In 2012, a neural network learned to recognize cats in YouTube videos. It needed 16,000 processors and three days to do it. The result made the front page of the New York Times.

In 2025, a neural network passed the bar exam, wrote production-quality code, diagnosed rare diseases from blurry scans, and composed music that made professional musicians genuinely uncomfortable. It ran on your phone.

The distance between these two moments is thirteen years. But the real story — the one that reshapes every creative brief you’ll ever write, every brand strategy you’ll ever build, every team you’ll ever manage — happened in the last six of those years. And one book tried to capture it while it was still happening.

That book is The Scaling Era: An Oral History of AI, 2019–2025 by Dwarkesh Patel with Gavin Leech. Published by Stripe Press. 248 pages. 170+ technical definitions. Eight thematic chapters. And a cast of characters that reads like the Avengers of artificial intelligence — if the Avengers disagreed on whether they were saving the world or accidentally ending it.

Here’s why you, specifically, should care.

Who made this and why it matters

Dwarkesh Patel is a 27-year-old podcaster. That sentence should make you skeptical. It made me skeptical. But here’s the thing: his podcast guests include Dario Amodei (CEO, Anthropic — the company behind Claude), Demis Hassabis (CEO, Google DeepMind — Nobel Prize winner), Ilya Sutskever (co-founder of OpenAI, now building Safe Superintelligence), Mark Zuckerberg, and Eliezer Yudkowsky, the person who has spent two decades arguing that AI might kill us all.

Jeff Dean — Google’s chief scientist — called the book a great distillation of conversations that help people understand modern AI. Patrick McKenzie posed the only test that matters: would you learn more reading this book than spending the equivalent time using an LLM? His answer was yes.

Gavin Leech, the co-author, has a PhD in AI and co-founded Arb Research. His job was to do what Patel’s interviews alone couldn’t: add 170+ margin definitions, annotate technical claims, fact-check timelines, and thread the excerpts into a coherent narrative.

The publisher is Stripe Press — Patrick Collison’s imprint. The same Collison who built a $95 billion payments company and publishes books about progress the way other billionaires collect yachts.

This lineage matters because it tells you what the book isn’t. It isn’t another “AI Will Change Everything” manifesto written by a consultant who discovered ChatGPT in January 2023. It isn’t a doom pamphlet. It isn’t a how-to guide. It’s an oral history — in the tradition of Studs Terkel, not Kai-Fu Lee. The people who are building the most powerful technology in human history, talking on the record about what they think they’re doing, what they’re afraid of, and what they don’t understand.

The only book of its kind (at least that’s how it feels)

Between 2020 and 2025, dozens of AI books hit the market. Most suffered from the same two problems.

Problem one: they were written by outsiders looking in. Journalists, consultants, futurists — people interpreting the technology through secondhand accounts and press releases. The resulting books read like restaurant reviews written by people who’ve never cooked.

Problem two: they aged like milk. A book finished in March 2023 was outdated by the time it shipped in September. The field moves at a pace that makes traditional publishing look like sending messages by carrier pigeon.

Patel’s book dodges both problems through its format. Oral history doesn’t claim to predict — it captures. When Dario Amodei says he was surprised by how much scaling kept working, that surprise doesn’t expire. When François Chollet argues that LLMs can’t truly reason — only fetch memorized patterns — that argument doesn’t become irrelevant just because GPT-5 launched. These are the frameworks, intuitions, and disagreements of the people building the systems. The thinking behind the technology ages better than any snapshot of the technology itself.

And as of April 2026, no other book has attempted this at this scale with this level of access. There are academic papers. There are podcasts. There are blog posts and Substacks (including, yes, this one). But a curated, structured oral history of the scaling era, with the actual protagonists? This is it.

The Translator’s Chapter Guide: What Each Section Means for You

I read this book with a specific lens: what does each chapter tell someone who builds brands, manages creative teams, or makes product decisions? Here’s the field guide.

Chapter 1: Scaling — “Why your AI budget will grow exponentially and there’s nothing you can do about it”

The core thesis: making AI models bigger, with more data and more compute, produces reliably better results. This is called the “scaling hypothesis,” and it’s the central bet that every major AI lab is making. Billions of dollars riding on a power law.

For creative professionals, the translation is this: the AI tools you’re using today are the worst they’ll ever be. Not metaphorically. Literally. The models get better with more resources, and the resources keep doubling. When Mark Zuckerberg says it’s worth investing $100 billion-plus on the assumption that scaling continues — that’s not a bluff. That’s a budget line item.

The practical implication: stop evaluating AI tools as fixed products. They’re trajectories. The question isn’t “can Midjourney do this today?” It’s “what happens to my competitive position when it can do this in 18 months?”

Chapter 2: Evals — “Nobody agrees on how to measure intelligence, and that’s your problem too”

The most fascinating debate in the book: François Chollet (creator of Keras, senior researcher at Google) arguing with Patel about whether LLMs actually reason or just memorize. Chollet’s position: if you show a model a problem it hasn’t seen before — genuinely novel, not a variation of training data — it fails. LLMs don’t synthesize new solutions. They fetch stored ones.

Patel pushes back: isn’t that what humans do too? We drill math for years before we can “reason” about it.

This debate should be required reading for every creative director. Because the exact same question applies to AI-generated campaigns: is the model being creative, or is it fetching the most statistically probable remix of everything it’s seen? If you’ve read my Creativity Gap research, you know the answer is nuanced. But the evaluation problem is real. When Adobe says “creative AI” and Google says “creative AI,” they may not be measuring the same thing. Chollet’s ARC benchmark — puzzles designed to resist memorization — remains unsolved by any LLM. Children can do them.

Chapter 3: Internals — “The machine is a black box, and the people who built it are the first to admit it”

Trenton Bricken from Anthropic’s interpretability team explains how they’re trying to understand what happens inside a model. The honest answer: they’re only scratching the surface. Mechanistic interpretability — figuring out which “neurons” correspond to which concepts — is the AI equivalent of neuroscience. We built the brain before we understood it.

For anyone making decisions based on AI output — brand guidelines, campaign copy, product recommendations — this chapter is a cold shower. The systems are powerful. They are also, in a fundamental sense, unexplained. When you use an LLM for ideation, you’re collaborating with something nobody fully understands. That’s not a reason to stop. It’s a reason to keep your judgment sharp.

Chapter 4: Safety — “The alignment tax you’re already paying”

This is where the book gets heavy. Carl Shulman calmly discusses the probability of AI takeover scenarios. Yudkowsky argues we haven’t earned enough time. Dario Amodei walks the line between building the most powerful models on Earth and worrying about what they might do.

The creative industry connection is less obvious here, but it’s real. Every AI model you use has been through RLHF — Reinforcement Learning from Human Feedback. That process makes models helpful, harmless, and honest. It also makes them safe. And safe, as I’ve written before, is the enemy of great. Every alignment optimization comes at a diversity cost. The NeurIPS 2025 “Artificial Hivemind” paper shows that even models from different families converge on similar outputs. That’s not a bug in the alignment process. It’s a feature — with side effects that hit the creative industry first.

Chapter 5: Inputs — “The data wall is real, and you’re part of the supply chain”

Dylan Patel from SemiAnalysis and others discuss the looming data wall: we’re running out of high-quality text to train on. The solutions — synthetic data, multimodal training, longer context windows — are all compromises.

For brand owners: your content is training data. Your campaigns, your copy, your visual assets — they’re part of the distribution these models learn from. The quality of AI output is downstream of the quality of human output. This is both a threat (your work gets absorbed without compensation) and a strategic insight (the scarcity of truly original content increases).

