Every week I talk with game founders, and the same moment keeps coming up. They have built something, the numbers are not there, and they already know they need to change direction. They do not ask me whether to pivot. They ask how long it will take to figure out what the pivot should be.

That question used to have a brutal answer: months.

I have lived it. In my early founder years, we would spend weeks on a single direction. We read reports, built spreadsheets, talked to a handful of users, and then argued about the little we had found. By the time we had something resembling an answer, we had burned real runway just to learn that we were pointed the wrong way. The pivot itself was usually correct. The cost of finding it was what hurt.

Here is the part I want you to sit with. AI has made being wrong cheaper. Not wrong after a year of production, after you have hired a team, built the game, soft-launched it, and discovered that nobody cares. Wrong before production, when you are still researching, comparing, testing, filtering, and changing your mind.

This is where I think most founders are reading the AI moment wrong. The headline everyone repeats is that you can build faster now. That is true. You can ship a prototype in a weekend. But building faster has a trap inside it. If you are building something nobody wants, building it faster just gets you to failure faster. The need to pivot does not go away because engineering got cheap.

What actually changes is the speed of the pivot. And the pivot was always bottlenecked on one thing: information. How fast can you research a market, slice the data, test an assumption, and see the next move clearly enough to commit? That bottleneck is what broke open.

Here is how I would use it if I were deciding what game to build today. The same method works for a browser game on Poki, a downloadable game on Steam, or a mobile title. The domain does not matter. The method does.

Fire bullets before cannonballs, and bullets just got cheap

There is a concept from Jim Collins and Morten Hansen in Great by Choice that I keep coming back to. I wrote about bullets, then cannonballs back in 2021, and AI has only made the idea sharper. They studied why some companies thrive in chaos while their rivals do not, and they found that the winners fired bullets before cannonballs. A bullet, in their words, is an empirical test that is low-cost, low-risk, and low-distraction. You fire small bullets to learn what actually hits. Only once you have a confirmed hit do you concentrate your resources and fire the cannonball.

Founders have always had one problem with this. Bullets were never as cheap as the theory wanted them to be. A real market test took weeks, so founders fired too few bullets, or they skipped straight to the cannonball on a hunch and called it conviction.

AI changes the economics of the bullet. Here is how I would actually fire them.

1. Fan out the research. Keep the conclusion, not the pile.

Say I am weighing a browser-based idle game on a platform like Poki or CrazyGames, where the audience runs into the tens of millions of monthly players. The old move was one research thread at a time. The new move is to run several at once.

I would ask, in parallel: how saturated is the idle genre on web right now, what mechanics are retaining players, what realistic revenue share and web ad eCPMs look like, how large the addressable audience actually is, and how the top three competitors monetize. Five bullets, fired together, back in minutes.

The skill is not asking more questions. The skill is refusing the wall of text. You want the conclusion, not the pile of raw pages. If what comes back is ten pages of notes, you have done it wrong. You want the one paragraph that tells you whether this direction is alive or dead.

2. Distrust every confident number.

Our industry runs on repeated benchmarks. “Day 1 retention has to be 50 percent.” “Hyper-casual is dead.” “Your CPI needs to be under fifty cents.” These get passed around until they feel like physics.

A lot of them are blog-lore. They were true for one genre, on one platform, in one year, and then everyone kept quoting them. The most useful thing I do with AI is not generate numbers. It is pressure-test them. Where did this benchmark come from? Is it still true in 2026? What was the sample? Make the model trace the claim back to its source and argue against it.

I learned this the hard way watching founders inherit a 2019 hyper-casual benchmark and build a 2026 game against it. The market had already moved. The number they trusted was stale. Do not bet a quarter of your runway on a statistic you have not cross-checked.

3. Make it attack your idea, not flatter it.

The lowest-value way to use AI is to ask it what to build. It will happily agree with you. A model that flatters you is worse than useless because it gives you the emotional feeling of progress without the information.

The high-value prompt is the opposite. Here is my concept: a merge game with a roguelike meta layer. Tell me why it fails. Give me the strongest argument that this audience will not care. Name the three games that already own this space and explain why my twist will not move retention.

Now you are getting somewhere. When the answer comes back with Travel Town and the other merge incumbents, and a clear case that a meta layer rarely fixes a saturated core loop, you may have saved yourself three months. Use AI as your sharpest critic, not your oracle.

4. Run the same question until you realize the filter is wrong.

This is the one that changed how I think. I was helping scope an underserved audience for a game. First pass, I asked for teen players. Too broad, useless. Second pass, I narrowed by genre. Still soft. Third pass, I realized the thing I had been getting wrong was not the search. It was the filter. The right cut was not age at all. It was platform behavior: who plays long sessions on web versus quick sessions on mobile?

Each pass corrected the question, not the answer. And here is what struck me: that is a pivot. A small one, happening in an afternoon, that used to take a research cycle of weeks. The pivot stopped being a dramatic event and became a thing I did four times before lunch. When the cost of changing your mind drops far enough, changing your mind stops being scary and starts being the work.

5. Build something you can actually look at.

You do not make good decisions staring at a chat window. I have AI build the artifact: a comparison table scoring six game concepts on market size, competition, build cost, and retention fit; a one-page dashboard; a rough mockup of the core screen.

The point is not that the artifact is perfect. The point is that it gives your judgment something to push against. The moment I can see six concepts ranked side by side, with the weak ones visibly weak, my feedback gets sharper and more specific. The format you evaluate in changes the quality of the call you make. A founder reviewing a real dashboard gives better notes than the same founder reading a paragraph.

6. Then take it to a human who has shipped.

None of this replaces an operator who has actually launched in your space. AI gives you breadth and a coherent first draft of a strategy. A real human catches the thing the model got confidently wrong.

I once ran a full analysis that looked airtight. Then I took it to someone who had shipped web games. In two minutes, he told me the platform would never feature that genre and my eCPM assumption was triple reality. The model did not know the platform politics, the featuring behavior, or the real monetization floor.

That is the division of labor. Use the model for breadth. Use the operator for the load-bearing facts.

The new failure mode is never firing the cannonball

Notice what none of this did. It did not tell me what to build. It made being wrong cheap, and it made the wrong answers show up fast. In a few days, I changed direction several times. Every change was an upgrade because each one ran on more information than the last.

But here is where the Collins discipline still bites. Cheap bullets tempt you to never fire the cannonball. You can research forever now. You can generate a hundred concepts, slice the market a hundred ways, and feel productive the entire time. Ask what Collins would ask: are you on an undisciplined pursuit of more?

At some point, you have to take the confirmed hit and put everything behind it. The information got cheaper. The conviction still has to be yours.

I think about Balatro, sold in the millions and built largely by one person. Or Vampire Survivors, a near-solo team and a price under five dollars. Those games are not examples of AI magically finding the answer. They are reminders that a small team does not need a giant organization behind it when the idea hits. AI does not guarantee that hit, but it gives small teams more shots before they run out of time.

That is the shape of the thing.

Final words

AI did not repeal the hard law of building. If you make something nobody wants, you still have to pivot, the same as founders always have. What it gave you is speed at the exact point that used to be slowest: the finding.

The pivot was never the problem. The time to find the pivot was the problem. That is the part that just got cheap.

If you are sitting on a game that is not working, you no longer have a runway excuse for taking a quarter to decide what comes next. You can fire ten bullets this week. Fire them.

My sleep book, Sleep Again, is now on Spotify. Please give it a listen and leave a rating! Also, I would highly appreciate reviews on Amazon.

Now, to this week’s piece.

For more than a decade, Finland was one of the most admired game development ecosystems in the world.

It was not only because of the companies. It was because of what those companies seemed to prove.

A small country could produce global games companies. A city like Helsinki could compete with San Francisco, London, Tokyo, Seoul, and Shanghai. Small teams could create games that reached hundreds of millions of players. A studio from Finland could generate billions in revenue without becoming a traditional corporate machine.

For a while, it felt like Finland had discovered a new way to build games.

Rovio had Angry Birds. Supercell had Hay Day, Clash of Clans, Boom Beach, Clash Royale, and Brawl Stars. Remedy had Max Payne, Alan Wake, and Control. Housemarque carried the arcade tradition into the modern console era. Small Giant built Empires & Puzzles into a major success. Seriously, Next Games, Fingersoft, RedLynx, Frogmind, Traplight, Futureplay, and others all contributed to the feeling that Finland had become a special place for games.

It was a strange and powerful moment.

The Finnish games industry was small enough that everyone knew everyone. But the ambition was global. The teams were small, but the revenue could be enormous. The culture was creative, technical, and product-led. People talked about games as craft. They talked about quality. They talked about player experience.

And for a time, that was enough.

But the difficult question today is this: what if the model that made Finland successful is no longer the model that the market rewards?

I don’t think the Finnish games industry has collapsed. That would be too dramatic and also inaccurate. There are still strong companies, strong developers, experienced operators, and a lot of knowledge in the ecosystem. Finland is still one of the most important games countries in Europe.

But something has clearly changed.

The momentum has slowed. The startup energy is not what it was. There has not been a new Finnish mobile games breakout in the 2020s comparable to what we saw in the previous decade. The industry is still alive, but the belief that Finland will naturally keep producing new global games companies has become harder to defend.

The easy explanation is that mobile games became harder.

That explanation is partly true. User acquisition became more expensive. Apple’s privacy changes made performance marketing more difficult. The market matured. The top charts became more stable. Live operations became more demanding. Investors became more cautious. The easy distribution window of the early App Store era closed long ago.

But those explanations are not enough.

Because while Finland slowed, other ecosystems accelerated.

The most interesting comparison is Turkey.

Over the past five years, Istanbul has become one of the most exciting mobile gaming ecosystems in the world. Peak Games was acquired by Zynga. Dream Games became a unicorn. Spyke Games raised major funding. A new generation of Turkish studios has followed, especially in puzzle, casual, hybrid casual, and adjacent mobile categories, with countless companies raising hundreds of millions in funding.

Turkey did not rise because mobile games were easy. It rose during the same period when Finnish founders and investors were saying that mobile had become too difficult.

That contrast matters.

It suggests that the problem is not only the market. The problem is also the operating model.

Finland became great at one version of the games business. Turkey appears to be better adapted to the next one.

Industries do not usually fall because people become stupid

Business history is full of industries where the old champion did not become stupid. It became overfit.

Japan did not lose consumer electronics because Japanese engineers forgot quality. Nokia did not lose smartphones because Finnish engineers forgot how to build phones. Premium Western companies did not struggle in China because they forgot how to make good products.

In each case, the market changed the definition of excellence.

That is the lens through which I now see Finnish mobile games.

The issue is not talent. It is not creativity. It is not that Finnish developers suddenly became worse at making games. The issue is that Finland became excellent at a version of mobile games that no longer compounds the way it once did.

This is where the comparison to other industries becomes useful.

Michael Porter’s work on clusters has always been one of the best ways to understand why some regions become unusually strong in a specific industry. A cluster is not just a collection of companies located near each other. It is a system that produces talent, knowledge, ambition, suppliers, investors, role models, habits, and new companies.

But there is another side to this.

A cluster does not only produce companies. It produces norms.

It teaches founders what a good company looks like. It teaches employees what kind of risk is admired. It teaches investors what kind of story sounds fundable. It teaches young people whether entrepreneurship feels like a natural path or an irrational gamble.

Helsinki and Istanbul are both gaming clusters.

But they do not seem to reproduce the same behavior.

Helsinki built a craft cluster.

Istanbul built a competition cluster.

And the current mobile games market rewards the second one more than the first.

Helsinki built a craft culture

The Finnish games industry grew out of a particular kind of developer culture.

It was technical. It was creative. It respected originality. It valued small teams and high autonomy. It did not like unnecessary management. It had a strong belief that great games came from talented teams who cared deeply about what they were making.

This culture produced remarkable things.

It helped create some of the most important games companies in Europe. It showed that a small country could create global entertainment products. It gave young developers a sense that they did not have to leave Finland to build world-class games.

But every industry culture has a hidden cost.

The same values that help an ecosystem win in one era can make it slower to adapt in the next.

The Finnish model was well suited to the first mobile era. That era rewarded teams that could combine product creativity with technical execution. Distribution was still open enough that a great game could break through. A small team with taste, skill, and a strong product instinct could create something globally meaningful.

That era created the myth of the exceptional game.

If the game was good enough, the market would make room for it. If the team was talented enough, originality would win. If the studio cared enough about quality, players would notice.

There is truth in this. Great games still matter. Quality still matters. Product still matters.

But in mature mobile games, quality is no longer enough. In many genres, quality is not even the starting question.

The starting questions are different now.

How large is the market? What are the dominant player behaviors? What is the benchmark CPI? How many ad creatives can we test? What is the expected day-one retention? What monetization model fits this audience? How quickly can we soft launch? What does the LTV curve suggest? How fast can the team respond to data? How much capital is needed to scale?

This is a different game.

It is not a game where craft disappears. But craft becomes one part of a much larger commercial machine.

And this is where Finland has struggled.

The Finnish ecosystem has often celebrated the craft mythology of games more than the commercial machinery behind them. We like the idea of small teams making great products. We like the idea of originality. We like the idea that player love creates business outcomes.

But the mature mobile market often works the other way around.

The business opportunity defines the product. The genre defines the constraints. The user acquisition environment defines the speed. The market data shapes the roadmap. The company is not built around the dream of a game. The game is built inside a commercial system.

For many Finnish developers, this feels uninspiring.

For many Turkish founders, it seems to feel obvious.

Istanbul built a challenger culture

One of the useful business comparisons here is the way local companies in emerging markets have competed against global multinationals.

The old assumption was that the advanced companies from rich markets had the best products, the best management practices, and the strongest brands. Local companies in emerging markets were expected to copy them, partner with them, or eventually lose to them.

But that is not always what happened.

In many industries, local challengers won because they understood the real market better. They did not compete on the old prestige dimensions. They competed on adaptation, speed, cost structure, distribution, customer understanding, and willingness to build around the actual opportunity in front of them.

This is how I now think about Istanbul.

The Turkish gaming ecosystem is not simply trying to become a second Helsinki. It is not trying to recreate the Finnish mobile games mythology from 2010. It is behaving like a challenger ecosystem.

It studies large markets. It follows money. It understands performance marketing. It enters proven categories. It moves quickly. It accepts competition. It treats gaming as a business opportunity, not only as a creative calling.

This does not mean Turkish games lack quality. Dream Games’ Royal Match is a highly polished product. Peak Games built world-class puzzle games. The best Turkish studios are not making careless products. They are making extremely disciplined products.

But the starting point is different.

The Finnish question has often been: what kind of game do we want to make?

The Turkish question seems closer to: where is the opportunity, and what kind of company do we need to build to capture it?

