was evaluating models. Two weeks later, staring at thirty cycles of evidence, I realized the models had been evaluating me the whole time. Every false alarm traced back to a contradictory instruction I wrote. Every missed defect traced back to a prompt that wasn’t brave enough to ask the real question. Every time the panel disagreed, it was because I’d given three different brains permission to read the same instructions three different ways.

Garbage in, garbage out. Turns out it’s not just a law for machines.

Share

Subscribe now

Artwell

Gracie found the scent again this morning. Same patch of grass, same obsessive circling, same conviction that the deer or rabbit she smelled there three days ago might still be waiting. I’ve watched my golden retriever copilot do this a hundred times on our early morning walks – past signal, possible reward, investigate again – and it never changes. She trusts the pattern completely.

At the same time I was on my phone working through a strategic question out loud with ChatGPT. At one point, the model responded by pointing out that I was relying heavily on pattern matching to reach my conclusion.

That caught me a little off guard. Pattern matching, after all, is one of the things that frustrates me most about these AI models. It’s the engine behind the hallucinations, the confident wrong answers, the moments where a model completes a pattern with something plausible rather than something true and presents it like settled fact. It’s also, to be fair, the thing behind the responses that feel almost unreasonably good – the leaps that land precisely because the model connected dots you hadn’t seen yet. I’ve complained about the misses more times than I can count. And here was the AI telling me I was doing the very thing that frustrates me most about it.

But standing there on a dirt path watching Gracie investigate that same patch of grass for the third time, I had to sit with an uncomfortable recognition: I was doing the exact same thing. Drawing on prior conversations, previous outcomes, familiar structures. Connecting incomplete dots and filling gaps with probability-weighted guesses. The behavior I keep criticizing in machines was running quietly underneath my own thinking the entire time. That irony was not lost on me.

Here’s the thing, though – once you see it, you can’t unsee it. And I don’t just mean in that one moment on a walk. I mean everywhere. We assume tone from phrasing in an email and act on that assumption before we’ve verified it. We predict how someone will behave based on one previous interaction. We interpret silence as intent. We build entire strategic narratives from partial information and then defend them in meetings like we arrived at them through rigorous analysis. Most of the time we call this intuition or gut instinct. Sometimes, if we’re being honest, it’s just a really confident guess.

So why does it bother us so much when AI does the same thing?

I think it’s because the seams are showing. When a model makes a leap that feels slightly off, you can see it reaching – extending a pattern past where the data actually supports it. When we make the same kind of leap, it feels organic. Earned. Our patterning happens internally, wrapped in years of experience and a healthy layer of confidence. The model’s happens right out in the open where anyone can poke at it.

And honestly, that visibility is more useful than most people give it credit for.

I spend a lot of my working hours inside these tools. Building with them, stress-testing them, occasionally arguing with them. I’m building a platform called Artwell that’s rooted in the belief that AI should sharpen human storytelling rather than flatten it, so this isn’t theoretical for me – it’s the work. And the thing I didn’t expect – the thing nobody really talks about at AI conferences – is how much that daily interaction has changed the way I observe my own reasoning. Not in some grand philosophical way. In small, practical, slightly annoying ways. A model overextends on an inference, and I catch myself thinking, “Wait, didn’t I do almost exactly that in a pitch last Tuesday?” It fills in missing context with a plausible guess, and I realize I’ve been doing the same thing in emails all week without a second thought.

It’s like having a sparring partner who occasionally throws a sloppy punch and forces you to notice your own footwork isn’t as clean as you thought.

Used passively – just accepting whatever the model generates and moving on – AI absolutely reinforces lazy thinking. I’ve seen it happen. I’ve done it myself on a Thursday afternoon when I just need a first draft out the door. But used with even a little bit of intentionality, something different happens. You start noticing the gap between what you actually observed and what you concluded. You start asking yourself questions you’d normally skip: Is this conclusion real, or am I overfitting to one example? Am I projecting intent where there might just be silence? Am I mistaking familiarity for truth?

Those questions aren’t comfortable. But they’re the kind of questions that make you a sharper thinker over time. And not just analytically – the same awareness that helps you catch a bad inference in a strategy deck starts showing up in how you read a room, how you listen to a colleague who’s frustrated, how you check your own assumptions before reacting to an email that rubbed you wrong. It shows up at the dinner table when your kid says something that irritates you and you catch yourself mid-reaction, realizing you’re responding to a pattern from last week’s argument, not to what they’re actually saying right now. When you get in the habit of interrogating your own pattern matching, it doesn’t stay contained to work. It changes how you show up for the people around you.

Maybe that shouldn’t surprise us. Humans built these systems, after all. We modeled them on how we process language and probability and context. But somewhere along the way, most of us forgot that the resemblance runs both directions. AI wasn’t designed to think like a strange alien intelligence. It was designed to think like us – and the reflection it sends back, if you’re willing to look at it honestly, is more revealing than most of us expected.

The broader AI conversation right now is almost entirely about capability – what models will generate next, how fast they’ll scale, which jobs they’ll absorb. And look, I get it. That’s where the money is, and it’s where the anxiety is. But there’s a quieter shift happening underneath that nobody’s writing breathless LinkedIn posts about. As these systems become embedded in how we brainstorm, plan, and make decisions, they’re externalizing something humans have always kept hidden: the mechanics of how we actually think. They’re making it harder to pretend that what we call logic isn’t mostly structured probability. That what we call certainty isn’t compressed experience. That what we call judgment isn’t, at its core, sophisticated guesswork refined over time.

Gracie doesn’t question why she keeps returning to that same patch of grass. The pattern works often enough. She trusts it. Honestly, most days, humans aren’t that different – we’ve just built more elaborate narratives around our version of sniffing the same spot.