Chapter 6: Impact — “What happens to your job, your team, and your budget”

Sholto Douglas from Anthropic describes how even he — a reinforcement learning infrastructure lead — can imagine AI automating most of his tasks within a few years. If the people building AI think their own jobs are at risk, the creative industry shouldn’t assume immunity.

But the nuance matters. The book doesn’t predict mass replacement. It predicts transformation. Zuckerberg talks about inference compute vs. training compute. Hassabis talks about the energy constraints that slow deployment. The intelligence explosion, if it comes, will be bottlenecked by power grids and permitting processes — not just algorithms. This means the transition won’t be instant. But it also won’t be optional.

Chapters 7 & 8: Explosion and Timelines — “How much time do you have?”

The range of predictions in this book is staggering. Some interviewees think AGI arrives by 2028. Others think decades. Carl Shulman puts the probability of catastrophic outcomes at 20-25% on a bad day. The Anthropic timeline of 2028 seemed plausible to several guests at the time of recording.

What’s useful here isn’t the specific dates. It’s the fact that the people closest to the technology disagree wildly about when transformative AI arrives — but almost none of them say “never.” For strategic planning in any industry, that’s the signal. Not when, but that the range of credible timelines is measured in years, not generations.

The expiry date problem (and why it doesn’t matter)

The book was published in October 2025. We’re in April 2026. Claude has shipped new capabilities. GPT-5 happened. The AI landscape has shifted again.

Does that make the book outdated?

No. And this is the key insight about oral history as a format: facts expire, but thinking patterns don’t. When Ilya Sutskever says he believes scaling will keep working but that the link between next-word prediction and reasoning is complicated — that framework for thinking about AI capabilities is still the right one, regardless of which model just launched. When Chollet says generality is not specificity scaled up — that’s a philosophical position, not a product spec. It doesn’t expire with a software update.

The book captures how the people shaping AI think, not just what they’ve built. And for anyone trying to navigate AI’s impact on their work, understanding the thinking is worth more than knowing the latest benchmark scores.

The creative industry’s homework

If you’re a creative director, a brand strategist, a product manager, or anyone who makes decisions about how humans and machines should work together — this book gives you something no AI tutorial, no LinkedIn hot take, and no conference keynote can provide: the unfiltered internal logic of the people building the systems you’re increasingly dependent on.

You’ll understand why the models behave the way they do. You’ll understand why alignment makes them conservative. You’ll understand why scaling creates both opportunity and homogeneity. You’ll understand why nobody — not Amodei, not Hassabis, not Zuckerberg — knows exactly where this is going.

And that understanding is the first step toward the only AI strategy that actually works: treating AI not as a tool to deploy, but as a system to navigate. With your judgment intact.

The book started with neural networks identifying species in blurry photos. It ends with people seriously debating whether superintelligence arrives this decade. Somewhere in between those two endpoints lives the future of every creative brief, every brand, and every team you’ll ever build.

Read it. Then read it again in a year. See which parts aged well. That’s the real test.

Before we begin: how this list was born

Two-three months ago, I walked into my first product management class and heard a sentence that almost made me leave.

“We need to validate PMF before defining our GTM and aligning OKRs with the North Star Metric.” Something like this. Or even worth,

Seventeen+ years in advertising. Ogilvy. Leo Burnett. A co-founded creative academy. Hundreds of campaigns, brand systems, creative workshops. And in that moment, I felt like a tourist who’d been airdropped into a country where everyone speaks in code — confident, fast, acronym-heavy code — and nobody hands you a dictionary at the border.

So I made one for myself.

Not a textbook glossary. More like a translation layer — a document I kept updating every time I sat in a meeting and thought: wait, I know this concept. I’ve been doing this for fifteen years. They just call it something else.

That document became this piece.

Because here’s what I discovered, term by term, meeting by meeting: the tech world and the creative world aren’t two different planets. They’re the same city with two dialects. The product people talk about “activation” — we call it the aha moment. They obsess over “retention” — we built careers on brand loyalty. They worship “positioning” — we literally wrote the positioning statements they’re trying to reverse-engineer from a feature list.

The gap isn’t knowledge. It’s vocabulary.

This is for anyone making the same crossing I did — or thinking about it. Creative directors curious about product. Brand strategists eyeing startups. Designers who keep hearing “sprint” and aren’t sure if anyone’s actually running. You don’t need to become a product manager. You need to realize you’ve been one all along — just without the job title and the Jira board.

Let’s translate.

I. The brief is the same. The name tag isn’t.

You know that moment in a brand strategy project when everything clicks? The audience research lands, the insight is sharp, the positioning feels inevitable, and the client looks at the creative concept and says: “This is exactly what we needed but couldn’t articulate.”

In tech, that moment has a name: Product-Market Fit — or PMF, because the tech world runs on three-letter abbreviations the way agencies run on coffee and passive aggression.

PMF is the moment your product satisfies a real market need so precisely that growth starts pulling instead of pushing. Teams spend months — sometimes years — trying to find it. VCs won’t fund scaling until they see it. It’s the single most discussed concept in startup culture.

And you’ve felt it before. Every time a campaign hit — when the brand insight unlocked a message that made the audience lean in — that was PMF’s creative cousin. Brand-market resonance. Same physics, different lab.

Once you see this pattern, it starts repeating everywhere.

Value Proposition — the thing founders agonize over in pitch decks — is positioning. Literally. It’s the “why us” distilled to one sentence. You’ve written these at agencies, except you called them positioning statements and the client changed them four times before launch.

ICP — Ideal Customer Profile — is the target audience persona, “a brand’s muse”, but sharpened to a surgical point. Not “women 25–45 interested in wellness.” More like “Head of Marketing at a Series B fintech in DACH, 3–5 years in role, reports to a CFO who doesn’t believe in brand.” In agencies, you’d call this a brief’s audience section done properly. In tech, it determines where every sales dollar goes.

North Star Metric is the one number that tells you whether your product actually matters to people. Think of it as the KPI that sits above all other KPIs — the single signal that captures whether you’re delivering real value. In brand terms, it’s the metric you’d tattoo on the CMO’s forearm if they’d let you. Not impressions. Not reach. The thing that proves the work is working.

And Positioning — yes, it exists in tech as its own term, separate from brand strategy — means exactly what you think it means. How your product occupies a distinct place in the customer’s mind relative to alternatives. Identical concept. But here’s the gap: most founders confuse positioning with features. “We’re like Slack but with AI.” That’s not positioning. That’s a feature comparison wearing positioning’s coat. Your job, if you cross over, is to translate features into meaning. You’ve been doing this your whole career. They just didn’t know they needed you yet.

II. The funnel you already know (they just gave it a pirate flag)

Here’s a gift from the growth marketing world: AARRR — also known as Pirate Metrics, because someone at a startup in 2007 realized the acronym sounds like a pirate and the name stuck forever. Welcome to tech naming conventions.

AARRR stands for: Acquisition → Activation → Retention → Revenue → Referral.

Now read this: Awareness → Trial → Loyalty → Revenue → Advocacy.

That’s the brand funnel. Your funnel. The one you’ve been building campaigns around for years. The mapping is almost embarrassingly direct.

Activation — the moment a new user first experiences the core value of the product — is your aha moment by another name. The instant someone goes from “I signed up” to “oh, this is why this exists.” In campaigns, you’ve designed these moments intentionally: the visual that stops the scroll, the headline that reframes the problem, the demo that makes the prospect say “wait, show me that again.” Same craft, different medium.

Retention and Churn are the yin and yang of product health. Retention measures who stays. Churn measures who leaves. Together, they’re the most important growth metrics after PMF — because acquiring users who leave is just expensive theater.

In brand terms? This is loyalty. The difference is precision. Tech measures retention weekly, with cohort analysis and behavioral data. Branding measures it annually, with brand trackers and NPS surveys. Same question — are people coming back? — asked at radically different clock speeds.