That difference sounds small. It is not.

It changes everything.

It changes the founder profile. It changes the team culture. It changes how quickly companies go to market. It changes what kind of talent is valued. It changes how disagreement happens inside teams. It changes how investors evaluate opportunities. It changes whether people see a familiar genre as boring or as a large market to attack.

In Istanbul, gaming is often understood as competitive commerce.

A founder can look at puzzle games, casual games, hybrid casual games, and say: this is a huge market, there is money here, there are known behaviors, there are benchmarks, and we can compete.

In Helsinki, many developers still carry a more romantic view. Games should come from personal passion. A team should care deeply about the world, mechanics, feel, and originality. Building within a proven market category can feel derivative. Following the data too closely can feel like selling out.

This mindset can produce wonderful games. But it may not produce the companies that win in today’s mobile market.

The Turkish ecosystem seems to have built a different norm. It is younger, hungrier, more commercially direct, and more willing to compete in crowded genres. It does not seem embarrassed by the business side of games. It does not treat paid marketing as something to figure out after the product is done. It does not seem to believe that a game deserves to exist because the team loves it.

The market has to prove it.

This is the critical difference.

Finland produced a generation of developers who believed in product excellence. Turkey is producing founders who believe in market discipline.

The market changed what excellence means

Another useful comparison comes from the way premium Western and Japanese companies were challenged in China.

For years, many incumbents believed that their advantage came from superior quality. They had better engineering, stronger brands, and more sophisticated products. But in large emerging markets, many customers did not want the most premium version of a product. They wanted something reliable enough, affordable enough, accessible enough, and better adapted to their actual needs.

The challengers understood the market’s real criteria. The incumbents kept optimizing for their own definition of excellence.

This is uncomfortable because it also applies to mobile games.

The analogy is not that Turkish games are low quality. They are not. The point is that mature markets often stop rewarding the old definition of excellence.

Incumbents keep polishing the dimensions they admire. Challengers optimize the dimensions the market actually pays for.

In mobile games, the market may not reward originality in the same way Finnish developers want it to. It may reward familiarity, clarity, fast onboarding, scalable acquisition, proven monetization, strong creative testing, live operations, and relentless optimization.

A Finnish developer might look at a puzzle game and ask: where is the originality?

A Turkish founder might look at the same category and ask: where is the inefficiency?

That is the difference.

The mature mobile market does not always reward the most original product. It often rewards the company that understands a large existing behavior pattern and executes against it better than others.

This is especially true in casual and puzzle categories. These genres look simple from the outside. They are not. They require deep understanding of player psychology, level design, monetization, creative testing, difficulty curves, retention, and user acquisition. The products may appear familiar, but the operating complexity is enormous.

Turkey has built strength in exactly this kind of operating complexity.

Finland, by contrast, has often preferred more original, midcore, or craft-led directions. That can produce excellent games, but it is also more capital-intensive, slower, and riskier. If the game fails, there may be less reusable learning. If the game takes years before market validation, the company may not have time to adapt.

The Turkish model is not necessarily more artistic. But it may be more adaptive.

And in a mature market, adaptability is often more valuable than originality.

The new model may come from the younger ecosystem

There is another business pattern worth considering: reverse innovation.

The old assumption is that mature, wealthy markets create the advanced model, and emerging markets copy it later. But in many industries, innovation has started to move in the opposite direction. Companies build new models in more constrained, more competitive, or more price-sensitive markets, and those models later influence the richer markets.

This is a useful way to think about Turkey.

The old hierarchy would place Helsinki above Istanbul. Helsinki had the global hits, the capital, the prestige, the conferences, the experienced operators, and the long history. Istanbul would be seen as the rising follower.

But in mobile games today, that hierarchy may be outdated.

Istanbul may now be ahead in the operating model that matters most for certain parts of the mobile market.

Not ahead in every dimension. Not ahead in every genre. Not ahead in artistic originality, technical tradition, or the long history of game development.

But ahead in market selection. Ahead in founder hunger. Ahead in performance marketing culture. Ahead in company formation. Ahead in talent recycling. Ahead in the willingness to build around proven demand.

That is why the Turkey comparison should make Finland uncomfortable.

It is not just that another country has produced new successful gaming companies. That happens.

The more important point is that Turkey may have developed a more modern mobile games company-building model.

In other words, Istanbul is not only catching up to Helsinki.

In some parts of mobile gaming, Istanbul may now be teaching Helsinki.

The Nokia lesson is closer than we want to admit

The Nokia comparison is obvious, maybe too obvious. But it is still useful.

Nokia did not lose because Finland ran out of engineers. It did not lose because Finnish people forgot mobile phones. It lost because the basis of competition changed.

The mobile phone market moved from hardware-centric excellence toward software, touch interfaces, app ecosystems, consumer experience, and platform control. Nokia had enormous strengths. But many of those strengths were built for the previous era.

This is the uncomfortable mirror for Finnish games.

Finnish mobile games are not struggling because Finland ran out of talented developers. They are struggling because the basis of competition moved from creative product invention toward commercial operating excellence.

The old advantage was small-team creative excellence in a growing global platform window.

The new advantage is commercial learning speed in a mature, crowded, performance-driven market.

Those are not the same capability.

A Finnish studio can still produce a brilliant game. But the question is whether it can produce a brilliant mobile games company under current conditions.

That requires more than design and engineering. It requires market selection, capital discipline, paid marketing, creative production, live operations, analytics, rapid decision-making, and a culture that is comfortable killing ideas not because they are bad, but because they are commercially insufficient.

This is where the craft mindset struggles.

A craft culture asks: is the game good?

A commercial learning culture asks: is this opportunity worth scaling?

Both questions matter. But they are not the same question.

Supercell became both inspiration and distortion

No discussion of Finnish mobile games can avoid Supercell.

Supercell is the best thing that ever happened to the Finnish games industry. It created global credibility, wealth, talent, ambition, investors, alumni, and belief. It showed that a Finnish games company could become one of the most successful entertainment companies in the world.

But every great success story also creates distortion.

Supercell became the ultimate expression of the Finnish craft mythology. Small teams. High autonomy. Kill the games that are not good enough. Build for the long term. Trust talented people. Avoid bureaucracy. Let the teams decide.

This was a profound contribution to the global games industry.

But it also created an ideal that was almost impossible to copy.

Many founders took the visible parts of the model without having the full machine behind it. They saw small teams, autonomy, and high standards. But they did not necessarily have the same talent density, capital base, marketing capability, organizational discipline, or hit-making judgment.

The mythology of Supercell may have made some Finnish founders believe that the highest form of a games company is a small autonomous craft team searching for greatness.

But Supercell was never only that.

It was also an incredibly sophisticated commercial company. It understood markets, scale, monetization, live operations, talent, and capital allocation at a level very few companies ever reach.

The Finnish ecosystem often remembered the craft part more than the commercial part.

Istanbul seems to have remembered something else from its own success stories.

Peak Games and Dream Games created a different kind of founder imagination. They showed that you could enter huge existing categories, execute with extreme discipline, and build enormous companies. They created wealth, talent, and ambition. They also created a founder mafia, a network of people leaving successful companies to start the next wave.

That matters enormously.

When employees leave successful companies to start new ventures, a cluster compounds. Knowledge moves. Capital moves. Confidence moves. Teams form faster. The next company starts with a higher baseline of competence.

In Turkey, this appears to be happening repeatedly.

In Finland, less so.

There are many reasons. Finland has an older and more mature industry. Senior people have families, mortgages, high salaries, and stable jobs. The risk-reward calculation is different. The ecosystem is comfortable. The cost of leaving is higher.

But comfort is dangerous in a startup ecosystem.

A cluster that stops producing hungry new founders eventually becomes a talent pool for existing companies, not a launchpad for new ones.

Government support can reduce urgency

One of the more uncomfortable differences is the role of public funding.

Finland has been very generous in supporting early-stage technology companies. Business Finland and earlier public funding instruments have played an important role in the development of the Finnish games industry. Many companies benefited from this support, and the ecosystem would probably be weaker without it.

But every funding model creates behavior.

If early public money is available before strong market validation, it can extend the life of companies that have not yet earned the right to scale. It can give founders time, but it can also reduce urgency. It can allow teams to spend longer in product development before facing the market.

This matters especially in mobile games.

A team can spend two years building something beautiful and only then discover that the CPI does not work, the retention is not strong enough, the audience is too small, or the genre is impossible to scale. By then, the company has used its energy, time, and capital.

In Turkey, public support has often been more connected to later-stage scaling and market activity. The exact mechanisms differ, but the cultural effect appears different. Founders are pushed toward the market earlier. They have to prove more before support becomes meaningful.

That changes the rhythm of company building.

The Finnish rhythm can become: fund, build, polish, launch, learn.

The Turkish rhythm seems closer to: test, prove, scale, optimize.

In mobile gaming, the second rhythm is often stronger.

This does not mean Finland should remove public funding. That would be too simplistic. But the design of support matters. Funding should increase commercial learning, not delay it. It should push founders closer to the market, not protect them from it.

The best support would force earlier confrontation with reality.

The passion trap

I have spent much of my career in and around games, and I still believe that passion matters. Games are hard to make. Players can feel when a team cares. A purely cynical game company will eventually produce lifeless products.

But passion can also become a trap.

Especially in Finland.

We often talk about games as something that should begin with love for games. Founders should care about what they are making. Developers should be players themselves. The team should be proud of the product.

These are good instincts.

But they can become limiting if they create contempt for market-driven opportunities.

In Helsinki, gaming is often perceived as a passion industry. In Istanbul, it seems more often perceived as a business opportunity.

That is a profound difference.

If games are a passion industry, then making a familiar puzzle game can feel like compromise. If games are a business opportunity, then making a familiar puzzle game can feel like intelligent entrepreneurship.

If games are a passion industry, then the founder’s taste matters enormously. If games are a business opportunity, then the market’s behavior matters more.

If games are a passion industry, then spending years polishing a product can feel responsible. If games are a business opportunity, then delaying market feedback can feel irresponsible.

This is not about one culture being better in every way. The Finnish passion for craft has created world-changing games. The Turkish hunger for business opportunity can also create risks: crowded genres, copycat thinking, short-termism, pressure, burnout, and too much dependence on paid marketing.

But right now, in mobile games, the Turkish mindset appears better matched to the market.

That is the hard truth.

What Finland should learn from Turkey

The wrong lesson would be that Finland should become Turkey.

It cannot. It should not. Ecosystems cannot copy culture that easily. Finland has its own strengths: trust, technical skill, design taste, stability, education, global reputation, and a long history in games. Those are real advantages.

The right lesson is that Finland must update its operating model.

First, Finnish founders need to confront the market earlier.

A mobile games company should not spend years building before it understands distribution. Genre selection, CPI assumptions, creative testing, retention benchmarks, and monetization potential must be part of the company from the beginning. Marketing is not something that happens after the game is ready. In mature mobile, marketing is part of product discovery.

Second, Finland needs more founder recycling.

Experienced people need to leave comfortable jobs and start companies. Not everyone should do this, of course. But a startup ecosystem depends on ambitious talent taking risk. If the best people remain inside established companies, the ecosystem becomes stable but less dynamic.

Third, Finnish games companies need to treat business innovation as real innovation.

There is still too much prestige attached to originality and not enough prestige attached to distribution insight, monetization design, creative testing, and operational speed. In today’s market, these are not secondary skills. They are core product capabilities.

Fourth, public funding should be designed around market learning.

Support should encourage teams to test faster, validate earlier, and become commercially sharper. It should not make it easier to avoid the hardest questions.

Fifth, Finland needs to be honest about what kind of games industry it wants to build next.

Maybe Finland’s next wave is not in casual mobile. Maybe it is in PC and console. Maybe it is in premium, simulation, strategy, tools, AI-assisted production, UGC infrastructure, or new forms of interactive entertainment. That is fine.

But if Finland wants to compete in mobile games again, it has to accept the reality of mobile games today.

The market is not waiting for a beautiful Finnish product to appear.

The market is being attacked every day by companies that move faster, test harder, spend smarter, and care less about whether the opportunity feels creatively noble.

The end of the craft era

I do not believe the Finnish games industry is finished.

But I do believe one era is over.

The era when Finland could rely on craft, originality, small-team mythology, and the afterglow of past successes is over. The global market has moved on. Mobile games have become a more mature, more competitive, more operationally demanding business.

Turkey’s rise makes this impossible to ignore.

If mobile had simply become impossible, then Istanbul would not be producing new winners. If the market had closed completely, Turkish founders would not be raising major funding and building companies in the same period when Helsinki has struggled to create new breakouts.

The difference is not that Turkey found an easier market.

The difference is that Turkey built companies better suited to the harder market.

That should make Finland uncomfortable. But it should also make the path clearer.

The future of Finnish games will not come from nostalgia. It will not come from waiting for the next Supercell. It will not come from telling ourselves that Finnish developers are uniquely creative and that the world will eventually reward quality again.

Quality matters. But quality alone is not a strategy.

The next Finnish games industry has to decide what it is optimizing for.

If it wants to build beautiful games, it can do that. If it wants to build sustainable creative studios, it can do that. If it wants to build premium PC and console companies, it can do that. These are all valid futures.

But if it wants to create the next generation of global mobile games companies, it needs to become more commercially aggressive, more market-led, more comfortable with competition, and much faster at learning.

Helsinki built a craft culture. Istanbul built a competition culture.

For the first mobile era, Helsinki’s culture was enough to create miracles.

For the next era, it may not be.

If you got my latest book, Sleep Again, already and read it, I would appreciate a review on Amazon! The Kindle version is still $3.99 for a few days, get your copy!

I was recently listening to Naval Ravikant talk about vibe coding.

I built my own app store. So if I want an app, I literally open up Claude on my phone. I can operate a remote terminal which is running on my desktop, or I can just use Claude in the cloud. It can connect to Xcode. I give it a two-line description. It builds me an app. It ships it to my app store. I open my app store app. The app is sitting there. I click install. 30 seconds later, I have a working app on my phone. That’s magical. You can literally be at dinner with someone having a conversation. They describe some app they want. You can describe it to Claude and five minutes later, you’re showing them that app on your phone.

Listen on Youtube or Spotify.

Something clicked for me.

Last year, when I got into vibe coding with Claude Code, I still had the old app economy in my head. I was thinking about consumer apps. Build something. Launch it. Get users. Run paid ads. Compete in the App Store.

That model still works. But I’m starting to think there is another model coming.

The old way

The old model is simple.

You build one app for many people.

That means you need onboarding. You need logins. You need analytics. You need pricing. You need customer support. You need distribution. You need the product to make sense for thousands or millions of users.