AI isn’t replacing thinking. It’s showing us what thinking actually looks like under the hood. And for those of us willing to sit with that instead of flinching from it, the real gift isn’t efficiency or automation. It’s the chance to become more deliberate, more honest, and maybe a little more human in how we use the minds we’ve had all along.

The next time you catch yourself making a confident assumption – in a meeting, in an email, in a conversation with someone you love – pause for a second. Ask yourself where the pattern came from. You might be surprised how often the answer is worth examining.

Tomorrow morning, I think I’ll take Gracie down a different trail. Same nose, same instincts – just new ground. She’ll have to trust the patterns she already knows in a place they’ve never been tested. Come to think of it, that makes two of us.

Share

Subscribe now

Artwell

Merry Christmas from Santa’s Vibeshop!

Share

Subscribe now

Artwell

The recent coverage of Copilot’s slow adoption has focused on familiar explanations: pricing resistance, unclear value, enterprise caution, change management. Those factors are real, but they are surface symptoms. They don’t explain why Microsoft’s AI efforts feel less convincing than they should, given the company’s scale, talent, and access to some of the most advanced models in the world.

The deeper issue sits lower in the stack. Microsoft has invested heavily in intelligence, but far less successfully in the conditions that allow intelligence to become indispensable. What users are reacting to is not a lack of capability, but a lack of cohesion. AI appears in Microsoft’s products as something added, rather than something assumed. It shows up as a panel, a button, a suggestion, a helper, not as the organizing principle of the experience itself.

This is where the idea of Artificial Integration stops being a clever phrase and starts being a diagnosis.

When Work Had a Starting Point

For most of Microsoft’s history, its strength came from owning the starting point of work. You opened Word because that was where documents began. You opened Excel because that was where analysis lived. Windows and Office were not just tools, they were defaults. Productivity flowed inward toward the platform.

Generative AI has reversed that gravity. Today, many people begin with intelligence first and worry about format later. They open ChatGPT or Claude to think, draft, summarize, or structure ideas before deciding whether a document even needs to exist. When AI sits outside the application, the application loses its privilege. Integration becomes the difference between relevance and bypass.

For most of the software era, tools defined how work happened. Interfaces imposed structure, and users adapted. That model held for decades. What AI has changed is the direction of that influence. Workflows now shape tools, not the other way around, and conversational interfaces collapse that feedback loop almost instantly. When people can express intent directly, they stop tolerating systems that require translation, navigation, or indirection.

The Sidebar Problem

This is where Microsoft’s challenge becomes visible. Copilot often feels adjacent to the work rather than embedded within it. The experience suggests augmentation instead of authorship, assistance instead of orchestration. That may have been a reasonable compromise under intense time pressure, but it carries consequences. Once users experience AI as something separate from the core workflow, they start asking a simple question: why am I here at all?

What users increasingly expect is not an AI they consult, but an application that behaves intelligently by default. Conversation has become the new entry point for work. When intelligence lives somewhere else, in a panel, a sidebar, or a separate mental step, users feel the friction immediately. They are being asked to meet the system where it is, rather than the system meeting them where they already work.

This isn’t an argument that platforms should blindly follow users. It’s an acknowledgment that the interface for intelligence has shifted. Users still work inside applications, but they expect those applications to respond directly to intent – not to route intelligence through a separate step.

The Coherence Gap

The difficulty of unwinding these decisions is amplified by ecosystem realities. Apple and Google both benefit from something Microsoft no longer has: a cohesive, end-to-end environment where hardware, operating system, and intelligence evolve together. Apple’s decision to delay public AI commitments while reinforcing its hardware ecosystem now looks prescient. When Gemini enters Apple Intelligence, it will arrive in a system already designed to feel unified. The intelligence will not compete with the interface; it will inhabit it.

Google enjoys a similar advantage through Android and Pixel. Intelligence moves with the device, not alongside it.

Microsoft, by contrast, is integrating AI into a product surface that has become fragmented across devices, platforms, and contexts. Windows no longer functions as an operating-system moat. Office no longer defines the default starting point for knowledge work. Productivity has become more fluid, more conversational, and less bound to a specific application.

In that environment, AI that feels bolted on is easy to bypass.

Smart Enough, But Not Native Enough

This is why Copilot’s adoption challenges should not be interpreted as a referendum on AI’s usefulness. They are a signal that integration quality now matters as much as intelligence itself. Users do not reject AI because it is unfamiliar. They reject it when it feels optional, interruptive, or redundant with tools they already trust.

Microsoft still has formidable strengths. Azure remains one of the most significant infrastructure moats in the industry, and the company’s enterprise footprint is unmatched. But infrastructure alone does not create habit. Habit forms where work begins. And increasingly, work begins in conversation, not in documents.

Until AI becomes the place where work starts inside Microsoft’s ecosystem, not the place users visit after the fact, adoption will continue to lag behind capability.

Beyond the Bolt-On

Artificial Integration is not a branding problem or a messaging problem. It is the structural gap between intelligence and experience, and until that gap closes, capability alone will not drive adoption.

The companies that lead the next decade of AI will not be the ones with the largest models or the loudest launches. They will be the ones that design systems where intelligence feels native, inevitable, and inseparable from the work itself. Microsoft is still capable of being one of those companies. But that future depends on whether AI remains something it adds, or something it builds around.

Bruce Bracken is the founder of Artwell.ai and former Head of Podcasts & Digital Experiences at Microsoft, where he built a content ecosystem that generated $59M in attributed Azure revenue. He writes about the gap between what AI companies say and what humans actually experience.

Share

Subscribe now

Artwell