CAC and LTV — Customer Acquisition Cost and Lifetime Value — are the unit economics that keep founders awake at 3 AM. If it costs you more to get a customer than that customer will ever pay you, you don’t have a business. You have a bonfire.

You understand this intuitively from agency economics: cost per lead, cost per conversion, ROI on media spend. CAC/LTV is the same math, applied to the entire customer relationship instead of a single campaign. And here’s a line you can use in your first startup meeting: “Brand is what makes your CAC go down and your LTV go up over time. It’s compound interest for trust.” Watch the room recalibrate.

III. The process is identical (they just ship faster)

If you’ve ever run a production cycle at an agency — brief lands Monday, internal review Wednesday, client presentation Friday — you already understand Sprints. A sprint is a fixed work cycle, usually two weeks, with planning at the start, execution in the middle, and a review at the end. Most tech teams run on this rhythm. It’s production scheduling with a standup meeting every morning and a retrospective every two weeks.

Agile is the philosophy underneath: build in short cycles, get feedback early, adapt constantly. If you’ve worked at any agency that wasn’t completely dysfunctional, you’ve practiced this. You just didn’t have a manifesto printed on the wall.

MVP — Minimum Viable Product — is concept testing. Not the final campaign. Not the polished brand book. The scrappiest version of the idea that lets you put it in front of real people and learn whether it works before you spend six months and half a million producing the full thing. You’ve done this every time you tested a creative concept with a focus group or ran an A/B test on ad copy. MVP energy is: show the sketch, not the oil painting.

Lean Startup formalizes this into a loop: Build → Measure → Learn. Build a small thing. Measure how people react. Learn from the data. Repeat. If this sounds familiar, it’s because agencies have been running a version of it forever: concept → test → feedback → iteration. The startup world just gave it a book deal and a TED talk.

Design Thinking — Empathize, Define, Ideate, Prototype, Test — maps so directly to the agency process that the first time I saw it on a slide, I thought someone had plagiarized a planning department’s workflow. Consumer research → brief → brainstorm → concept → testing. Identical structure. IDEO and Stanford’s d.school packaged it beautifully, but the bones are the same bones you’ve been building on.

And Pivot — the word that makes founders sound brave and investors nervous — is rebranding triggered by market feedback. You’ve watched brands pivot their positioning when the original one didn’t land. Same logic, higher emotional stakes, because in startups the pivot isn’t just a messaging change — it might mean rebuilding the product. But the mental model is identical: the market told us something, and we’re listening instead of arguing.

IV. The strategy layer (where you’re already fluent)

This is where creative people have the biggest unfair advantage — and don’t know it.

GTM — Go-to-Market — is the strategy for launching a product: who’s the audience, what channels, what messaging, what pricing, what sequence. Read that sentence again. Now tell me it’s not a campaign brief combined with a media plan and a launch strategy. You’ve built GTMs your entire career. You called them launch plans, comms strategies, or “the deck for the client.” The deliverable is the same. The acronym is new.

Jobs To Be Done (JTBD) is the framework that says: people don’t buy products, they hire them to do a job. A person doesn’t buy a drill because they want a drill. They buy a drill because they want a hole in the wall. And they want a hole in the wall because they want to hang a photo. And they want to hang a photo because they want their apartment to feel like home.

Sound familiar? It should. This is consumer insight — the deep, motivational layer underneath the surface behavior. At Ogilvy, you called it “the real reason people buy.” Clayton Christensen gave it a framework and a name. Same excavation, different shovel.

Blue Ocean Strategy — creating uncontested market space instead of fighting in a bloody “red ocean” of competition — is category design meets positioning. Finding the white space for a brand to own. You’ve done this: identified the gap no competitor occupies and built a brand to fill it. In tech, they write entire business plans around this insight. In advertising, you put it on slide 7 of the strategy deck and moved on.

Category Design takes Blue Ocean further: don’t just find an empty space — create a new category and define the rules. This is naming and positioning at the highest level. When a startup says “we’re creating a new category,” what they mean is: we’re going to make people think about a problem differently, and then position ourselves as the only solution to the problem we just defined. If that’s not strategic creative work, nothing is.

Brand Architecture — how a company organizes its brands (branded house, house of brands, endorsed, hybrid) — exists in tech too, and it becomes critical when startups grow, acquire companies, or launch sub-products. The founder who started with one product and one name suddenly has three products and no idea whether to put them under one brand or give them separate identities. You’ve solved this problem before. They haven’t. That’s your leverage.

And the Hook Model (Nir Eyal’s framework: Trigger → Action → Variable Reward → Investment) is behavioral design meets creative direction. You’ve created campaigns that trigger behavior — the ad that makes you want, the packaging that makes you reach, the CTA that makes you click. The Hook Model applies the same psychology to product design: how do you make someone come back without asking them to? You already think this way. You just apply it in 30-second intervals instead of product lifecycles.

V. The growth engine (where brand is the unfair advantage nobody talks about)

Three concepts that run every growth conversation in tech — and where creative people are chronically underestimated.

Flywheel is a self-reinforcing growth loop. Good product → happy users → word of mouth → more users → better data → better product → happier users → more word of mouth. Each rotation accelerates the next. Amazon’s flywheel is legendary. So is Apple’s ecosystem lock-in.

Here’s what most tech teams miss: brand amplifies every stage of the flywheel. It’s the lubricant that makes the loop spin faster. A strong brand means lower acquisition costs (people come to you), higher activation rates (trust is pre-built), better retention (identity, not just utility), and more referrals (people share things they’re proud to be associated with). If you can articulate this in a product meeting, you’ll watch the “brand is a nice-to-have” crowd go very quiet.

PLG — Product-Led Growth — means the product itself drives acquisition. No sales team cold-calling. No paid ads (or fewer of them). The product is so good that people try it, love it, invite colleagues, and upgrade. Slack, Notion, Figma, Calendly — all PLG.

The PLG crowd sometimes believes brand doesn’t matter in their model. They’re wrong. In PLG, the product IS the primary brand touchpoint. Every onboarding screen, every empty state, every error message, every tooltip — that’s brand. UX writing is copywriting. The upgrade prompt is a sales page. The loading screen is a billboard. PLG doesn’t eliminate brand. It makes every pixel of the product a brand decision.

Moat — a sustainable competitive advantage that’s hard to copy. Network effects, proprietary data, switching costs, patents, economies of scale. Warren Buffett coined the business metaphor. Every investor asks about it.

Here’s the line I keep using: brand is a moat. Not the only moat. Not always the deepest moat. But a real one. Brand equity compounds over time, like product quality. It creates switching costs that aren’t contractual but emotional. It makes your CAC lower than competitors’. It makes talent want to work for you. When a founder tells you “we don’t need brand at this stage,” what they’re really saying is “we don’t understand that brand isn’t a stage — it’s a layer.”

VI. The money conversation (where you stop feeling like a tourist)

The last cluster. The one that makes creative people’s eyes glaze over — until you realize it’s the same conversation you’ve been having, just with different units.

MRR / ARR — Monthly and Annual Recurring Revenue — is subscription income. The heartbeat of every SaaS business. If you’ve worked with any subscription-based media company (I spent years at MEGOGO, one of the largest streaming platforms in Eastern Europe), you already know this model. Revenue isn’t a one-time transaction. It’s a relationship measured in months.

Burn Rate and Runway — how fast a startup spends money (burn) and how long until the money runs out (runway). If you’ve co-founded anything, you know this feeling in your bones. It’s the creative budget conversation, but existential. “We have 14 months of runway” means “we have 14 months to prove this works or we’re done.” Concentrates the mind wonderfully.