This is why so many utility apps become bloated. A workout app tries to support every kind of lifter. A note-taking app tries to support every kind of thinker. A screen-time app tries to support every kind of family.

The app has to generalize.

The new way

With vibe coding, the app doesn’t need to generalize.

It can be built for one person. You.

That changes the whole equation. The app doesn’t need a growth strategy. It doesn’t need a monetization layer. It doesn’t need to support edge cases from thousands of users.

It only needs to solve your problem. This is where I think the shift will happen. We’ll move from external apps to internal apps.

External apps are the ones we download from the App Store.

Internal apps are the ones we build for ourselves.

My first assumption was wrong

A few weeks ago, I still had an old assumption in my head.

I thought that if I built a mobile app for myself, it would take a few weeks.

That was true for the first ones, especially the ones I built in 2025.

But the timeline keeps collapsing. Now I can start something in the afternoon and have a working version the same day.

That matters. When something takes three weeks, you ask whether it’s worth building. When something takes one evening, you just build it.

LiftMate

The first example is my workout tracking app, LiftMate.

I’ve had a shoulder that used to dislocate. It was operated on, but it still affects what I should do in the gym. I’ve had the same issue with my knee. It was also operated on. It works, but it’s not a perfect knee anymore.

Most workout apps don’t know this. They give me programs that assume my body is generic.

LiftMate doesn’t do that. It knows my constraints. It can rule out movements that aren’t good for me. It can suggest alternatives. It can keep the plan focused on what I can actually do.

I don’t need a login system. I don’t need analytics. I don’t need a privacy policy written for millions of users.

It’s my app. It stores my data. I improve it when I need to.

Earned

The second app is called Earned.

I had done an online course on attachment issues. I realized I had lived with anxious attachment for a long time, and that it had affected my relationships and how I lived.

So I built an app around that. The goal is simple: move from anxious attachment toward earned secure attachment.

The app gives me exercises, reflections, and prompts. It helps me work on the thing I’m actually trying to fix.

This is where personal software becomes interesting. As I change, the app can change.

I don’t need to build a therapy product for a market. I need a system that supports me at the stage I’m in now. Then, when I gradually move away from being anxiously attached, I can modify the app to cater to where I’m on the spectrum of moving from anxious to securely attached.

My family screen-time app

This weekend, I built a screen-time app for my kids.

The existing products try to do too much and they are good at asking for more money when your kids have many devices.

I only needed four things.

1. Lock the device.

2. See the child’s location on a map.

3. Lock specific apps on demand.

4. Start with Android and iOS, later add Windows and Playstation.

That’s it. So I built that.

If we need another feature later, I’ll go back to Claude Code, describe what we need, and Claude pushes a new version to TestFlight.

This is where it became clear to me.

TestFlight is becoming my personal App Store.

My own DocuSign

Today I started building my own DocuSign for my own use cases.

After two hours, the first version was already running.

I don’t need the full DocuSign product. I don’t need the enterprise version. I don’t need all the features built for every company in the world. I don’t need to pay for anything.

I need a signing flow that works for the types of documents I send. Starting with a few types, adding more later if needed.

That is a much smaller problem. And smaller problems are now easy to turn into software.

The cost structure is different

The cost of all the services is also close to zero.

There are now many backend services with free tiers that are more than enough for personal use. Like free databases. Free email-sending tools. Free hosting tiers.

Here’s the tools that Claude Code suggested for free tier use:

1. Resend for free email sending (100 a month for free)

2. OpenFreeMaps for mobile app maps

3. Vercel for backend hosting (extremely generous free tier)

4. Turso for SQL database hosting (very generous free tier)

The free tiers are more than enough. You are not building for scale. You are building for personal utility.

That removes a lot of the old “startup” logic. No fundraising. No go-to-market. No paid ads. No growth loops.

Just a problem, an interface, and an evening of building.

Where this goes next

Right now, this works especially well for utility apps.

Workout tracking. Notes. Family tools. Document flows. Personal therapy. Small internal systems.

But I think it will go further. Learning is the obvious next area.

You could build a personal learning app for one concept. A podcast app that understands what you are trying to learn. A YouTube interface that organizes videos around your goals. A reading system that connects to your notes, emails, and past questions.

The AI already has context.

It knows what you ask about. It knows what you’re trying to build. It knows what you’re worried about. If you give it access, it can know your notes, emails, calendar, documents, and second brain.

And you can update the app as you progress to be more challenging in the learning content.

So why would the software stay generic?

The caveat

This is still early.

If you’re not vibe coding, this probably sounds far away. You still go to the App Store. You still download the best available app. You still adapt your workflow to someone else’s product.

But the interface is getting easier. At some point, building an app will feel close to talking to ChatGPT.

You describe what you need. The AI builds it. You test it. You ask for changes. It ships the next version.

When that happens, the default changes. You won’t always ask, “Which app should I download?”

You’ll ask, “Should I just build this for myself?”

Final words

The App Store isn’t going away. But its role will change, and many people will stop treating it as the default place to look for solutions.

In many categories, especially utility apps, the best app for you won’t be the one with the most downloads. It will be the one that understands your specific needs and context.

In the old model, software was built outside your life and you adapted to it. In the new model, software gets built from inside your life.

That’s the part I find exciting.

My Sleep book is out! Get your copy from Sleepagain.co.

And now for this week’s piece.

In January 2026, Steam updated its AI disclosure form. Developers now answer three questions instead of one: does your game use generative AI at all, does it pre-generate content that ships with the game, and does it live-generate content during gameplay. Coding helpers are exempt. The focus is on what the player actually sees.

One in five new Steam releases now carries an AI disclosure. That number will keep rising.

I run an AI game development course, and one of the lessons walks through level design specifically. The Steam policy maps almost exactly to the three modes I teach: pre-generated levels, real-time levels, and player-generated levels. Each mode has a different relationship with AI, and each has a different answer to the same underlying question.

That question is: can AI design fun levels?

Fun is pattern recognition

Raph Koster gave the clearest answer I’ve read in A Theory of Fun for Game Design. His argument is that fun is the brain’s reward for learning. Humans are pattern-recognition machines. Games present patterns to master, and when your brain is learning, it releases a reward signal. That signal is what we call “fun.”

AI can handle patterns. It can build rule systems, calculate constraints, and validate that a level is solvable. What it cannot do is feel whether the learning curve is right. It cannot play your level and know it’s boring at minute eight. That part still requires a human.

This split, math on one side and creativity on the other, is why the three modes of level design behave so differently in practice.

Pre-generated levels

These are the levels the designers build during development and ship inside the game. There is a linear path, difficulty spikes, and new mechanics that get introduced as the player levels up in skill and gear. AI is extremely useful here. You can generate dozens of level variants that satisfy a constraint set, validate them for solvability, and let a human designer pick the ones that feel right.

The key insight is that AI does the math. The human still does the judging. If you skip the human pass, you ship boring levels that technically work.

Real-time levels

These get generated while the player plays. The game reads what the player is doing and builds the next room, the next wave, the next puzzle on the fly. In theory this is personalization. In practice it is hard, because AI does not know what “fun” looks like for this specific player in this specific moment.

Most real-time level generation that works well today is not pure AI. It is a rules engine with AI assistance. The game has a library of hand-tuned building blocks and uses AI to arrange them. Same pattern again: AI does the combinatorics, humans designed the blocks.

Player-generated levels

This is the mode I find most interesting. Players have always wanted to build their own levels. The problem is that level design tools are cumbersome. You need to learn the editor, understand the constraints, and spend hours iterating.

With AI, players can describe what they want in a prompt. “A desert map with three choke points and a hidden shortcut through the canyon.” The AI builds a first pass. The player tweaks it. The cycle that used to take two hours takes ten minutes.

This lowers the barrier for creativity without removing the creativity. The player is still the one reading whether the map is fun, because they play it.

The crossword experiment

The clearest example I have of the math-versus-creativity split is outside of games. It is a crossword puzzle project I’ve been running since summer 2025.

I wanted a 5x7 grid with clues, valid words, and proper crossings. I tested the same prompt against ChatGPT, Gemini, and Claude. All three struggled. In Finnish, they invented words that do not exist. In English, they produced grids with mismatched letters at the crossings. Single prompts were never going to solve it.

In the fall, I built three separate projects pointed at how professional crossword constructors actually work, step by step. None of them produced a valid puzzle.

The thing that eventually worked is Claude Code with a set of guard rail scripts. The scripts do the deterministic work: check the grid, flag duplicate clues, enforce word-length constraints, reject words I’ve already used in earlier puzzles. Claude does the search and the filling. When Claude tries to take a shortcut, the script catches it and tells Claude exactly what to fix. Claude Code became the software operator, not just a code generator.

That framework has now built 140 puzzles for a couple’s app I’m working on. Each puzzle takes 10 to 15 minutes. I usually kick off a run and go do something else. The grids are valid. The words are real. The crossings hold.

But the clues still need a human pass. The AI will happily ship “COCOA” clued as “like cola” because letter-fitting is satisfied. I read every puzzle, mark the weak clues, and paste the fixes back into Claude Code, which updates the grid. That step is not optional. It is the creative judgment that the scripts and the model cannot do.

Level design is the same shape of problem. Parts of it are math. Parts of it are creativity. The modes differ in where the line sits.

What this means if you’re shipping

If you are building a game right now and thinking about where AI fits, the shortest version is this. Use AI for the math. Keep the creative judgment for yourself.

For pre-generated levels, generate many candidates and hand-pick the ones that feel right. For real-time levels, design the building blocks yourself and let AI arrange them under constraints. For player-generated levels, give players a prompt interface and let the AI handle the tool complexity.

Today is the day my sleep book comes out. You can buy it on my website or on Amazon. It’s available as an audiobook, paperback, hardcover and ebook.

I started writing this book in 2024 after a year of digging into all sorts of things to improve my sleep. When we started our venture fund in 2023, I noticed that my sleep started getting really bad. It was an extremely exciting time; I had never felt so excited in my entire life before about starting a new company like this, but at the same time the daunting unknowns were hitting me in a way that they hadn’t done before. I believe partially it is due to my burnout in 2019 that I had developed a more sensitive nervous system for stress, which meant that when stress was high it was higher than previously whenever I had been a founder, like during the times when I started my previous two games companies, Ironstar Helsinki and Next Games.

There were many realizations along that journey of understanding how to better my sleep. From sleeping environments to wind-down routines to supplements, apps, mattresses, and earplugs.

When I look back now on what I wrote in the book and all the experiences, I do feel that good sleep is then achieved by paying a price of leading a less stressful life. Now that I’m back in the entrepreneur’s role after several years building my AI startup, I feel that I had to make the decision of constant good sleep versus the startup, and I have now picked the startup. This book tells you that you can make those choices. But there’s also the possibility that you’re the person who can get enough benefits to enjoy being a crazy entrepreneur and still sleep well.

How I used AI to publish my book

I finished my sleep book manuscript early this year. I had 44,000 words and it was a lot of work ahead to get this out. My first book, which came out in 2020, was something where I hired at least three people to help me out with different self-publishing tasks which were very time-consuming. This time around I felt that we have AI capabilities to do a lot of the heavy lifting which comes after the manuscript is ready. Here’s a list of things that I applied with AI to publish my book.

I used Claude Code for all of this. I had the ideas, wrote the prompts and it implemented it the best way it could.

1. Markdown as the Foundation

The first decision I made was to convert the entire manuscript into plain text markdown files, one per chapter, managed in Obsidian. No Word documents, no Google Docs, no proprietary formats. Markdown is simple enough that an AI agent can read, search, and edit it without any compatibility issues. Every chapter, every endnote, every appendix lives in its own file. That meant Claude Code could operate on any part of the book at any time without me having to copy and paste text back and forth. It sounds like a small thing, but it made everything else on this list possible.

2. EPUB Creation (ebook)

I didn’t use Vellum or Atticus or any of the standard book formatting tools. Claude Code built the entire ebook pipeline from scratch. It wrote a preprocessor that converts my chapter files into Kindle-compatible HTML, handling epigraphs, section breaks, drop caps, and a custom table of contents. It took 12 rounds of iteration to get right. Kindle strips CSS classes, ignores small-caps, overrides text alignment on paragraph tags. Each of those was a dead end that Claude Code hit, diagnosed, and rewrote around. The final output is a 163KB EPUB with embedded fonts and bidirectional endnote links.

3. Print Interior

The print interior went through four proof cycles. The first version looked fine on screen but the body text was rendering at 9.68pt instead of 11pt in print, because neither Claude nor I knew that we had done something wrong: Chrome’s headless PDF generator, which we were using, was silently scaling the content to fit the page. Claude Code found the bug, constrained the viewport width to the exact content area, and regenerated. Every time the page count changed, it recalculated the spine width, updated both cover templates, and produced new PDFs. The final interior is 277 pages with industry-standard margins.

4. Audiobook

The audiobook was generated using ElevenLabs text-to-speech. Claude Code managed the entire production. It split each chapter into chunks, generated dual takes for each one, ran spectral analysis to pick the cleaner generation, then verified the output against the source text using Whisper transcription. If a chunk had artifacts or word errors, it flagged it and regenerated. The final audiobook is 30 tracks, just under 5 hours. It’s included in the website bundle.

5. Cover Design

I started with an AI image for the book cover from Google’s Nano Banana Pro and spent weeks dialing in the typography, based on what Nano Banana had picked for the book cover, since gen AI doesn’t really tell you that this is the font to use.

Claude Code built a 54-font specimen page so I could compare capital letterforms side by side. I picked Cormorant for its flared serifs on the A and its thick-thin contrast. From there it built HTML-to-PDF cover templates for KDP paperback and Lulu hardcover, each with precise bleed and spine calculations. Every time the page count shifted, the spine width changed, and the covers had to be regenerated. By the end I had print-ready PDFs for three formats: paperback, hardcover case wrap, and ebook.

6. Consistency Audit

Before locking the text, I needed to know whether the book contradicted itself. I felt like the text was fine and had read it multiple times, but I still had this nagging feeling that maybe there was too much repetition, or inconsistencies I was blind to.

Claude Code extracted 746 factual claims from all 33 files, cross-referenced them, and flagged 38 potential issues. After verification, 18 turned out to be real. Three were critical: a market size number that didn’t match its source, an alcohol timing recommendation that conflicted between chapters, and a heart rate threshold that was off. The rest were moderate things like meal timing inconsistencies and duplicate paragraphs. All fixed before text lock.

7. Endnote System

The book has 87 endnotes. I had most of the citations listed already, but I wasn’t confident I’d caught every claim that needed a source. Claude Code went through each chapter systematically, found the gaps, matched claims to the right studies, and wrote full academic-style citations for the ones I’d missed. It numbered them sequentially across the entire book, built a master mapping document, and added bidirectional hyperlinks in the EPUB so readers can tap a superscript to jump to the endnote and tap back to return.

8. Website

Sleepagain.co is a static HTML/CSS site. Claude Code built all of it. The homepage, the about page, the chapter list, the FAQ, and a podcast guest landing page. It handled all the technical plumbing to make the site discoverable by search engines and AI: structured data, social card previews, a sitemap, and crawler permissions. Later it added a free chapter lead magnet funnel, a Lemon Squeezy checkout for launch day, and a launch-day branch that I’ll merge on April 9.

9. TikTok and Instagram Carousels

I wanted to promote the book on TikTok and Instagram using carousel posts, the swipeable slide format that works well for short, visual content. I picked the topics, the humor styles, and the visual direction for each one. Claude Code handled the rest. It built the slides as HTML, screenshotted them in the right aspect ratios for each platform, and scheduled the posts via API. Over the course of a month we produced more than 30 carousels, over 240 slides total. I couldn’t have done that volume manually in the time I had. See on of the carousels here.

10. Sales Analytics Pipeline

I wanted a single daily KPI email that told me everything: site traffic, subscriber counts, and sales across every platform. Claude Code built it. It wrote a fetcher that pulls data from five APIs every morning: Kit for subscribers, Umami for analytics, Lemon Squeezy for direct sales, Apple Books, and Lulu. For platforms without APIs, like KDP, it built Chrome browser scrapers that log in and extract the numbers. Everything gets pushed to Supabase, and a GitHub Action assembles the daily email at 10 AM Helsinki time. I open my inbox and the full picture is there.

11. Google Drive Backups

I used Obsidian and Github as backups for the project, but I also wanted a third one. So I asked Claude Code to help me out.

I had over 200 files across the project: covers, audio tracks, research, marketing docs, distribution paperwork. Claude Code created a structured folder hierarchy on Google Drive, uploaded everything in two passes, and built a manifest-based sync script that detects new or modified files and pushes them. The manifest tracks each file’s Drive ID and last sync timestamp. I run one command and everything is current.

12. Finnish Translation

I’m publishing a Finnish edition in June. Claude Code built the translation pipeline. Each chapter goes through DeepL first (which is a text language translation tool), then Claude reviews the output against a locked glossary of Finnish terms that I’ve created. It highlights every change with == markers and writes a separate review file with a markdown table explaining each edit. I open both files side by side in Obsidian and go through them. Twenty-two chapters are done. The quality has been improving steadily. Early chapters had over ten flags each. Recent ones only average two or three, so it’s getting better over time.

13. ARC Campaign and Email Automation

The advance reader campaign ran entirely through Kit’s API. Claude Code tagged subscribers, scheduled broadcast emails, and sent reminders at each milestone. I drafted the email: the initial ask, the weekly check-ins, and the final “post your review on April 9” reminder. Then Claude scheduled them. When I needed to update a broadcast that was already scheduled, it hit the API to swap the content. No logging into dashboards, no clicking through forms. Just Claude talking with Kit’s API.

Final words

This book wouldn’t exist without the people around me. My family, for their patience through the early mornings and late nights. The founders who shared their sleep stories for the book: Julia Palatovska, Jenny Xu, Vlad Rannik, Emily Yim, Anders Leicht Rohde, Kristian Metzger, Martial Valery, and Robin Squire. Their honesty made it better.

The biggest thing I learned from this process is that the hard part of publishing a book isn’t any single task. It’s the sheer number of them. AI didn’t replace my judgment, but it let me act on it at a pace I couldn’t have matched alone.

Go get it: https://sleepagain.co

We still talk about tools like Claude Code as if they’re code generators. The name says it: code. Generative AI for producing code. But after 6 months of working with these tools daily, I’ve shifted how I think about them entirely. They’re not coding tools. They’re a new kind of computer operating system.

Think about what you do on a computer today. You open files. You edit files. You move things around. You search for things. With AI, you don’t do any of that anymore. Your interface is speech. The output is visual: text, interactive interfaces, whatever the task requires. The whole paradigm of pressing buttons and modifying files is going away.

One of the most interesting unlocks I’ve seen in the past 6 months is how AI operates your device on its own. It can install software. It knows how to use virtually every tool out there. You don’t need to learn any of it. This was a big deal for me personally. I was fluent at coding, but I’d taken close to a 20 year pause from active software development. Coming back, I realized I didn’t need to go study every new technology or read documentation on how to use specific tools. I just needed to know what a tool could achieve. Then I’d tell Claude Code: I want this, use this tool. And it would do it.

I’ve been editing video with Claude Code since November. I discovered FFMPEG, which is a command line video editing tool that can do basically anything. I’ve done green screen work with it. I take a video clip of myself on my phone, bring it to the computer, and ask Claude Code to put a black hole behind my head so I’m hovering in space. It does that.

This is what I mean by operating system. The change isn’t about producing code. It’s about how you interact with the computer.

The second thing I want to focus on here is QA. Quality assurance for games and mobile apps. This has been a really big unlock, and I cover it in my AI game dev course.

Here’s the typical workflow when you’re building a game with Claude Code. The code is ready. The default next step is: install it on your phone and try it out. Or have someone on your team, a QA person, test it. They file bug reports. There’s a whole rigmarole of humans interacting with the build. But now, the first steps of automating this process are at our fingertips.

I created a chess game. Then I asked Claude Code to play chess against the computer on the Android simulator. It can simulate taps on the screen at specific coordinates. It takes screenshots and determines where to tap next. It was fluent at playing chess. More importantly, it was testing that things work.

This is a simple game, sure. But the principle scales. How we produce bugs, how we reproduce bugs, how we navigate different screens and gameplay elements: all of that can be textually recorded, the same way we’ve done QA for years. The difference is that AI is now doing the work.

The challenges are still there. Think about a game like Hexa Sort, where hex tiles rapidly move around the screen from pile to pile. AI takes a screenshot every 5 seconds or so. That means it’s still limited to turn-based interactions. Fast animations are hard to evaluate from static screenshots.

But developments are happening. Android simulators now support video recording. So the AI plays the chess game, which has animations that need testing. It records a video, breaks it down frame by frame, maybe 5 frames per second, and analyzes each frame to see what the animation is doing.

We’re getting to a point where you don’t need to assume this is still a humans-only area. I think the only place humans will be needed soon is the creative work: having the taste to understand what works well in a game. All the non-creative work has now, or will in the next 12 months, be handled by AI.

Two quick mentions before this week’s piece:

• My book, Sleep Again, is coming out on April 9th. It has all the lessons I learned from struggling with chronic insomnia as a founder, and more. Join the waitlist to be the first to know when it’s live.

• I just pushed out a new lesson to my free AI Game Development course, check it out here. It’s meant for non-technical and technical people alike.

Now to this weeks piece.

Something strange is happening in the startup world right now.

AI is changing the economics of building companies faster than almost any technological shift we’ve seen before, yet if you look at how venture capital operates, you would barely notice anything has changed.

Round sizes are the same. Valuations are the same. The expectation that a startup needs to raise capital, hire a team, and scale headcount as quickly as possible is still the default script.

It’s as if the startup ecosystem is collectively pretending we’re still living in the pre-AI world. The more I think about it, the more I wonder why founders are still following that script.

Because if AI is doing what it appears to be doing, the entire logic of how companies are built might be changing underneath our feet.

The Venture Capital Gravity Field

There has always been a certain type of founder that venture capital loves.

A strong team with an incredible background is basically like the most attractive person at the party. Everyone wants to date them. If you’re that founder, you will always have the option to raise money.

And here’s the uncomfortable truth: many founders raise simply because they can.

Not because they actually need the capital, but because the system rewards it. The fundraising announcement creates credibility, momentum, and social proof. It’s the default badge of legitimacy in the startup world. I know because I’ve also felt proud in the past about raising for Silicon Valley top tier funds for my own startup.

But raising venture capital is not just a financing decision. It’s a strategic decision about what kind of company you’re building. Once you take venture money, you are no longer just building a company. You are now participating in a financial instrument designed to return venture-scale outcomes on a specific timeline. That changes everything.

I say this as someone who has been on both sides of the table. I’ve been a angel investor turned VC and I’ve been a venture backed founder.

There are companies that require funding to exist. Infrastructure-heavy businesses, breakthrough tech and certain kinds of services genuinely need capital to get off the ground.

But AI is quietly moving the boundary of what actually requires funding.

Startups Needed Money Because They Needed People

Historically, startups needed capital because execution required headcount.

You needed engineers to build the product. You needed designers to design it. You needed marketers to promote it. You needed operations people to keep the system running. Money was essentially a way to buy execution capacity.

AI is starting to dismantle that equation.

A growing portion of the work that previously required teams is now becoming software-driven. Code generation, design generation, marketing assets, content, research, analytics, customer support. These things are increasingly handled by machines.

The result is that the cost of execution is collapsing. And it’s not collapsing linearly. It feels exponential.

My Own Productivity Has Gone Vertical

The easiest way I can describe this shift is through my own workflow.

If I compare what I could build in the summer of 2025 with what I can build today at the end of Q1 2026, the difference is honestly absurd. The amount of output I can produce in the same amount of time has increased dramatically.

I would roughly characterize it as my efficiency doubling every few months.

And the truly interesting part is that we’re still very early in this transition.

Right now AI mostly behaves like a tool. You ask it to do things and it helps you execute faster.

But it is clearly evolving toward something else: autonomous collaborators.

Agents that can write code, run experiments, generate assets, analyze data, and deploy new features. As these systems continue to develop at breakneck pace, the marginal cost of building new products drops even further.

At that point, execution stops being the main constraint.

The One-Person $1BN Company Is No Longer Science Fiction

If you extrapolate this trend far enough, you arrive at a conclusion that still sounds radical but increasingly feels plausible: the one-person company operating at massive scale.

A founder orchestrating systems rather than managing employees. AI agents acting as the execution layer while the founder focuses on strategy, product direction, and distribution.

Ten years ago the idea of a billion-dollar company with one employee would have sounded ridiculous. Today it sounds improbable. In another ten years it may sound obvious.

We have already seen very small teams build extremely large companies. AI simply pushes that trajectory further.

The Real Problem Is Lack of Curiosity

What worries me is that many founders don’t seem curious enough about what AI is actually solving. A lot of startup thinking is still based on the old playbook. Raise capital, hire people, scale headcount, grow revenue.

But AI is attacking one of the biggest constraints startups have always faced, which is execution bandwidth. If the amount of things a single founder can build in a week doubles every few months, the entire structure of startup economics changes.

You might not need funding to scale people. Because you might not need people.

The Question Founders Should Be Asking

Instead of asking how much money they should raise, founders might want to ask a different question entirely: what can I build without raising anything?

AI is making experimentation radically cheap. Founders can build faster, test faster, and reach early revenue faster than ever before. And if you can reach meaningful revenue before raising capital, the power dynamics shift dramatically.

Funding becomes optional. Not a prerequisite.

The biggest thing AI unlocks is not automation. It’s leverage. A single motivated founder with the right tools can now operate at a scale that previously required entire teams.

That should fundamentally change how founders think about building companies. And yet much of the startup ecosystem still behaves as if none of this has happened.

Which raises an uncomfortable question: are founders raising venture capital because they actually need it? Or because they haven’t yet realized that the rules of the game have already changed?

So I’m experimenting with a new newsletter format. Occasionally, instead of focusing on a single theme, I’ll cover a few ideas that have been on my mind.

Today I’m writing about the AI Game Dev Course that I just launched, which is a free course for my subscribers. And the sleep book that I’ve been writing for the last two years, which is finally coming out in April.

Then on OpenClaw, which is this personal AI assistant system that has been in the news everywhere for the past month, especially after OpenAI bought it after three weeks of the project going live. I’ve been setting up a system and I want to share my learning so far from the first two weeks of building a cluster of OpenClaw bots.

AI Game Dev Course

Last year I started putting together a course on building mobile games with Claude Code. It didn’t come together, because I felt the tooling was changing so fast. Every week something changed in how Claude Code worked, and what felt like a solid approach on Monday looked incomplete by Friday. I decided to wait.

Now, in Feb 2026, things have stabilized. Using AI to build mobile games has gone from something experimental to something with a clear, learnable workflow. That’s the right moment to teach it.

I’ve spent 15 years making mobile games. I’ve watched the industry go through several major shifts: from premium to free-to-play in the early 2010s, from feature phones to smartphones before that, from small teams to large studios and back again. Each time, the shift looked uncertain from inside it and obvious in retrospect. Each time, the people who figured out the new way early had a real advantage. Claude Code feels like one of those shifts, if not the biggest, and 15 years of building games gives me a useful frame for evaluating what’s actually different now.

I’ve been deep in it. Learning what works, testing what breaks, figuring out where the real leverage is when you’re collaborating with an AI on a codebase. The workflow is different from anything I’ve used before.

That experience is what the course is built on. 12 lessons across 4 parts, starting from the first prototype and ending at a published App Store build. The case study is Read Clash, a speed reading game for iOS. Every lesson uses real working sessions, real decisions, and real code.

7 lessons are live now and the rest will follow over the coming weeks. What the course is actually about isn’t which buttons to press in Claude Code. It’s how to think about building with AI: how to break down a problem, how to communicate intent clearly to a model, how to stay in control of everything. Those skills hold regardless of what ships next month.

You don’t need to be a developer to follow along. You need to want to build something, even if it’s not the right thing, and be willing to work in a new way. The tools are capable. The workflow is learnable. The gap between an idea and a working prototype has never been smaller.

The course is free. Sign up here:

https://joakimachren.com/courses/ai-game-dev

More lessons are coming. I’ll keep you updated here as each one goes live.

Sleep Again book

My sleep book comes out April 9. I’ve been working on it for two years and I want to tell you what it is, because I haven’t said much about it here.

In 2019 I burned out.

My sleep had been fragile for years, but that was when I could no longer manage around it. I was waking at 4am, lying there running through decisions I could not act on until morning. I had been building companies for 15 years. Sleep was always the constraint I worked around instead of fixed. During burnout recovery, I slowed down and started sleeping again.

Then we started F4 Fund, and it deteriorated again. Everything that had worked stopped working. So I went deep into the research. Sleep science, chronobiology, and how founder stress disrupts sleep in ways standard advice does not address. I interviewed other founders. I ran experiments on myself. This book is the result.

It covers why entrepreneurs sleep worse than most people, and what to do about it given the life founders actually lead. There is a chapter on chronotypes and scheduling around your biology. One on the 3am wake up and what is happening physiologically. One on caffeine timing, because most founders shorten their sleep window without realizing it. One on overtraining, which took me three years to understand in my own case.

This is for founders whose brains do not switch off at night and who have already tried the usual tips. Most sleep advice assumes a predictable schedule and a stable cortisol curve. Founders do not have those.