Series A, B, C — funding rounds, each with different expectations. Seed is the idea stage (you have a thesis, maybe a prototype). Series A means you’ve found PMF and need money to grow. Series B is scaling. Series C is market dominance.

Here’s why this matters for creative people: each stage needs different brand work. A seed-stage startup needs naming, core positioning, maybe a logo that doesn’t embarrass them in a pitch deck. A Series A company needs an identity system that scales. A Series B company needs brand architecture because they’re launching a second product. A Series C company needs a brand that can recruit 200 people in six months and make them all feel like they joined something that matters. If you can map your services to these stages, you stop being “the brand person” and start being “the person who understands what we need right now.”

OKRs — Objectives and Key Results — is the goal-setting framework most tech companies use. An Objective is ambitious and qualitative (”Become the most trusted AI tool for creative teams”). Key Results are measurable (”Reach 10K weekly active users,” “Achieve NPS > 60,” “Reduce onboarding time to under 3 minutes”).

It’s a creative brief. Objective = what we want to achieve. Key Results = how we’ll know it worked. You’ve been writing these your entire career. The format is tighter, the cadence is quarterly, and there’s a scoring system. But the thinking is identical.

The real translation

I could keep going. There are dozens more terms — cohort analysis, feature flags, technical debt, API integrations, A/B testing at scale. But the point of this piece isn’t to be exhaustive. It’s to be liberating.

Because the hardest part of crossing from creative to tech isn’t learning new concepts. It’s the imposter syndrome that comes from hearing familiar ideas in an unfamiliar accent and concluding that you don’t belong in the room.

You belong in the room.

The creative industry trained you in consumer psychology, narrative structure, visual systems, audience segmentation, campaign architecture, and the discipline of generating ideas under pressure with a deadline and a skeptical client. That’s not a background you need to overcome. That’s a background most product teams are missing — and paying consultants to approximate.

The vocabulary gap is real, but it’s narrow. A few weeks of immersion and you’ll stop translating in your head. A few months and you’ll start catching things the native speakers miss — because you see the brand layer they keep forgetting is there.

Learn their language. Don’t abandon yours.

The tech world doesn’t need another product manager. It needs people who can look at a product roadmap and see a story. Who can look at a GTM plan and see a campaign. Who can look at a North Star Metric and ask: “But does anyone feel anything about this?”

That’s not a skill you need to acquire.

That’s the skill you’ve been sharpening for years.

Serhiy Vovk is the founder of VOVK (Creative) Consulting and co-founder of Kyiv Academy of Media Arts (KAMA). He writes about what happens when the creative industry learns to speak machine at aibubbledotcom.com.

The Vanishing

I didn’t plan to stop using my browser. It just happened.

A few weeks ago, I opened my laptop to check something — an email thread, a flight price, a fact I needed for a piece I was writing. Standard Tuesday morning. Except somewhere between the intention and the action, I realized I’d already done all of it. Inside a single conversation window. No Chrome. No Gmail tab. No toggling between six apps like a pianist playing separate instruments with each finger.

The browser didn’t crash. It didn’t freeze. It just… became unnecessary.

Sit with that for a second. Because this isn’t a story about one person’s workflow getting slightly more efficient. This is a story about the most fundamental change in how humans interact with computers since the graphical user interface shipped in 1984. And it’s happening right now, on your machine and mine.

You’re probably already feeling it. Even if you haven’t named it yet.

* * *

The Screen You Stopped Seeing

Let me describe your day. Tell me if this sounds familiar.

You wake up. You check your phone. Not for one thing — for a cascade of things routed through a cascade of apps. Messages here. Email there. Calendar somewhere else. News in a fifth place. Weather in a sixth. You haven’t brushed your teeth and you’ve already performed fifteen context switches across six interfaces, each designed by a different company, each demanding you learn its grammar, remember its layout, tolerate its ads.

Multiply that by a workday. Slack for messaging. Notion for docs. Google for search. Figma for design. Jira for tasks. Zoom for calls. Each one is a toll booth between your intention and its execution. You don’t think of it that way because you’ve been trained, over two decades of smartphone culture, to accept fragmentation as the price of digital life.

But it’s not a law of nature. It’s a business model.

Every app exists because a company claimed one slice of your intent and built a walled garden around it. Uber owns your ride. Gmail owns your inbox. Spotify owns your music. Each one said: “To do this one thing, you must enter our space, on our terms, through our door.” And for fifteen years, we complied. Because there was no alternative.

That alternative is arriving. And it looks nothing like an app.

* * *

A 35-Year-Old Prophecy Coming True in the Wrong Order

In 1991, a computer scientist named Mark Weiser published a paper in Scientific American called “The Computer for the 21st Century.” He was CTO of Xerox PARC — the lab that invented the modern desktop, the mouse, and ethernet. And he was already tired of all of it.

Weiser’s thesis was deceptively simple: the most profound technologies are those that disappear. They weave into the fabric of everyday life until they’re indistinguishable from it. Writing was his example — once a rare, expensive, tightly controlled technology, now so ubiquitous you don’t notice the text on a candy wrapper. He wanted computers to follow the same arc. Not bigger screens. Not more personal devices. Invisible computing, woven into the environment, requiring zero attention.

He called it “ubiquitous computing.” Then, with John Seely Brown, he gave it a more poetic name: calm technology. Technology that informs but doesn’t demand your focus.

Weiser died in 1999. Before the iPhone. Before apps. Before the attention economy turned his vision inside out. Instead of calm, we got chaos. Instead of invisible, we got screens competing for every second of our waking lives. The entire arc of consumer technology from 2007 to 2024 was the opposite of what he proposed: louder, brighter, stickier, more demanding.

Here’s the twist. In 2026, Weiser’s prophecy is finally coming true. Just not the way he imagined.

It’s not arriving through smart walls and embedded sensors. It’s arriving through language.

* * *

The Instruct/Verify Era

The HCI field — Human-Computer Interaction, the discipline that studies how people and machines communicate — has names for what’s emerging. Several names, because the field loves a good taxonomy war.

Ambient intelligence — context-aware environments that anticipate needs without explicit commands. The convergence of AI, IoT, and edge computing into systems that sense, adapt, and respond. The global market hit $36 billion in 2025. Projected: over $440 billion by 2034, growing at 25%+ annually. That’s not a forecast. That’s a gravitational field.

Zero UI / The Invisible Interface — the design philosophy that the best interface is no interface at all. At CES 2026, this wasn’t a panel topic. It was the premise of the entire show. The conversation shifted from “AI as a feature” to “AI as the environment.”

Agentic AI — systems that don’t just generate text but execute actions autonomously. Book flights. File reports. Navigate APIs. Gartner projects 40% of enterprise applications will embed task-specific AI agents by end of 2026. IDC says AI copilots will land in 80% of enterprise workplace apps. The paradigm shifts from “ask and receive” to “command and execute.”

But forget the taxonomy for a moment. Here’s what’s happening on the ground.

Carl Pei, CEO of Nothing, stood at SXSW three weeks ago and said apps will become obsolete. His example: meeting a friend for coffee currently requires launching four separate apps — messaging, maps, ride-hailing, calendar. A single AI agent should resolve the entire intent in one sentence. Nothing raised $200 million on this thesis.

In January 2026 alone, $2 trillion in SaaS market capitalization evaporated. Not a correction. Structural disruption. When one AI agent replaces dozens of software licenses, the per-seat pricing model that built an entire industry starts to collapse.

The industry phrase catching on: we’re entering the “Instruct/Verify” era. You state intent in natural language. The system executes. You verify the result. That’s the loop. Everything else — navigation, clicking, switching, searching — disappears.

Everything else disappears.

* * *

The Race to Build the Last Device

If the software layer is collapsing into a single conversational interface, the hardware question follows: what do you hold? Wear? Place in your room?