I am looking for first readers. If you read it before launch and leave an honest review on Amazon or Goodreads by April 9, I will send you a free ebook copy now. A few sentences is enough. A new book lives or dies on its first reviews, and early ones from real readers matter more than anything else I could do.

I’m looking for 40 ARC readers and its first come, first served. Sign up at https://BookHip.com/SWHWWTL and I’ll send a copy in early March.

If you want to wait for the book to come out, sign up on the Sleep Again website to get notified. Hardcover, paperback, ebook and audio, all coming April 9th.

Your AI Co-Founder

It feels like the entire OpenClaw discussion is orbiting the wrong center.

People are excited about email triage, scheduled automations, timed events. But none of that is new. We could do most of it with Zapier and cron jobs years ago. Automation isn’t the breakthrough.

The real shift is this: we are getting close to fully autonomous AI founders.

Running agents 24/7 isn’t new either. I’ve had Claude Code grinding away on my machine for hours, executing long task chains without me touching the keyboard. That was already possible. OpenClaw doesn’t invent autonomy. It lowers the friction and expands the surface area.

The inflection point is when we stop thinking about AI as a productivity assistant and start treating it like a lean startup operator.

AI is natively built for experimentation. It can run A/B tests, collect data, optimize funnels, rewrite landing pages, iterate pricing, test positioning, all without ego and without fatigue. That’s where the leverage is. A founder that never sleeps and optimizes purely on signal.

My focus now isn’t on more clever workflows. It’s on turning these systems into AI founders. Entities that launch, test, measure, and adapt continuously.

There are two bottlenecks:

1. I’ve experimented with OpenClaw as an AI founder already. Here’s one of the many systems I’ve been experimenting with for the AI founder: ideate mobile games, many of them. In each stage, the AI founder does the work with several sub-agents.

   

   But the struggle is real. It can come up with ingenious ideas worth prototyping, but it doesn’t feel like it is inventing anything new. It’s very much like throwing stuff at the wall and seeing what sticks. This is where the bottleneck is: you need to human in the loop to point the direction.
   
   In the latest Naval podcast episode, “Motorcycle for the Mind,” Naval Ravikant talks about AI’s creative abilities. AI is good at connecting existing things, but Naval says that real creativity is something else. It is arriving at an answer that is not predictable or foreseeable from the known elements. Something so far out of bounds that you could search forever and not find it by accident. Naval points out that this kind of creativity is rare even among humans, which is why the human in the loop becomes more important. I can definitely feel that with game ideation.

2. The other bottleneck for this early wave of “AI founders” will definitely be the token cost. A 24/7 autonomous founder cluster running experiments nonstop won’t be cheap. But that feels like a temporary constraint, not a structural one, since the race for AI dominance isn’t only giving us better LLMs but the affordability has been coming in “all-you-can-eat” packages like Claude Code for $180 a month, which I use and have not ran into the usage limits for months now, even though I constantly have 3-4 Claude Code sessions running simultaneously for 8-10 hours a day.
   
   Memory is another similar complaint, the model forgets what it was supposed to do. This can be a bottleneck like price, but at the pace we are moving, all the memory issues will be solved soon.

What I am focused on now is not the OpenClaw cluster itself or getting the infrastructure to work. That part is increasingly solved. What I am actually building is the system, and that’s not something that needs OpenClaw. I can prototype and experiment on the system inside Claude Code, locally on my machine. The goal is an AI founder that can execute large tasks end to end. Creating documents, writing, coding, shipping, doing stuff online like filling forms.

The execution layer is no longer the bottleneck. The real human edge, for a long time still, will be in creative direction. But the leverage to build AI co-founders is already here.

News flash! My new book is coming out!

It’s called Sleep Again, and it launches on April 9th. I wrote it for busy people who know what it’s like to build something amazing while your sleep falls apart. If that sounds familiar, head over to the website to sign up for updates and get notified at launch. More on that soon!

Now, to this week’s piece!

I recently spent time with AppMagic’s Mobile Market Landscape 2026 report. It’s one of those documents that rewards slow reading. Not because every chart is surprising, but because when you step back, the overall shape of the market becomes much more clearer.

What stood out to me wasn’t a single “hit-category-of-the-year” or a random growth spike. It was more the sense that the mobile market has crossed an invisible line. It hasn’t slowed down. It has matured more than I expected. And maturity changes what matters.

The invisible majority of products

In 2025, more than 1.4 million apps and games were released, up roughly 25% year over year. At the same time, only around 10% or less ever attract meaningful user attention. This data point explains a large share of what many of us are experiencing.

Most products do not fail because they are bad. They fail because they lack the critical mass of users to take them off the ground. In a market where release volume keeps accelerating, quality alone has stopped being a differentiator. It is necessary, but no longer sufficient. AI slop or not, the quality is not a defining factor.

What this forces is a reordering of priorities. Generating attention is the real bottleneck, and it shows up far earlier in the lifecycle than many teams expect. • ⁃ Because of AI capabilities, agents talking with advertising platform APIs, analyzing data, running ad variants, 24/7 with no humans in the loop, the data-driven approach of game development will win even bigger than before.

As an investor, this has changed how I think about risk. I am less interested in what an app does in isolation and far more interested in why it will be discovered at all. And what the founder sees as the mechanisms for discovery.

Games did not disappear, but apps quietly took the lead

One of the most important moments in the report is easy to miss. In September 2025, non-gaming apps overtook games in monthly revenue for the first time, generating $4.8B compared to $4.5B for games. Since then, the gap has continued to widen.

Games are still enormous, and they are definitely not going away. But economically, something meaningful has shifted. Apps have become the more predictable, compounding revenue engine. Growth is spread across many categories rather than concentrated in a small number of hits. Monetization is still subscription-led. Teams can be smaller, more focused, and still build very real businesses.

From an investment perspective, this matters a lot. To have venture outcomes, games are often capital-intensive, hit-driven, and operationally complex beasts. Apps, by contrast, allow for a calmer kind of ambition. You can build something durable without needing a single explosive moment of success. You are building utility, versus entertainment, with fierce competition for attention.

This does not make apps easier. It makes them different. The upside now comes from execution quality, monetization innovation, and retention mechanics which many B2C app developers have previously neglected, causing lower LTV.

Generative AI has moved past curiosity

Generative AI apps is the fastest-growing segment in the entire report. Downloads grew by 178% year over year. Revenue grew by 273%, reaching roughly $3B in 2025. On the surface, this looks like pure acceleration.

What makes it more interesting is what happens underneath. Retention is declining across the board. Revenue is highly concentrated, with ChatGPT alone accounting for the majority of chatbot revenue.

But there is a clear, yet more nuanced story here.

We are past the novelty phase. Users are no longer impressed just because something is powered by AI. They try it, and if it does not provide value in a repeatable way, they leave. That is exactly what you would expect as a category moves from experimentation into utility.

As an investor, this shifts the question entirely. I am no longer asking whether a product uses AI. I am asking: what habit does it cater to? What repeated problem does it solve? Why does it deserve a monthly subscription rather than a one-time experiment?

The report shows that most AI revenue concentrates at around $6 to $40 per month. That is not impulse pricing. That is trust pricing. You only get there if the user come back often enough to justify it. The next wave of winners will not be horizontal chatbots or ones who justify high CAC on customers “forgetting” to unsubscribe from a $29.99 monthly subscription. The winners will be vertical, opinionated tools that users depend on rather than play with.

What I now expect from developers

I do not expect founders to predict the future. But I do expect them to understand the market they are entering.

After reading Appmagic’s report, my expectations have become simpler, but also stricter.

First, I expect a real distribution hypothesis. Not a vague plan to test ads, but a clear explanation for why this product earns attention at all. Whether it is content, creators, SEO, platform leverage, or community, there needs to be a reason discovery will happen.

Second, I expect monetization literacy. Founders should know their likely price band, their primary revenue lever, and what success looks like in their specific category. The venture-scale outcome requires that the incentive for repeat spending is there. Treating monetization as something to figure out later is increasingly a liability, more than ever.

Third, I expect category awareness. Retention, growth, and revenue behave very differently across segments. This is especially true for founders, moving from gaming to make apps. A good retention curve in one category would be mediocre in another. Founders who do not know their benchmarks are not being bold. They are being blind.

Closing thought

The biggest takeaway for me is not optimism or pessimism. It is clarity.

The mobile market in 2026 rewards teams that are intentional about distribution, engagement and monetization, and category dynamics. The era of accidental success is shrinking. The upside now belongs to founders who understand the constraints they are operating within and design accordingly.

That is not a bad thing. It simply means the market is asking for more deliberate products, built with eyes open.

We all recall, back in the 2010s, especially in mobile games, how popular the concept of “independent teams make all the decisions” was. Inside larger studios, the prevailing belief was that if you just gave teams enough autonomy, enough ownership, enough trust, the results would take care of themselves. And to be fair, in many cases that model worked quite well.

Lately I’ve been thinking that we are again at the cusp of a new era of team-led game development inside larger studios, but with a very different definition of what a “team” actually is. The idea I keep coming back to is an AI-native games company organized around product managers running what are effectively solo teams, supported by AI agents. Each PM would own a game, end to end, and work directly with the agents to design, build, iterate, and operate the game. True independence by default, not just in theory but in practice.

We are probably still early for this to work perfectly, but we are much closer than people seem to realize.

• On the coding side, I honestly think we are already there. Over the last six months of using Claude Code to build mobile apps, the progress has been staggering. Every month the capabilities take a meaningful step forward. Things that used to require a full team and a long timeline are now realistic weekend projects, projects I could never have imagined even attempting before. Having coders at your fingertips, constantly building, refactoring, and reasoning through problems while you watch, fundamentally changes your sense of what is possible.

• On the graphics side, we are getting there as well. Recently I’ve been creating game assets with ChatGPT by simply asking for them and explicitly asking for transparent PNG backgrounds so that the assets are immediately usable inside a game. That alone removes a surprising amount of friction. Sprite sheets are almost there too. There are still clear challenges, especially in getting each frame to contain exactly what I want it to contain, but the trajectory is obvious. And this is not limited to one model either. Grok and Gemini are very close as well. Scenario and Layer have been at it for years already.

So the question becomes: what is actually hard now?

The challenge is not execution anymore. The real challenge is ideas, specifically great game concepts that are grounded in a kind of 80/20 mindset where proven patterns are combined with just enough innovation. There is a lot of talk right now about this being the age of the idea person who can quickly build lots of cool things, and that is true to an extent. As I wrote in my previous piece, it has never been as cheap or as fast to generate software and content as it is right now. You can build multiple games at once, spin up prototypes effortlessly, and move lightning fast.

But that abundance shifts the bottleneck rather than removing it. The hard part becomes validation. Knowing which ideas deserve a place in the pipeline. Developing a gut sense for what makes sense to pursue and what should be killed early. That kind of judgment has competitive value, and it has not been automated away. In many cases, you know nothing until you have ample cohorts of users, repeat spenders, etc. and AI can’t make that part faster.

Then there is marketing, which is an entirely different beast. I can imagine a future where marketing AI agents exist that can take over most of the process of launching and scaling a mobile game, but we are not there yet. The difference compared to coding is striking. Claude Code works so well because it is essentially everything in one place. A year ago, people were wiring together MCP tools and open source projects just to make these systems usable. Now most of that is becoming redundant, and the functionality is converging inside a single tool.

It naturally makes me wonder when we get the equivalent of Claude for marketing, because that is something I am very much looking forward to.

My route

This brings me to the second part of the article, and the more personal question of why I am not starting an AI-native games company organized this way.

The honest answer is that it would still require a lot of humans, and I do not see a strong enough competitive advantage for myself to get into that kind of a business. In games, I would still be competing with extremely sharp, fast-moving operators, including the kind of genius hustlers coming out of places like Turkey. They can adopt the same AI-native approaches just as quickly as I can, if not faster. In that environment, simply having access to the same tools does not create a meaningful edge.

This is why I keep coming back to the one-person app studio model instead. On the app side, the dynamics are different. I can make myself the user. I can take something that already exists and that I already use, and make it significantly better for a very specific niche. I can make it more AI-native, more optimized and more closely aligned with my own needs. From there, it is possible to scale outward. In games, that personal wedge is much harder to establish.

Final words

The AI-native games company will exist. I am convinced of that. But for now, it is not the thing I feel compelled to build myself, and being honest about that distinction feels important.

I’ve written before about Duolingo, mostly not as a language app but as a game. Probably the most successful game-like utility app in the world.

And to be clear, it’s very good at certain things. It gives you goals. Streaks. Things to chase both short term and long term. You open the app and you know exactly what you’re supposed to do. You see how other people are doing, what league they’re in, whether you’re moving up or down. From a game design point of view, a lot of this is really solid.

But I keep coming back to the same question, which is what value it actually creates.

I honestly don’t think I’ve met anyone who spent a massive amount of time in Duolingo and then could actually speak the language. I had a friend in the VC world who had a daily streak for over a year, learning Finnish. When I tried to talk to him, he basically knew individual words. That was about it. No conversation, no flow.

That feels telling.

You’re probably learning something. You’re likely picking up vocabulary. But you’re not really learning the thing you think you’re learning, and probably not the thing the app implicitly promises. To get conversational, you have to leave the app and do other things.

Which is fine, but then the product is selling progress without outcomes, or at least progress that only really exists inside its own system. An illusion of utility more than the utility itself.

And I think that illusion matters more now than it used to.

Why this suddenly feels relevant

The reason it feels more important now is that the way apps are built has completely changed. With vibe coding and AI, building something usable in a weekend is normal. Shipping it is easy. Generating content is easy.

What’s interesting is that a lot of people building these consumer apps don’t come from game development at all. They don’t have that background or intuition.

Meanwhile, people in mobile games have spent years obsessing over engagement, retention, motivation, progression, and what keeps something interesting after day thirty instead of just day one. They’ve internalized this stuff.

You can feel that difference when you look at consumer apps.

Everything optimizes for the paywall

Most consumer apps still optimize for one thing above everything else, which is converting you to a subscriber as fast as possible. Short onboarding, a bit of novelty, then a paywall.

That’s the loop.

The whole idea of building habits over time, or turning something into a hobby, or giving people a reason to come back beyond obligation just isn’t the priority. Games think about that constantly. Consumer apps mostly don’t.

And that’s why so many of them feel shallow once the initial novelty wears off.

Content is basically free now

This part still surprises me a bit.

In the last six months of building vibe-coded apps myself, I realized just how cheap content has become. You can generate enormous amounts of it. You can create agents that behave like level designers. Things companies used to raise funding for are now just normal parts of the toolbox.

Which makes the shallowness of a lot of apps feel even stranger.

If content is cheap, there’s no real excuse not to build depth.

Some categories already hint at this

There are a few categories where you can already see hints of what’s coming.

Couples apps are interesting here, partly because I’m building one myself. “The Truth or Dare” category in particular has exploded. A lot of these apps are making real money.

If you look at them closely, they’re basically hypercasual games. The experience is shallow, but the promise is very clear. You’re on a date or out with friends, you open the app, and it delivers immediately.

It’s essentially a board game compressed into a mobile app. Easy to build, easy to understand, easy to monetize.

Music learning apps are another category I keep thinking about. They’ve been around for a long time. Some are profitable and well designed. They already use gamification. But none of them have really produced a breakout on the scale of Duolingo.

Maybe music learning is just a smaller market. Or maybe the engagement never quite crossed some threshold. I’m not sure.

Things from games that feel inevitable

When I step back, it feels like a few things from games are going to move into consumer apps whether people want them or not, not because they’re trendy but because they solve problems that consumer apps keep running into.

The first one is real PVP. Not leaderboards or streak comparisons or being ranked somewhere on a list, but actual head to head interaction inside the content. The moment you put two people into the same piece of content, engagement changes. It stops being analytical and starts becoming emotional. You’re no longer optimizing a number or trying to keep something green. You’re reacting to another person. Duolingo’s streaks already hint at this. Trying to keep a longer streak than someone else is a soft form of PVP, but it never really crosses the line into direct interaction. Once it does, the app stops feeling like a tracker and starts feeling like a game, and that shift is bigger than it looks.

The second thing is live ops. Daily things, weekly things, monthly arcs, something that is happening right now rather than something you can grind whenever you feel like it. Games live on this. It’s how they avoid feeling finished. There’s always a reason to show up today instead of tomorrow. Consumer apps mostly don’t do this, and I think that’s because they’ve been optimized around early subscription conversion rather than long term presence. Live ops only really make sense if you care about people showing up again and again. But when content is cheap, live ops stop being scary. You don’t need a big team or handcrafted everything. You can generate events, challenges, and variations continuously, and once you do that the app stops feeling static.

The third thing is collaborative gameplay, and this one feels especially underexplored. Games figured out a long time ago that doing something together is fundamentally different from doing it alone. Guilds, alliances, teams, shared goals. Most consumer apps are still single player experiences, and even the social ones are usually about comparison rather than cooperation. I’m building around this idea right now, two people progressing together, sharing goals, unlocking things together, and once you start thinking this way it feels obvious that this shouldn’t be limited to couples. Families, households, groups of friends, small communities, even strangers who opt into doing something together. Collaboration changes motivation. You’re no longer only accountable to yourself. You’re accountable to someone else, and that turns out to be a very powerful shift.

Ending thought, not a conclusion

I don’t think the next wave of consumer apps is about more engagement tricks.

I think it’s about generating actual value. Measurable value. Things that matter outside the app, not just inside the metrics dashboard.

If you generate real value, people stay. They talk about it. They pay.

And now that building depth is cheap, it feels like we’re still barely scratching the surface. That’s the part that keeps pulling me back to this.

Here are my predictions for the games industry in 2026.

1. Breakout gaming winners will largely come from Turkey and China.

Hard times hit the softest players first. In Europe, there hasn’t been a single breakout games company founded in the 2020s. Compare that to the 2010s and the contrast is brutal. This is no longer a game of brute force production. The winners will understand that making games is not about making games. It’s about building a product and distribution engine capable of challenging incumbents in a believable, smart, and capital-efficient way. Creativity still matters, but only when it aligns with the structural shift in how games are discovered, distributed, and consumed.

2. Venture capital will continue to retreat from early-stage gaming.

The reasons are familiar: studios struggle to raise follow-on rounds, something I’ve covered extensively on my Substack. But there’s another dynamic quietly reshaping the market. Asian strategic investors, long active as LPs in gaming funds, have built internal teams to invest directly into studios. They’ll still value exposure to Western markets, but not through a dozen funds, where they’ve in the past been limited partners in, to gain access. They’ll concentrate capital into a small number of vehicles that deliver real intelligence and results. The rest will be left behind, meaning less bets in new games companies.

3. The EU will make moves against Roblox and other platforms targeting children.

After my November piece on Gen Alpha and their gaming habits, I received several DMs pointing toward an inevitable regulatory push in Europe. I predict that dree-to-play platforms targeting adolescents will face stricter age-verification requirements. It’s uncertain how strongly the regulatory push will be in 2026, but directionally, this feels like an unavoidable move. Look at Norway, who is not even an EU country, has just created the most strict social media age restrictions in Western countries.

4. AI-driven efficiency will unlock an explosion of bootstrapped startups.

Generalist, jack-of-all-trades founders will increasingly choose to bootstrap. Why raise a pre-seed when $200 a month buys you an AI coding partner that outperforms what used to be a six-figure hire? The escalating competition between major LLM companies is making these tools smarter and more creative at a staggering pace. Updates arrive weekly. Bootstrapping for a year, building real traction, and skipping pre-seed entirely will become a rational default. And once you do that, the question becomes uncomfortable but obvious: why raise dilutive capital at all? Why invite VC overlords into the cap table? 😉

5. Mid-sized studios will grow as leadership moves closer to product and marketing.

The era of distant management is over. In order for gaming companies to grow, leadership must engage directly with product decisions and distribution strategy. That means removing blockers, systems, and sometimes even people they once trusted but who can’t operate at the required level anymore. Growth won’t come from more process. It will come from sharper judgment, faster decisions, and brutal prioritization.

That’s a wrap. Happy new year 2026!

It’s the time of the year to reflect and review, and to take a look into what next year will hopefully bring to me. For the past five years, I’ve been in the habit of reflecting on the past 12 months, what have gone well, what hasn’t gone well and what things I look to improve. Previous reviews from 2020, 2021, 2022, 2023 and 2024.

Some handy templates for annual reviews that I’ve utilized over the years:

• Steve Schlafman’s Annual Reflection Guide

• James Clear’s annual reviews

• Chris Guillebeau’s annual reviews

So let’s get started.

What went well in 2025?

A lot of things went well this year. First, we completed several investments with F4 Fund, and I’m genuinely happy about how that chapter progressed.

But more importantly, during the summer I was fully bitten by the “AI bug.” That moment helped me see something very clearly: I am an entrepreneur first. I love building things, going deep into products, and watching ideas turn into something real. AI pulled me back into that mindset in a way I hadn’t fully felt in years.

I’ve experienced this kind of timing advantage before. When we started Next Games, mobile free-to-play gaming on smartphones was just taking off. The App Store was still early, distribution was changing fast, and new behaviors were forming. That timing mattered. We ended up building a company that eventually exited to Netflix, and in hindsight, the biggest leverage wasn’t just execution, but being early in a real platform shift.

I increasingly believe that the highest leverage an entrepreneur can have is to ride a genuine new wave, not a trend, but a fundamental change in how people behave and how products are experienced. For me in 2013, that wave was mobile free-to-play gaming. In 2025, it is clearly AI.

You could argue that moments like this resemble gold rushes, and in some sense they do. But AI feels different. This isn’t just another speculative cycle. It is a structural shift in how work gets done, how products are built, and how people interact with technology. In many ways, it feels comparable to the smartphone revolution, and arguably even larger. Smartphones changed what we carry in our pockets and how often we engage with a screen. AI changes how we think, create, and solve problems on a daily basis.

When I compare this to other “timing moments,” like the crypto and Web3 wave a few years ago, the difference is stark. That movement never felt like a true platform shift to me. It didn’t fundamentally reinvent everyday behavior, beyond a narrow set of financial use cases, and the incentives for mass adoption were never strong enough. With AI, the incentives are immediate and obvious. Anyone using generative AI, whether for writing code, creating art, answering complex questions, or building presentations, feels the value instantly. There is no friction in understanding why this matters.

What did I change my mind about in 2025?

That I could actually be a founder again. I had thought that my life of building and launching product was behind me, but it’s been the best mind change I’ve done in years. And that decision has felt better and better as more time goes on. I’m learning so much about myself in the midst, which is probably the most gratifying thing about being an entrepreneur again.

What didn’t go well in 2025?

There are quite a few things I could list, but one of the biggest was my expectation around AI and how much heavy lifting it would do for me when building products. When I started producing apps and experimenting seriously with AI-assisted coding, I genuinely believed the tools would take care of most of the work. When I picked up Claude Code during the summer, it felt like something entirely new. For a moment, it even felt like I didn’t need to do everything myself anymore.

At the same time, that illusion quickly broke. It felt like the tools could do everything, but in reality, they couldn’t. Especially when working on larger applications with multiple interconnected systems, client-server architectures, and real-time user interactions, the limitations became obvious. AI coding agents struggle to truly understand complex products as a whole. They have limited context windows, and unless you explicitly design your workflows to compensate for that, they simply cannot keep all the critical pieces of a project in mind at once.

I learned that AI does not anticipate product needs in the way a human architect does. To make it effective, you need very deliberate structures, rules, and methods that force the agent to repeatedly re-anchor itself to the core principles of the project. Without that, things slowly drift, and complexity becomes a liability rather than a strength.

Another thing that didn’t work out was my plan to create a coding course around AI-powered game development. I had hundreds of people on a waiting list, but the course never materialized. The core reason was simple: everything was changing too fast, including my own understanding. I spent most of the summer actively coding and learning, and by late August it felt like I should finally share what I knew.

But the problem was that I was still running into new issues every single week. AI coding constantly introduced new frustrations and unexpected limitations. I didn’t want to teach something that would immediately become outdated, nor did I want my students to hit the same walls I was hitting without having better tooling and solutions already figured out for them. Teaching felt premature when I hadn’t yet stabilized my own workflows.

Finally, after deciding to leave F4 Fund, I completely underestimated how consuming starting a new company would be. I didn’t expect it to affect my ability to write and reflect as much as it did. Building a startup is an all-consuming effort, and I felt that deeply this fall. After kicking off my new company project in September, my focus narrowed almost entirely to execution.

Right now, that company is centered on building mobile consumer apps with AI, drawing heavily on everything I’ve learned from two decades in the games industry and from being both an entrepreneur and a VC. But the trade-off was real. Writing, teaching, and stepping back to reflect became much harder than I expected.

What habits or interests held me back?

I went through my notes and reflections from the year and here are the things that stood out and things that could have held me back.

1. Too many parallel tracks. I consistently carried more projects, ideas, and experiments than my attention could truly support. Even good initiatives became diluted when run side by side.

2. Perfectionism disguised as preparation. I often stayed in optimization mode, refining systems, workflows, and setups instead of committing and compounding with something good enough.

3. Tool curiosity replacing execution. Exploring AI tools, models, and workflows sometimes felt like progress but delayed shipping and real-world validation.

4. Underinvesting in recovery. Stress, sleep issues, and insufficient recovery quietly reduced my cognitive capacity, creativity, and patience over long stretches.

5. Turning learning into pressure. I treated learning as a race, creating unnecessary urgency and self-imposed expectations that reduced enjoyment and clarity.

6. Execution crowding out reflection. Once I entered full building mode, I failed to protect space for writing, thinking, and synthesis, assuming it would happen automatically.

7. Opening loops faster than closing them. I started more initiatives than I finished, slowing momentum and limiting the compounding effect of completed work.

Top lessons learned in 2025?

Besides the fact that entrepreneurship is my way and not investing? Let me think.

No, actually this year delivered more learning than I expected, mostly through friction rather than smooth execution. These aren’t theoretical insights. They’re lessons earned by building, hitting limits, and adjusting in real time.

The biggest realization was that AI is leverage, not autonomy. It doesn’t replace thinking, judgment, or architectural responsibility. When I assumed it could “just handle things,” especially in complex products, it failed. The real power came from pairing AI with clear constraints, explicit rules, and strong product intent. Experience didn’t matter less this year. It mattered more.

I also learned that focus compounds more than intensity. Pushing hard across too many parallel efforts created motion without depth. Progress only accelerated when I narrowed the scope and stayed with a problem long enough for effort to compound. Intensity without focus produced noise. Focus produced momentum, and I’m still trying to learn it intensely.

What is my purpose for 2026?

I really want to ship a consumer app and run marketing for it, all by myself. I’m now in a spot, with all the AI tooling available, to make this a reality and to scale a business like that. The goal isn’t to ship and grow; rather it’s to actually show that you don’t need to raise funding to make it. Too many founders obsess over raising VC before they get going, and I want to prove that this is not necessary.

In 2026, my purpose is simple and very concrete: I want to ship a consumer app and take it all the way to meaningful traction entirely on my own. I want to build it, ship it, and run the marketing myself, end to end.

The point isn’t just to launch something, or even to grow it aggressively. The deeper goal is to demonstrate that, with today’s AI tooling, it’s now genuinely possible to build and do some early scaling for a real consumer business without raising venture capital. I want to prove this by doing it, not by arguing about it.

Too many founders fixate on fundraising before they’ve built anything of substance. Capital becomes the plan instead of a tool. I want to show a different path: start with users, distribution, and revenue, and let everything else be optional. If funding ever becomes relevant later, it should be a choice, not a prerequisite.

Do not get me wrong, this isn’t an anti-VC statement. VC has its place in growing companies past a certain point, when you’ve shown that things work and that there is a massive outcome to be pursued.

This is an efficiency statement. AI has lowered the cost of experimentation, development, and iteration so dramatically that individual founders can now operate at a scale that previously required teams and capital. My purpose for 2026 is to fully lean into that reality and make it visible through a real product in the world.

What 2-3 goals do I want to accomplish in 2026?

1. Ship and grow one consumer app to real usage and revenue, solo

I want to take a single consumer product from idea to real users and recurring revenue without raising funding or building a team. What matters isn’t just shipping, but owning the full loop: product, iteration, and distribution. This goal forces focus and removes excuses. If I can make this work alone with modern AI tooling, it proves that building a sustainable business no longer requires capital as a starting point.

2. Build repeatable solo-founder systems for product and marketing

Beyond one app, I want to develop reliable systems for how I build, ship, and market products as a solo founder. This includes AI-assisted development workflows, fast iteration loops, and distribution tactics that don’t depend on paid growth. What’s important here is compounding. If the systems work once, they should work again, and that turns one success into a durable operating model.

3. Make independence visible through execution, not commentary

I want to demonstrate, through a real product in the market, that founders don’t need permission or capital to get started. This isn’t about being anti-VC. It’s about shifting the default mindset from fundraising to building. What matters is credibility earned by execution. The product itself should be the argument.

Finally, what do I need in 2026?

More than anything, I need fewer moving parts and clearer constraints. I don’t need more ideas, tools, or opportunities. I need a tighter operating environment where focus is protected and decisions are simpler. One primary product, a small set of metrics that matter, and a clear definition of what “good progress” looks like.