The answers are already in prototype.

OpenAI + Jony Ive are building a screenless AI device codenamed “Sweetpea.” A pill-shaped wearable that hooks behind the ear. No display. No keyboard. Voice, sensors, contextual awareness. Custom 2nm chip. Manufactured by Foxconn. Target: 40–50 million units in year one. Launch: second half of 2026. Altman calls it “more peaceful than a smartphone.” The man who built ChatGPT is betting the next interface has no screen at all.

Apple is developing three AI wearable categories simultaneously: smart glasses (codename N50, production by December 2026, launch 2027), an AI pendant with dual cameras and microphones, and camera-equipped AirPods. All built around a radically overhauled Siri — retooled as a conversational chatbot powered by models co-developed with Google. iOS 27, shipping later this year, debuts this new Siri. Apple is accelerating because it knows: if someone else builds the last interface first, the iPhone becomes a peripheral.

Meta’s Ray-Ban smart glasses are in market and gaining traction. No screen. Voice-first. Camera-equipped. They work because glasses are something people already wear — and the AI makes them useful enough to become necessary. For vision-impaired users, they read signs and describe surroundings. For everyone else, the first mass-market proof that screenless AI works in daily life.

The pattern: every major player converges on the same thesis — screenless, voice-first, context-aware, always present. The interface dissolves. The machine becomes an ambient companion.

Weiser’s calm technology. Except it’s not in the walls. It’s in your ear.

* * *

The Trillion-Dollar Funnel

Now for the part that should make you uncomfortable.

Right now, your digital life is fragmented across 40+ apps because each company controls one slice of your intent. That fragmentation is annoying — but it’s also a form of distributed power. No single entity owns your entire digital existence. You choose your email provider, your music service, your ride-hailing app, your bank. The friction is real. So is the freedom.

The moment one entity provides a single natural-language layer that executes across all domains, the power dynamics invert.

You don’t need Uber’s app if your agent books rides. You don’t need Gmail if your agent handles mail. You don’t need a browser if your agent fetches and acts. You don’t need a screen if your agent speaks results into your ear.

Each sentence eliminates a multi-billion-dollar industry. And funnels its value into a single chokepoint: the conversational layer.

Whoever owns the last interface owns the relationship between humanity and its tools.

The contenders:

Apple — ecosystem lock-in, a billion active devices, hardware-software integration nobody matches. If Siri becomes genuinely intelligent, Apple controls the full vertical: chip, device, OS, agent.

Google — data, infrastructure, models. Search was already the closest thing to a universal intent layer. Gemini is their bet that conversational AI replaces the search bar — and with it, the browser, the web, the entire architecture of information flow.

OpenAI — models, distribution (300+ million ChatGPT users), and now hardware via Ive. No ecosystem yet. Building one from zero. The audacity is staggering: they’re trying to leapfrog the smartphone entirely.

Meta — social graph plus a head start on wearable AI people actually use. The Ray-Ban glasses are quietly becoming the Trojan horse.

Anthropic — building the deepest integrations into professional workflows: email, calendar, documents, Slack, code. Not hardware. Not consumer devices. The workplace as the entry point.

Five companies. Five strategies. One prize. Not market share. The funnel through which all human digital intent passes. A position so powerful it makes Google’s search monopoly look like a regional advantage.

* * *

The Oxygen Problem

Mark Weiser wanted technology to be like literacy: everywhere, invisible, calm.

What we might get instead is technology like oxygen: everywhere, invisible — and controlled by whoever owns the supply.

In the app era, if Uber annoyed you, you switched to Lyft. If Gmail felt invasive, you moved to ProtonMail. The exits existed because the interfaces were separate. Competition happened at the app level.

In the single-interface era, what does “switching” even mean? If your AI agent handles everything — rides, email, shopping, banking, health, work — the switching cost isn’t inconvenience. It’s identity loss. The agent knows your preferences, your history, your patterns, your relationships. Leaving it means leaving yourself behind.

We’ve seen this before. Social media created it with social graphs. Cloud storage created it with data gravity. But a universal AI agent creates something qualitatively different: cognitive lock-in. Not your files or friends. Your decision-making apparatus.

The antitrust framework doesn’t have vocabulary for this yet. Regulators think in “markets” and “competitors.” But when one company becomes the medium through which a person interacts with reality, market is no longer the right word.

Platform isn’t the right word either.

* * *

When, Not If

A realistic timeline — not from futurist keynotes, but from product roadmaps and supply chains:

2026 (now): OpenAI’s Sweetpea debuts. Apple ships chatbot-Siri in iOS 27. AI agents handle multi-step tasks in professional tools. Browser usage begins measurable decline among power users. You notice you open fewer apps per day. The Instruct/Verify loop becomes default for routine tasks.

2027: Apple launches glasses and pendant. Second-gen screenless devices arrive. Voice-first becomes normal in professional settings. The first major app companies report declining DAU — not because users left, but because agents replaced the need to open the app.

2028–2029: Consolidation. One or two companies emerge as the dominant conversational layer. Regulatory battles begin. The phrase “cognitive infrastructure” enters policy debates. The smartphone doesn’t die — it becomes a battery and antenna for the wearable that handles everything else.

2030+: The interface disappears for most daily tasks. Computing becomes ambient in the Weiser sense — but concentrated in a way he never envisioned. The question is no longer technology. It’s governance.

* * *

The Translator’s Take

I’m watching this from both sides of the stack.

For 17 years, I built brands inside a world where the medium shaped the message. You designed for screens, for scroll, for tap. The interface was the canvas. Now I’m inside AI product thinking, and I see the canvas dissolving in real time.

Personally? The number of apps I open per day is declining — not by design, but because the conversational layer absorbs their functions. Email, search, scheduling, writing, research, code. One window. I catch myself reaching for a browser and then not reaching, because the answer is already in front of me.

It’s faster. It’s calmer. It’s better. And that’s exactly what makes the monopoly question so urgent.

Because the best monopolies don’t feel like monopolies. They feel like convenience. Like relief. Like progress. You don’t fight them because fighting means going back to the friction, the fragmentation, the fifteen-app Tuesday morning. Who would choose that?

That’s the design. That’s the trap. And that’s the most important conversation in tech right now — one that almost nobody in the creative industry is having, even though it will determine who controls the interface between their ideas and the world.

Weiser dreamed of a world where technology serves humans so seamlessly we forget it’s there. We’re about to get that world. The only question — and it’s not a technical one — is whether we’ll also forget who’s running it.

* * *

Serhiy Vovk is the founder of VOVK (Creative) Consulting and co-founder of Kyiv Academy of Media Arts (KAMA). He writes about what happens when the creative industry learns to speak machine at aibubbledotcom.com.

Last Friday I walked into The Future of AI — a new event series at a local startup hub here in Valencia. One of the organizers opened with something that immediately earned my attention: “There’s too much AI hype. Too many headlines. It’s hard to tell what’s serious and what isn’t. So we’re going to do something radical — talk specifics.”

No manifestos. No evangelism. Just practitioners dissecting what’s actually happening.

I took notes. These are them.

Note #1: The footprint you’re not tracking.

Speaker one reframed AI not as a software product, but as an energy problem. Every query, every generated image, every “make it more punchy” prompt — there’s an electricity cost behind it. And when AI reaches full-depth penetration across industries and geographies, the countries leading the race — the US and China — won’t just be competing for chips. They’ll be competing for power grids.

The real insight wasn’t doom. It was this: alongside the companies building AI products, an entirely new layer of infrastructure businesses will emerge — companies that don’t build models but keep them running. Cooling, power, logistics.