I also need to deliberately protect energy. Sleep, recovery, and mental space can’t be treated as background variables anymore. They need to be designed into the year, not hoped for between sprints. High-quality output depends on sustainable input, and this year made that non-negotiable.

Happy new year everybody!

At the end of the year, I have a habit of covering the books I really liked. In 2021 and 2022, I managed to read over 60 books. In 2023, it was only 19. Last year, I got back on track with 31 books. This year, I have finished only 15 so far, with a few still in progress. I have clearly spent more of my time planning, building, and studying things for my new startup, and I had less time for reading. But I still got into some truly interesting books.

This time, I am taking a slightly different approach. I will cover four of the five books in a concise way, and then focus on my favorite of the year in more depth.

By the way, here are my favorite books from previous years.

• 5 Favorite Books of 2024

• 5 Favorite Books of 2023

• 5 Favorite Books of 2022 (easier to pick 5 😅)

• ​10 Favorite Books of 2021​

• ​10 Favorite Books of 2020​

• ​10 Favorite Books of 2019​

• ​10 Favorite Books of 2018

The Four Good Ones

Here goes.

The Wealth Ladder, Nick Maggiulli

“The Wealth Ladder” feels like a straightforward manual for building financial freedom through repeatable, practical steps instead of chasing a pipe dream or lottery wins. Maggiulli breaks wealth down into stages: earning, saving, investing, and scaling, and explains how each stage requires a different mindset and set of behaviors.

The best part is that the book does not try to impress with complex theories. It shows how ordinary people can move up the ladder by being intentional, building skills that raise earning power, and consistently putting money to work in simple, long term investments. It is basically a roadmap for turning disciplined habits into compounding advantages.

Why it deserves a top 5 spot:

The book addresses a topic that is close to any entrepreneur. Once you start making a living and paying your bills, the itch often appears to actually build real wealth. Not everyone feels it, but I have noticed that the wealth creation that comes from building your own company is a side product that rarely gets talked about. This book explores that better than any I have read.

Meditations for Mortals, Oliver Burkeman

“Meditations for Mortals” reads like a gentle but sharp reminder that life is short, and that accepting our limits is the fastest way to stop wasting time on the wrong things. Burkeman takes the big existential ideas from Four Thousand Weeks and turns them into short reflections that help you step out of anxiety, perfectionism, and overthinking.

Why it deserves a top 5 spot:

I rarely come across a book that makes me think this much while I am reading it. It kept pulling me back to the simple truth that we have a limited number of days, and clarity comes from choosing what truly matters rather than trying to do everything. It is both practical and philosophical, offering grounded prompts for focusing your attention.

No More Mr. Nice Guy, Robert Glover

“No More Mr. Nice Guy” is a wake up call for men who hide their needs, avoid conflict, and hope that being agreeable will solve everything. Glover explains how these habits begin as childhood coping strategies and later turn into resentment and weak boundaries. The book gives concrete steps for changing the pattern by being honest, setting limits, and acting from your own values instead of seeking approval.

Why it deserves a top 5 spot:

This book hit me because it put words to behaviors I had never really recognized in myself. It made me see how much I had tied my self worth to being agreeable and keeping the peace, even when it meant suppressing my own needs with my spouse and kids. It has given me tools to change that, from expressing what I want to setting limits without guilt. It has helped me feel more solid and authentic in my relationships.

Mark Twain, Ron Chernow

What I liked most in Chernow’s Mark Twain biography was how it revealed the messy side of a writer most of us imagine as effortless. Twain struggled with money, bad investments, perfectionism, burnout, and the pressure to keep producing. Seeing that side made me feel more connected to my own creative process. His accomplishments felt less mythical and more like the result of resilience and persistence.

Why it deserves a top 5 spot:

The book reads like the biography of a creative entrepreneur who kept reinventing himself. Twain took risks, made bad bets, rebuilt his finances, and continued innovating while managing his public persona. Seeing those cycles up close made me think about my own work in a grounded way. It reinforced how persistence and reinvention matter more than perfection.

My Favorite of the Year

The Art of Spending Money, Morgan Housel

“The Art of Spending Money” became my favorite book of the year because it helped me see money, both holding it and not spending it, from new angles. Housel shows how disposable income can quietly start shaping your identity, and how you may begin defining yourself by what sits in your accounts or your assets. The book helped me see more clearly how cash in the bank, and other assets, interacts with daily life without turning it into a performance, something I have definitely noticed in myself.

The biggest insight is that the book is less about spending and more about not spending. It is about preserving autonomy. Like in his previous books, Housel explains the freedom that comes from choosing not to buy, not to signal, and not to let money pull you into lifestyles you never consciously chose. That shift moves the focus away from optimizing purchases and toward protecting autonomy.

Get this book, and get his other books as well. He has a unique way of thinking about how we should approach money without letting it define us. This book helped me clarify what truly matters, cut the noise, and appreciate the power of restraint.

Out of all the books I read this year, this one gave me the cleanest and most useful shift in perspective. In my opinion, that is what matters most in books.

When I played games in the 1990s, things were simple. You bought a game, or more often, got a pirated copy from a friend, installed it, and played until you finished it, or you got a new one to play. Games were offline and finite. You could play to the end, see the credits roll, and that was it. There were no updates, no notifications, no constant live events.

Games were also premium products. You paid for them once, and the transaction ended there. The business model rewarded making great experiences that people wanted to buy. There were also more games; at least the perception for me was that you’d cycle through a lot more games in a year.

Today, everything is the opposite. Games are online, persistent, and designed to be played endlessly. They are updated weekly, layered with progression systems, and surrounded by content ecosystems that make stopping difficult. And the best ones keep you playing for years, restricting much of your attention to this one entertainment product.

In this piece, I want to talk about how this shift affects the latest gamer generation, Generation Alpha, and how modern game design is shaping the way they play, think, and spend their time.

Who Is Gen Alpha and How Much Do They Play

Generation Alpha includes children born roughly between 2010 and 2024, the first generation to grow up entirely in a digital world. They learned to swipe before they could read, and gaming has been part of their lives from the start.

Many studies and surveys show that around 90 percent of Gen Alpha play video games in some form. It is an extraordinary number, especially when compared with their parents’ generations. For Gen X, roughly 30 percent played video games when they were children. For Millennials, the figure was closer to 40 percent.

The leap is massive. Gaming has overtaken nearly every other form of entertainment for kids. It is more common than watching TV (dah!), more engaging than social media, and in many cases, more social than real-life interaction.

Developer Perspectives on What Has Been Happening

When my first gaming company, Ironstar, developed a game inspired by The Sims and Nintendogs, it was not meant for kids. The idea was to create a virtual character living in a small world, performing tasks while you were away.

To my surprise, the audience turned out to be much younger than expected. The game started attracting pre-teens who played constantly. This was in the mid 2000s and it was the first sign for me that something was changing in the gaming landscape.

Over the past decade, I have watched mobile games shift toward younger and younger audiences. Ten years ago, I knew plenty of adults who played mobile games daily, during commutes, breaks, or evenings at home. Now that adult audience has nearly vanished.

The new growth is clearly coming from children and from their parents’ wallets. Making games for kids used to be niche and sometimes frowned upon. Today, it is one of the most profitable corners of the industry. Much of the record-breaking revenue in gaming comes from products designed for Gen Alpha, whether intentionally or not.

Why Gen Alpha Gamers Are Addicted

1. The Influence of YouTubers and Streamers

There are now thousands of YouTubers and streamers who publish new gaming content daily. They have created a nonstop flow of excitement around games. Watching someone else play is now almost as engaging as playing yourself.

Platforms like Twitch and YouTube work in tandem with the games themselves. When a game becomes popular, creators rush to make videos about it. Those videos drive more players into the game, and that engagement, in turn, fuels more content. It is a powerful loop that keeps games constantly visible and exciting for kids.

2. Outdated Age Ratings

The current age rating systems were built for a different era. They focus on violence, blood, and language, but those are not the main drivers of today’s gaming addiction.

What really keeps kids hooked are free-to-play mechanics: rare item drops, endless trophy roads, “one more match” nature of the games, and seasonal reward systems. These features are designed both to keep players from closing the game and to pull them back in after they do. Gen Alpha is especially vulnerable to these mechanics because they have the most free time and are still adolescents learning how to manage it.

The industry and regulators have largely placed responsibility on parents, assuming they can monitor and limit gaming habits. In practice, that is not working. As a parent of two Gen Alpha kids, I have seen how easily children get drawn in, and how common it is for parents to allow unlimited play without realizing what is happening.

Addiction is not just about time spent; it is about how the games are built to make not playing feel like missing out.

3. Skill and the Loss of Real-World Time

One positive aspect of gaming is that it teaches mastery of things like dexterity. I.e. this “second-to-second” gameplay where you need to take action at exactly the right time. Many games reward this effort strongly, showing kids that practice leads to improvement. That lesson, that you can get better at anything with persistence, is valuable.

But it also comes at a cost. Time spent improving virtual skills often replaces time spent outdoors or with friends in the real world. This is one of the key themes Jonathan Haidt discusses in The Anxious Generation, that children are losing the unstructured play and face-to-face experiences that once defined childhood.

When the majority of free time is spent in digital worlds, real-world activities start to feel optional or even boring by comparison.

Conclusion

Gen Alpha is the first generation to grow up with games that never end. For them, gaming is not something you do and finish; it is a continuous experience that evolves, updates, and asks for more of their time.

The problem is not that kids love games. It is that modern games are built to love them back, with carefully engineered systems that reward constant engagement.

When I think back to the 1990s, I remember games that ended when you turned off the computer. For today’s kids, there is no such ending. And that might be the hardest level for them to beat.

In my previous article, Building for Everyone: The UX Challenge of Consumer AI, I argued that as AI tools move from niche prototypes to mass-market apps, the single biggest barrier is not model size or dataset quality, but interface quality.

If a system cannot deliver fast, fluid, intuitive interaction, it does not matter how smart the model is or how broad its access. It will fail to capture everyday users.

Now, we are entering a new frontier. With the launch of ChatGPT Atlas, we are beginning to see how large language models (LLMs) are not just helping us search or summarize information. They are becoming the interface to the web itself.

In this essay, I want to explore what that means: how the interface of computing is being rewritten around language, and why, despite this technological leap, one truth remains constant: great UX still wins.

From Mobile Apps to Model Interfaces

ChatGPT Atlas just launched, and it feels like the beginning of something big. It is a browser where you don’t really browse. You type what you want, and the large language model (LLM) goes out into the web to find, summarize, and structure it for you. You can still open individual websites, but the LLM has quietly become the interface to the World Wide Web.

It is an early glimpse of a new paradigm. For decades, the browser was our window to the internet. Now, the window itself is learning to think.

From Mobile Apps to Model Interfaces

Today, the smartphone is our dominant computing device. And if the web is being reshaped by LLMs, mobile interfaces will not be far behind. Imagine if, instead of tapping and swiping through apps, you described what you wanted, and an on-device model rendered it instantly on screen, like Siri, but actually useful.

It is easy to picture: you tell your phone, “show me my unread messages, then summarize today’s market news,” and it dynamically composes the screens, fetching the right content and layout on demand. The LLM becomes not just a voice assistant but the UX layer itself.

Even in gaming, I have seen early pitches where LLMs generate what appears on screen frame by frame, worlds rendered in real time at 120 frames per second, based on what the model “knows” should happen next. It sounds futuristic, but the underlying technology is already possible, and it is improving fast.

I’m Torn Because Good UX Still Wins

But I am torn. History has shown that elegant abstractions do not always win; great user experience does.

In 2011, Facebook launched its first dedicated mobile app built on HTML5. It was ambitious: one codebase to run everywhere. But it was also slow, clunky, and frustrating. Scrolling lagged. Transitions stuttered. Users hated it.

When Facebook switched to fully native apps, Objective-C on iOS and Java on Android, everything changed. The UI was fluid, load times dropped, and engagement soared. Native performance simply felt better.

That lesson still holds: smoothness, speed, and tactile feedback matter.

ChatGPT Atlas, as exciting as it is, reminds me of Facebook Mobile in 2011, visionary but not yet delightful. It is a glimpse of the future through a foggy window.

The Next Paradigm: The Everything-UI

If the 2010s were about native apps, what wins in the 2030s? How do we evolve from a unified, text-based chat interface into something that satisfies our deeper needs, such as play, communication, creativity, and exploration?

The challenge is not technical. It is experiential.

An “everything-UI” would need to support gaming, messaging, social feeds, video watching, and personalized creation, all seamlessly mediated by an LLM.

But that raises new questions:

• How do intellectual properties like Fortnite exist in a world without fixed app ecosystems?

• How do creators and influencers maintain reach when the LLM algorithms define discovery?

• What happens to distribution when the interface itself curates what we see?

We are standing at the edge of another platform shift. Atlas is the proof of concept, the HTML5 Facebook of this era. Somewhere ahead lies the native equivalent, an LLM-driven interface that feels instant, alive, and deeply human.

The future is not only about smarter models. It is about how those models reshape what interacting with technology feels like.

Because even in the age of AI, good UX still wins.

I’ve spent over twenty years in the games industry. If there’s one question I’ve been getting a lot lately, it’s this:

“Why didn’t you start another mobile games company?”

It’s a fair question. I’ve been part of every major wave in gaming, from the early days of Facebook games to the mobile revolution. But this time, I chose a different path. Instead of starting another gaming company, I decided to build an AI startup. You can read about it in this LinkedIn post.

Here’s why.

1. The Timing Just Isn’t Right

Every major shift in gaming has been driven by a distribution revolution.

When Facebook opened its platform to developers in 2007, it created an incredible opportunity. A few years later, the rise of the App Store sparked the mobile boom, the best possible time to build a gaming startup.

But right now, there’s no clear “new platform moment.” The market feels mature, and there isn’t that same asymmetric opportunity that allows small teams to challenge incumbents. Without that, starting a new gaming studio feels like running uphill.

2. I Don’t Have a Competitive Edge

In my previous startups, I always had a unique angle, a sense of timing, experience, and a view into something others hadn’t yet spotted.

Today, I don’t feel I have that edge in gaming. The industry is saturated, and the established players are incredibly strong. New studios often survive only by

1. copying fast and adapting even faster.

2. raising massive amounts to build something a massive game.

That’s not how I want to spend my time. I’ve done it before. I know what it takes. And I know what it costs.

3. I’d Rather Support Founders Than Compete With Them

Through F4 Fund and as an angel investor, I’ve invested in close to 50 gaming studios. I get to see the next generation of founders up close, their creativity, their ambition, their struggles.