The sharpest line of the evening came during the Q&A. Someone pointed out that the next unicorn might not be an AI company at all. It might be whoever solves energy-efficient cooling for data centers. Not the brain. The body temperature.

Note #2: What do AI and crystals have in common?

Speaker two opened with this question and I genuinely didn’t know where it was going.

Here’s where: when you compress an AI model’s weights to make it smaller and faster, the standard method maps data points onto a uniform grid — think of it as a flat, square lattice. Simple, but wasteful. A lot of bit-strings end up representing points that don’t even exist in the actual data distribution.

The alternative? Lattice vector quantization — arranging data points in crystal-like geometric structures (hexagonal, or even 8-dimensional lattices like the E8 Gosset lattice). It’s the same principle nature uses to pack atoms: a crystal is denser and more efficient than a cube. Applied to AI, this means smaller models with less quality loss. There are multiple NeurIPS 2025 papers on this exact approach — and it’s part of a bigger push toward on-device AI: models that run autonomously on your phone or laptop without calling a server.

The endgame isn’t AI in the cloud. It’s AI in your pocket.

Note #3: Read the papers before you read the headlines.

The final speaker delivered something between a talk and a stand-up set. The core message: the most transformative AI breakthroughs are rooted in research papers, not product launches. Transformers started as a paper. Diffusion models started as a paper. RLHF started as a paper. The distance between a white paper on arXiv and a feature in your favorite app is shorter than most people think — and longer than most founders admit.

The argument: before you take AI to the masses, take it to the science first. Understand what’s proven, what’s speculative, and what’s marketing dressed as innovation.

Post-event note (written at 1am):

That last point hit personally. I walked home and immediately fell into a rabbit hole on a topic that’s been pulling at me for weeks: the Creativity Gap in modern LLMs. There’s a growing body of academic research — from NeurIPS best papers to Science Advances to CHI — showing that AI models are getting smarter but also narrower. They raise the floor for average creative output while compressing the ceiling. The implications for everyone in the creative industry are massive. I’m reading through 15+ papers on this right now and will publish a full breakdown soon.

The bigger thought:

It’s Friday night. I could’ve gone to a concert or a movie. Instead, I walked to a coworking space and left three hours later with a head full of lattice structures and energy footprints.

I genuinely believe the quality of a city is no longer measured by its nightlife alone — but by whether you can walk out your door on a Friday and stumble into a room full of people building things.

Valencia is becoming that city. And the fact that the ratio of builders to consumers in rooms like these keeps growing is, to me, the most optimistic signal of 2026.

More notes from the next session soon.

Sergey Vovk is the founder of VOVK (Creative) Consulting and co-founder of Kyiv Academy of Media Arts (KAMA). He writes about what happens when the creative industry learns to speak machine at aibubbledotcom.com.

Before we start: this is not a prescription

Let me save you the suspense. No book on this list will make you creative. Not a single one.

I know — strange opener for a reading list. But after 17 years of building brands, training creative teams, and co-founding an academy dedicated to creative education, I’ve earned the right to say this without it sounding like false modesty: books don’t produce ideas. You do. Books are, at best, the equipment in the gym. You still have to show up. You still have to lift.

What these 20 titles have been for me — at different points in my career — varies wildly. Some were crutches during seasons when I couldn’t generate anything without a framework to lean on. Some were workbooks I literally handed to teams on Monday mornings, with specific chapters marked for the week’s challenge. Some I read once, a decade ago, and can still feel their fingerprints on how I brief a campaign or evaluate a concept.

None of them were a silver bullet. Every single one was a reminder.

A reminder that creativity is not a mood. It’s not a talent you either have or don’t. It’s not the mythical “spark” that self-help culture has been packaging and selling since the 1990s. Creativity, in my operating experience, is three things stacked on top of each other:

A state — something you enter, not something you wait for. Like a musician warming up before a session, or an athlete stretching before a race. There’s a protocol. It’s repeatable. It works even on the days you’d rather do anything else.

A habit slash mindset — the decision to treat your ability to generate and evaluate ideas as a serious professional competency, not a personality trait. Engineers don’t say “I’m just not a debugging person.” Creatives shouldn’t say “I’m not inspired today.”

A rhythm slash discipline — the part nobody wants to hear. The best creatives I’ve worked with — across agencies, across countries, across disciplines — have harder processes than most developers. The difference? Their process isn’t optimized for predictable output. It’s optimized for the maximum probability of an unexpected one. That’s still engineering. Just with a different metric.

But we’ll get into all of that in future editions. For now — here’s the shelf.

This is my personal library on thinking, creativity, and the mechanics of ideas. I compiled it originally for IT professionals who wanted to think beyond their default information field, but it applies to anyone who makes decisions for a living. Which, if you’re reading this newsletter, is you.

Every annotation below is mine — not from the back cover. Every book I’ve either read deeply or used directly in work with teams. Deliberately different genres, deliberately different eras — from 1939 to 2024. Because if you only read people who think like you, you’ll only produce ideas that think like you.

The List

1. The Artist’s Way Julia Cameron • 1992

The book that launched the modern conscious creativity movement. Cameron built a 12-week program to unblock creative potential through “morning pages” and “artist dates.” Core premise: creativity isn’t a gift reserved for the chosen few — it’s a natural function that most people have simply locked behind fear and self-criticism. Works regardless of profession. I’ve seen it land with designers and with database engineers.

2. The Creative Act: A Way of Being Rick Rubin • 2023

The legendary producer — whose artist roster spans from Beastie Boys to Adele to Johnny Cash — wrote a book that isn’t about music at all. It’s about perception. Rubin treats creativity as attention to what already exists around you, not a struggle to squeeze out something new. Reads like a meditative manifesto: minimal advice, maximum shift in how you see.

3. Six Thinking Hats Edward de Bono • 1985

The classic parallel-thinking methodology. De Bono proposes splitting discussion into six modes (”hats”): facts, emotions, criticism, optimism, creativity, process management. Instead of chaotic arguments — a structured rotation through perspectives. The ideal framework for teams where analysts and creatives need to find common language. I’ve deployed this in rooms where people didn’t agree on anything and left with a shared plan

4. The Creative Habit: Learn It and Use It for Life Twyla Tharp • 2003

One of America’s most accomplished choreographers, with 35+ years of practice, proves that creativity is not a flash of inspiration — it’s discipline, ritual, daily practice. 32 concrete exercises for building a creative routine. This book kills the muse myth and replaces it with a system that works for any profession, from choreography to product development

5. Just Kids Patti Smith • 2010

Punk rock legend’s memoir about early years in New York with Robert Mapplethorpe. This isn’t a book about music — it’s a book about how a creative identity forms: through poverty, friendship, and choosing to live for art. National Book Award. Reads like a novel and rewires your understanding of what “being creative” means at its rawest.

6. Creativity, Inc. Ed Catmull • 2014

The Pixar co-founder reveals how to build an organization where creativity isn’t killed by process. Catmull opens the studio that created Toy Story, Finding Nemo, and Up: how to work with failure, how to give honest feedback, how to protect an idea from corporate erosion. If you’re building a product team, this is the manual.

7. Where Good Ideas Come From Steven Johnson • 2010

Johnson traces innovation patterns across centuries and concludes: great ideas rarely arrive as “eureka” moments. They mature over years in a “liquid state,” at the intersection of disciplines and environments. The book for anyone who wants to understand why some ecosystems produce breakthroughs and others produce stagnation.

8. Creativity: A Short and Useful Book John Cleese • 2020

The Monty Python founder wrote a small, witty, and surprisingly deep book about how the creative process works. Cleese distinguishes between “open” and “closed” modes of thinking and explains why most people get stuck in the closed one. In one hour of reading — more practical insight than most 300-page guides deliver in a weekend.

9. Story: Substance, Structure, Style, and the Principles of Screenwriting Robert McKee • 1997

The narrative structure bible that long ago transcended Hollywood. McKee dissects the mechanics of story: conflict, the gap between expectation and reality, character arc. For a product manager, this is a direct framework: every pitch, every user story, every presentation is a narrative. Whoever understands the structure of story controls attention.

10. Creative Confidence David Kelley & Tom Kelley • 2013

The founders of IDEO and Stanford d.school dismantle the myth of the “creative class”: creative confidence is a skill accessible to everyone. Built around design thinking, the book shows how engineers, managers, and scientists can enter creative mode without an arts degree. Dozens of cases from companies that changed products by changing how their people think.

11. A Technique for Producing Ideas James Webb Young • 1939

48 pages, written by an advertising professional in 1939, and still the clearest idea-generation algorithm in existence. Five steps: gather raw material, process it, let go, wait for the “eureka,” bring it to working condition. A micro-book you can read over lunch that will permanently change your approach to brainstorming. The ROI per page is unmatched.

12. Steal Like an Artist Austin Kleon • 2012

Ten things nobody told you about creativity. Kleon argues that nothing original exists — and that this is liberating. Everything new is a remix; the art is in how you combine influences. Short, visual, provocative. The book that makes you want to immediately go make something. Hand it to anyone who says “I’m not creative” and watch what happens.

13. Thinking, Fast and Slow Daniel Kahneman • 2011

The Nobel laureate explains two thinking systems: fast intuitive (System 1) and slow analytical (System 2). For creativity, this is the foundation: understanding your own cognitive biases is the first step to better decisions. If you’re making product decisions, read this before everything else on this list.

14. Originals: How Non-Conformists Move the World Adam Grant • 2016

The Wharton organizational psychologist examines how people generate and champion new ideas. Grant shatters the stereotype: original thinkers don’t take reckless risks — they strategically manage uncertainty. A book about defending unconventional ideas in organizations where conformity rules. Especially valuable for anyone working inside corporate structures and wondering why their best ideas die in committee.

15. Creativity: Flow and the Psychology of Discovery and Invention Mihaly Csikszentmihalyi • 1996

The author of the “flow” concept studied 91 exceptional creative individuals — from Nobel laureates to poets — to understand how creative productivity works. Key conclusion: creativity doesn’t live in the genius’s head — it lives in the system of interaction between person, environment, and knowledge domain. The scientific foundation for everything you’ll read about creativity afterward.

16. The Opposable Mind: How Successful Leaders Win Through Integrative Thinking Roger Martin • 2007

The Rotman School of Management dean studied 50+ successful leaders and found a common thread: they don’t choose between two opposing ideas — they hold both simultaneously and synthesize a third, better one. Integrative thinking is the antidote to binary “either/or.” The book for anyone who wants to move beyond trade-offs and start creating options that shouldn’t exist but do.

17. The Disruptors: How 15 Successful Businesses Defied the Norm Sally Percy • 2024

A fresh analysis of 15 companies — from Spotify to Nintendo, from TikTok to A24 — that broke the rules of their industries. Percy, a Forbes business journalist, dissects the anatomy of disruption: why some businesses conform while others deliberately choose risk and redefine the market. Shortlisted for Business Book Awards 2025. Read it to understand the difference between “interesting idea” and actual structural change.

18. Antifragile: Things That Gain from Disorder Nassim Nicholas Taleb • 2012

Taleb introduces the concept of “antifragility” — systems that don’t just withstand stress but get stronger from it. This is the philosophical foundation for thinking about innovation: instead of defending against uncertainty, learn to feed on it. For product managers and entrepreneurs — an operating system upgrade for how you process chaos.

19. Make Every Day Creative Marion Deuchars • 2024

Award-winning illustrator and Royal Designer for Industry, Deuchars created something deceptively simple: 100+ projects and exercises that make creativity tactile, physical, screen-free. Hand printing. Drawing with sticks. Painting with a mop. This isn’t “art instruction” — it’s a manual for rebuilding your relationship with making things by hand. In an era where most creative work happens through a screen and an LLM prompt, this book is a necessary counterweight. Sometimes the fastest way back to your best thinking is through your hands, not your keyboard.

20. The Dialectic of Creativity Hermann Vaske • 2022

For over 30 years, filmmaker and producer Vaske has asked one question to some of the most influential creatives of our time: “Why are you creative?” This book compiles those conversations — with Marina Abramovic, David Bowie, Björk, Zaha Hadid, David Hockney, Jim Jarmusch, Damien Hirst, Yoko Ono, and dozens more. But the real value isn’t the answers. It’s the structure: Vaske maps both the stimuli of creativity (spirituality, sex, money, fear, ambition) and its killers (censorship, bureaucracy, compromise, distraction). And then makes the sharpest observation of all: often, it’s the blockages that make creativity thrive. The threats are the fuel. A dialectical synthesis of opposites. This book pairs perfectly with Taleb’s Antifragile — one gives you the philosophy, the other gives you the human testimony.

The shelf is a starting point, not a finish line

Twenty books. 85 years between the oldest and newest. Advertising, choreography, neuroscience, punk rock, Pixar, crystal-packing lattices of human experience.

But here’s the thing I need to say clearly, because I’ve spent enough years teaching to know the pattern: some of you will read this list and feel a dopamine hit of “I should read all of these.” You’ll screenshot it, save it, maybe order two. And then the list itself becomes the accomplishment. The curator’s trap.

Don’t fall for it.

These are my 20. They reflect my trajectory — 17 years in advertising agencies, a co-founded academy, a consulting practice, and now a master’s degree in AI Product Management where I’m learning to think about creativity from the other side of the stack. Your 20 will be different. Your “textbook” might be a podcast, a conversation with a colleague, a GitHub repo, a failed project you dissected at 2 AM.

The medium doesn’t matter. What matters is whether it changed how you generate and evaluate ideas — not just once, but repeatedly, on command, under pressure, when the brief is bad and the deadline is real.

That’s the test. Not “did it inspire you.” Inspiration is a sugar rush. The question is: did it give you a tool you can reach for on a Tuesday morning when you’ve got nothing?

Creativity is not inspiration. It’s discipline. And the best way to develop that discipline is to steal an hour from scrolling and give it to one of these books — or to whatever your equivalent turns out to be.

More on the system behind the discipline — the protocols, the warm-ups, the frameworks that actually make creative thinking repeatable — coming in future editions.

Sergey Vovk is the founder of VOVK (Creative) Consulting and co-founder of Kyiv Academy of Media Arts (KAMA). He writes about what happens when the creative industry learns to speak machine at aibubbledotcom.com.

I was scrolling a digital nomad group—the kind where people post co-working recommendations and “visa hacks”—when a job listing stopped me cold:

“If you are in Valencia and either (1) have your own application that uses GenAI in its back-end, rather than just using GenAI to develop it, or (2) have developed such complex application and can prove you did it—please DM me for a work opportunity.”

Read it again. Notice the parenthetical: “rather than just using GenAI to develop it.”

That single clause is doing more analytical work than most AI trend reports published this quarter. It draws a line. On one side: people who use AI tools—prompt, generate, automate fragments of their workflow. On the other: people who build with AI—who architect systems where a model is a structural component, not a convenience layer.

This isn’t a technical distinction. It’s a market fault line. And if you’re a creative professional, a strategist, a consultant, or anyone currently navigating the AI landscape from the outside in—the side you’re standing on is about to matter a lot more than your LinkedIn bio suggests.

* * *

1. Two Types of People in AI

Here’s the taxonomy I keep coming back to. It’s crude, but it’s honest.

Category A: Uses AI. You prompt ChatGPT for copy. You generate images in Midjourney. You run meeting transcripts through an AI summarizer. You are a consumer of AI outputs. Your workflow got faster, maybe 20–40% on certain tasks. You tell people you’re “working with AI.” And technically, you are.

Category B: Builds with AI. You’ve integrated a language model into a product’s back-end. You’ve designed an architecture where the model handles specific tasks within a pipeline—classification, extraction, generation, decision support—not as a chatbot, but as a functional component. You’ve dealt with context windows, token costs, latency, hallucination management, and model selection. You didn’t just use an API. You built something around it.

The job post in that nomad group wasn’t looking for Category A. It was explicitly filtering for Category B. And that filter is spreading.

Six months ago, “I use AI daily” was a differentiator on a resume. Today, it’s table stakes—the equivalent of “proficient in Microsoft Office” circa 2009. The market is recalibrating. Fast.

The question is no longer whether you use AI. It’s whether AI uses you—or whether you’ve taught it to do something it couldn’t do without your architecture.

* * *

2. Why Wrappers Die

To understand where this fault line comes from, you need to understand the word that professional AI circles now use as a pejorative: wrapper.

A wrapper is a product that puts a user interface on top of someone else’s model. Chat with your PDF. Summarize your emails. Generate social media posts from a brief. In 2023, this was a legitimate startup play. You could raise seed funding on the back of a GPT-4 integration and a clean UI.

In 2026, the term “wrapper” has become a verdict. Here’s why.

The Sherlocking problem. Every time OpenAI, Google, or Anthropic ships an update, they absorb features that wrappers were selling. Your “summarize this document” tool? It’s now a native feature in the model’s own interface. Your UI wasn’t the product. It was a temporary gap in someone else’s roadmap.

The commoditization of models. The model itself—GPT, Claude, Gemini, Llama, DeepSeek—is increasingly a commodity. Like electricity. You don’t build a business on the fact that you have access to electricity. You build it on what you do with the power.

The interface mismatch. There’s a growing argument in HCI (human-computer interaction) circles that the chat interface is the worst possible way to interact with AI. It places the cognitive load on the user: you have to know what to ask, how to ask it, and how to evaluate the answer. The future isn’t “talk to AI.” It’s AI embedded invisibly into tools you already use—your CRM, your design software, your project management stack—doing work without being asked.

The analogy I keep using with clients: imagine a designer who sells access to Figma. Not a design service. Not a design system. Just a login to Figma with a nicer landing page. That’s what a thin wrapper is. You’re not a business. You’re a reseller of someone else’s capability with a markup and a prayer.

* * *

3. What IS a Business, Then?

If wrappers are dying, what survives? Three things. All of them require going deeper than an API key.

Proprietary data as moat. The model is the engine. Your data is the fuel the engine can’t get anywhere else. A legal-tech tool built on top of GPT is a wrapper. A legal-tech tool trained on 15 years of Spanish contract law precedents, integrated with local court filing systems, and serving lawyers who trust it because it speaks their regulatory language—that’s a product. The model provides reasoning. The data provides irreplaceability.

Spain’s legal system is specific enough that OpenAI won’t build for it any time soon. Ukraine’s defense-tech and OSINT ecosystem runs on operational data that no foundation model will ever have access to. These are “right” wrappers—where local context creates a moat that global platforms can’t cross quickly.

Workflow integration as moat. Being inside the user’s daily process is worth more than being the best model. Notion’s AI isn’t the smartest. But it’s embedded in where people already work, and switching costs are enormous. The product isn’t the AI. The product is the AI plus the context of everything you’ve already built inside the platform.

Trust architecture as moat. This is the one nobody in tech talks about, but every branding professional understands instinctively. In a world where every tool has access to the same underlying models, the differentiator becomes: who do I trust with my business context? Who has earned the right to see my internal data, my strategy documents, my messy first drafts? Trust isn’t a feature you ship. It’s an asset you build over months of demonstrated competence. And it’s the one moat that can’t be copied by a better-funded competitor overnight.

In a world of commoditized intelligence, the moat isn’t the algorithm. It’s the data you’ve earned the right to hold, the workflow you’re embedded in, and the trust you’ve built around both.

* * *

4. Switcher’s Compass: How I Navigate This

I spent more than 17 years building brands, training creative teams, co-founding an academy. Now I’m inside a master’s program in AI Product Management, reading papers on lattice vector quantization and agentic workflows. I am, by every definition, a switcher—someone crossing from one professional world into another.

And the most useful thing I’ve developed in this crossing isn’t a technical skill. It’s a set of questions I ask myself when I encounter any AI product, opportunity, or hype wave. Call it a compass. Here’s what’s on it.

Question 1: If the model provider ships this feature tomorrow, does the product still exist? This is the wrapper test. If OpenAI adding a button kills your startup, you never had a startup. You had a feature request that OpenAI hadn’t prioritized yet.

Question 2: Where’s the data that the model can’t get on its own? Every defensible AI product I’ve seen has a data layer that isn’t in the training set. Industry-specific. Client-specific. Geography-specific. Operationally specific. If the answer is “nowhere—we use the general model,” that’s a red flag.

Question 3: Does this reduce cognitive load, or just shift it? Most AI tools I evaluate don’t eliminate work. They move it. Instead of writing, you’re editing. Instead of researching, you’re verifying. Instead of thinking, you’re prompting. The genuinely valuable products are the ones where the user’s effort goes down, not sideways. They don’t ask you to become a prompt engineer. They ask you to press a button and trust the system.

Question 4: Am I learning, or am I outsourcing my thinking? This one’s personal. I came from an industry where the quality of your thinking was the product. When I use AI to write a first draft, am I getting faster—or am I getting lazier? There’s research now (CHI 2025) showing that AI-assisted creative work can atrophy your independent creative ability over time. The tool makes you better while you’re using it, and worse when you’re not. That’s not a tool. That’s a dependency. I watch for this constantly.

Question 5: Would I bet my reputation on this output? The ultimate quality filter. If a piece of AI-generated work went out to a client with my name on it and no human edit—would I be comfortable? If the answer is no, the AI isn’t ready for that task. And if I’m editing 80% of it anyway, I need to be honest about where the real value is coming from.

These five questions won’t make you a machine learning engineer. They’re not meant to. They’re meant to keep you from being a passive passenger in a market that’s moving faster than most people’s ability to assess it.

* * *

The Fault Line Is a Mirror

Back to that job post.

A year from now, the distinction it draws—between using AI and building with AI—will not be a niche hiring filter in a nomad group. It will be a standard screening question across industries. Not because everyone needs to be an engineer. But because the market will increasingly demand that professionals understand the architecture of the tools shaping their work, not just the interface.

I say this as someone who is still learning the architecture. I’m not writing from the other side of the fault line. I’m writing from the middle of it. Every week I understand a little more about how models work, how products are built around them, how the infrastructure underneath is evolving. And every week I also see how my 17 years in creative strategy give me a lens that pure technologists don’t have—because I know what trust looks like to a client, what brand coherence costs to build, and why “creative AI” is two words that most people haven’t thought about carefully enough.

The fault line isn’t just between two types of professionals. It’s between two attitudes toward the same technology: passive consumption and active construction.

The question isn’t which side you’re on today. It’s whether you’re moving.

Serhey Vovk is the founder of VOVK (Creative) Consulting and co-founder of Kyiv Academy of Media Arts (KAMA). He writes about what happens when the creative industry learns to speak machine at aibubbledotcom.com.

17+ years in the creative industry — agencies, brands, an academy. Now I’m deep inside the AI stack studying product management, and writing about what I’m finding at the intersection.

AIBUBBLEDOTCOM is where creative professionals learn to work with AI — and where tech people learn why “creative AI” is harder than it sounds.

First edition is live. Would love to hear from anyone else navigating this crossing.