I can add more value by helping them than by becoming one of their competitors. I understand their challenges, the funding environment, and the brutal realities of game distribution and player retention. My experience serves them better from the outside than from within.

4. AI and Gaming Don’t Yet Fit Naturally

Yes, AI brings massive productivity gains. Smaller teams can build faster and smarter than ever before. But that doesn’t necessarily translate into new distribution power, which is where the real breakthroughs in gaming historically come from.

AI will absolutely reshape how games are made. But I don’t see it creating the kind of once-in-a-decade opportunity that mobile or social platforms did. Not right now at least.

5. I Know How Hard It Really Is

Making a great game is brutally hard. You have to kill countless prototypes. You never really know if something works until thousands of players are playing it.

It’s an emotional rollercoaster that takes years of your life.

After two gaming companies and two decades in the industry, I’ve learned how much persistence and luck it requires. I simply don’t feel the same drive to go through that cycle again.

6. AI Feels Like the Right Place for Me Now

When I build AI products, I feel like I’m the end user. I know instantly if what I’m making works or not. I can iterate daily, talk directly to my users, and see results fast.

That feedback loop gives me energy. It reminds me why I started building in the first place, not to chase market timing, but to create something useful and meaningful.

Final Words

In many ways, I’m still building products. I’m still solving problems for users. It’s just that now, my playground isn’t the games industry, it’s the world of AI.

And that feels like the right place to be.

I recently shared a LinkedIn post about how VCs often avoid saying outright that they do not like your team or that the market you are attacking is not attractive. One commenter asked if I could share a framework for how founding teams are evaluated by VCs.

Here is my attempt at such a framework, based on my experience as an angel investor and VC.

1. Motivation

This is about the founders’ hunger and motivation toward what they are building. The perception among investors is that in a competitive market, the founders who are willing to work until 2 a.m. will win over those who only want to work nine to five. I know this is unhealthy and unsustainable, but founders need to show that at critical times, working late into the night is not frowned upon. Building a startup is never easy, and taking it easy will not improve the odds of success.

2. Founder-Market Fit

Do the founders deeply understand a market opportunity that fits their background and experience? The strongest cases are when the opportunity feels almost inevitable for this team to pursue, whether it draws on what they have built before or on a pain they have lived through. Founder–market fit is not just about one person but the entire team. Even in saturated markets with entrenched incumbents, a team with insider knowledge, strong relationships, and relentless energy can break through. I have also noticed that cultural context plays a role, as I wrote about in my piece comparing Helsinki and Istanbul.

3. Venture Journey Fit

Not every founder or business is suited for venture capital. VCs ask a simple question: “Can this founder return the fund?” The company must be able to reach an outcome large enough to matter in the venture model, where only a few investments drive most of the returns.

This fit is about mindset as much as market size. Are the founders ready for the pace of scaling and the pressure of aiming for a billion-dollar outcome? Many strong companies are better off bootstrapped or financed differently, and misalignment here can waste years. It is not only about exit potential but also the likelihood of raising Series A, B, and C rounds. That is what the founder is signing up for, and VCs will assess those possibilities.

In short, a strong fit shows up when founders understand the expectations of hyper-growth, show ambition that matches the opportunity, and demonstrate resilience and leadership to manage large amounts of capital. Weak fit shows up when the market is niche, the ambition is capped, or the founders are unsure about the trade-offs of the VC path.

4. Team Complementarity

What matters is not just raw talent but how the founders’ skills fit together. Investors look for coverage across product, technology, go-to-market, and leadership. In most cases, a team of only product thinkers or only engineers leaves critical gaps. A balanced team can execute faster and adapt better.

Complementarity also shows in how founders divide responsibilities and handle conflict. The best founding teams are greater than the sum of their parts, while redundant skills often signal hiring gaps or future struggles. It is true that raising financing makes it easier to bring in missing talent. But motivation may be weaker if people only join once a paycheck is guaranteed.

5. Execution Track Record

Ideas are abundant, but the ability to deliver separates strong founders from the rest. VCs look for past entrepreneurial projects, relevant jobs, side projects, or even small wins that show follow-through.

What VCs look for:

• Evidence of follow-through: past startups, relevant jobs, side projects, or even small wins

• Shipped products, launched features, or milestones hit under pressure

• Clear bias toward action and turning vision into reality

• Measurable results from prior roles (growth metrics, revenue impact, user adoption)

What raises concerns:

• No history of shipping products or delivering outcomes

• Frequent role changes without tangible achievements

• Overemphasis on ideas or strategy with little proof of execution

• A track record that suggests talk exceeds follow-through

6. Learning Velocity

Startups move fast, and the best founders learn even faster. VCs look at how quickly a team adapts when KPIs stall or customers say no. Learning velocity is about resilience as much as speed. Strong founders treat setbacks as data and sharpen their focus with every signal. Teams that iterate rapidly, admit mistakes, and turn feedback into gains compound progress. The opposite is founders who cling to failing plans or show little improvement. For VCs, learning velocity is often the clearest predictor of survival in early growth.

In gaming, I have seen teams build a game, launch the MVP to get the KPIs, and then pivot straight to another game when the numbers disappoint, without stopping to reflect on what they actually learned beyond “this game wasn’t fun.” I have also seen founders ignore lessons from others, like trying to re-skin a women’s casual puzzler into a male-oriented game with swords and combat, only to burn money when CPIs shoot past $20. The knowledge is already out there, with countless examples of those experiments failing, yet people repeat them. And when someone points to Empires & Puzzles as a counterexample, the answer is not that Match3 magically worked for men, but that the real hook was the PVP competition and character collection layered on top. That is what strong learning velocity looks like: not just reacting to your own KPIs, but synthesizing insights from the market and from others before repeating the same mistakes.

7. Clarity of Vision

Investors want founders who can see beyond today’s product. Strong founders describe where the category will be in five or ten years, why it matters, and how their company can lead it.

Clarity of vision also applies internally. The best founders set a clear north star for the culture, values, and collaboration inside the company. Without this, teams chase short-term wins and lose sight of the bigger picture. VCs back leaders who can inspire customers, employees, and investors with both a market vision and a vision for the company itself.

8. Founder Dynamics

The chemistry between co-founders is often just as important as their individual skills. Investors want to know if the team can trust each other, make decisions under pressure, and resolve conflict in a way that strengthens rather than weakens the company. Strong founder dynamics create stability in the chaos of startup life, while weak dynamics are one of the most common reasons early companies fall apart.

Believe me, I have been there. My first venture-backed startup I started alone, and when the hard times came, all of the weight fell on my shoulders. At Next Games I corrected that mistake. I had several co-founders, and together we were pushing the same boulder up the hill.

What VCs look for:

• High level of trust and aligned motivations between co-founders

• Clear role division and complementary responsibilities

• Healthy decision-making that balances speed with input from all sides

• Evidence of constructive conflict, where debates strengthen ideas but end with unity

• Consistent communication and shared direction

What raises concerns:

• One founder dominates every discussion or decision

• Unbalanced equity splits that suggest future tension

• Slow or indecisive decision-making under pressure

• Visible tension or unresolved conflict during meetings or pitches

• Founders who avoid hard conversations instead of addressing them

9. Integrity & Self-Awareness

Integrity is non-negotiable for investors. A founder who exaggerates numbers, hides bad news, or spins half-truths may survive a pitch, but they will not build lasting trust. VCs look for honesty in how founders present their progress and challenges, and for transparency in how they handle setbacks. The early signals often come in small details: are they consistent in their story, do they own their mistakes, and do they give credit where it is due?

Just as important is self-awareness. No founder can do everything, and the best ones know where their strengths end and where others should take over. They are willing to recruit people better than themselves, to ask for help when needed, and to learn from those around them. Investors back leaders who are both confident and humble, who can inspire others while admitting they do not have all the answers. A founder who combines integrity with self-awareness creates an environment where teams trust leadership, and that trust compounds into stronger execution.

10. Early Talent Magnetism

Hiring great talent is one of the keys to success for VC-backed startups as they begin to scale. One of the strongest signals VCs look for is whether founders can attract great people before the company has money, traction, or brand recognition. Early hires and advisors usually take a leap of faith, betting more on the founder than on the product itself. If talented individuals are willing to join early, it shows that the founder can inspire belief in the mission and build the kind of loyalty that carries a startup through uncertain times.

Early talent magnetism is also a proxy for future recruiting ability. Scaling a company requires hiring dozens or even hundreds of people, and no investor wants to back a founder who struggles to convince anyone to come on board. Founders who can pull in high-caliber engineers, operators, or advisors early are demonstrating that they will be able to build the networks, teams, and culture needed for long-term growth. Weak talent attraction, by contrast, often points to future hiring struggles that can slow down even the best product ideas.

Red Flags

Charlie Munger once said, “All I want to know is where I’m going to die so I’ll never go there.” Founders can apply the same logic to their own journey. It is not only about knowing what makes a strong team, but also about being brutally honest with yourself about the patterns that destroy one. If you recognize the places where startup teams most often fail, you can steer clear of them before they drag you down.

Here are the top six red flags seasoned VCs look for:

Shaky Motivation

If the founder’s main driver seems to be money or chasing a hot trend rather than a deep connection to the problem, it raises doubts about long-term resilience.

Weak Founder-Market Fit

A team without relevant experience or unique insight into their chosen market looks interchangeable, which means higher execution risk.

Dysfunctional Team Dynamics

Unequal commitment, poor equity splits, or visible tension between co-founders often foreshadow bigger problems down the road.

Lack of Learning Velocity

Overconfidence without evidence of iteration is a warning sign. VCs want founders who adapt quickly to feedback, not those who dig in stubbornly.

Hiring Blind Spots

If a team struggles to attract strong first hires or insists on doing everything themselves, scaling becomes a bottleneck.

Mismatch with Venture Scale

Sometimes the founder is building something solid, but not venture-backable. If the opportunity cannot plausibly return a fund, VCs will pass no matter how good the product is.

In short: Red flags often boil down to misalignment. Misaligned motivations, mismatched markets, or misfit ambitions can sink a startup long before the product sees the light of day.

Case Study: Lovable

The vibe coding startup Lovable shows how these traits come together in practice. When investors lined up to fund the company, they were not just betting on a product. They were betting on a team that embodied founder–market fit, early execution signals, learning velocity, vision, and talent magnetism.

Founded in 2023 by Anton Osika and Fabian Hedin, Lovable built on the momentum of GPT Engineer, Osika’s open-source project. According to some sources, Osika left his earlier startup Depict because it was not cutting edge enough, and he set out to create something bolder. By late 2024, Lovable had raised a $6m pre-seed round, and within months was reporting $17M ARR and 30,000 paying customers.

Founder-Market Fit

Co-founder Anton Osika had already built GPT Engineer, an open-source project with strong developer adoption. That gave him credibility and insider insight into AI-assisted coding, which was a natural springboard into Lovable. This was equivalent to building something and showing that demand existed long before “vibe coding” was even coined.

Early Execution Signals

With minimal funding, the team turned GPT Engineer’s momentum into a commercial product and quickly converted thousands of users into paying customers. By early 2025, Lovable reported $17M ARR and 30,000 customers, an unusually strong traction curve for a pre-Series A company.

Speed and Learning Velocity

Investors noted how quickly the founders iterated on feedback. Lovable’s team shipped features at a pace that matched the fast-moving AI landscape, a trait VCs see as critical in frontier markets.

Vision and Ambition

Lovable’s stated mission to make “everyone a software engineer” is big enough to support a venture-scale outcome. VCs saw founders who were not afraid to take on incumbents and reimagine how software is built.

Talent Magnetism

The team was able to attract strong early hires and high-profile investors, including Accel, Creandum, and Hummingbird. That kind of validation creates a reinforcing loop where great people want to work with other great people.

Final Words

When VCs evaluate founding teams, they are not looking for perfection. They are looking for evidence: hunger, fit with the market, the ability to learn fast, and the strength of the dynamics around them. The checklist may look simple, but living up to it is hard.

As a founder, you do not need to score ten out of ten on day one. What matters is that you show the seeds of these qualities, that you are aware of your strengths, and that you are honest about where you need to grow. The best investors are not just betting on your product, but on your capacity to become the kind of founder who can scale a company into the future.

There have recently been a lot of discussions in the games industry about the high salaries we’ve been seeing. People have struggled with the fact that someone might need to make $300,000 or $400,000 a year in a startup. Of course, salaries vary across different markets based on supply and demand for talent. Different societies have different costs of living; for example, the US is much more expensive than the Nordics, Turkey, or other parts of Europe (excluding London).

I’ve looked at thousands of pitch decks over the last six years as an investor in the games industry, and I’ve observed countless US investment rounds being raised with the knowledge that they’ll have a massive burn rate before shipping a game. Then the question becomes: is this something I want to participate in? Basically, I’m questioning whether it’s a better place to put venture dollars with a very expensive team and high valuation, because you can’t really dilute the founder’s ownership. Those two issues go hand in hand, and I wonder if it makes sense to place bets in these “expensive” startups.

From my experience as a founder and investor for the past 20 years, I know that Silicon Valley VCs traditionally didn’t invest in European companies because attending board meetings across the world was too burdensome. Zoom calls and the pandemic made it easier for investors to invest globally, since everyone was on video.

Still, US investors tend to prefer US companies over European or Asian ones due to geographic pattern recognition and familiarity with legal jurisdictions. Talking to founders during office hours doesn’t require super early or late calls around the globe. That said, they do invest in great European companies when they see them. It has happened countless times, especially in the later stages.

But they aren’t concerned about massive salaries or million-dollar monthly burns. US VCs don’t care if founders pay themselves $400K. They expect most of their portfolio to fail, so the only metric that matters is whether the winners can return the fund. In that lens, founder salaries are noise compared to market timing and product potential.

Here’s what one US VC who has invested in gaming told me, with permission to post:

We invest in outlier teams with the potential to build massive, enduring IP or platforms. A US team may be expensive, but if they can ship a genre-defining game, own a new niche, or build a billion-dollar community, they’re worth every dollar.

Many European founders take pride in running lean, while in the US, ambition is often equated with spending boldly. This isn’t just about money, it’s about narrative. A US founder asking for $400K may be framing themselves as a proven operator building a billion-dollar company, while a European founder would be seen as out of touch if they asked for the same.

There’s also a trade-off question that rarely gets asked: does paying yourself a high salary as a founder make you less hungry, or does it make you more resilient? I’ve seen both sides play out. On one hand, a founder who draws $350,000 a year might lose the urgency to prove their company’s worth before the cash runs out.

On the other hand, being financially secure can enable better decision-making. A founder who doesn’t need to worry about rent or school tuition might think longer term and avoid desperate pivots. In some cases, the high salary is what keeps a founder in the game when the grind gets tough.

The question is really about alignment. In a startup, every dollar you take out as salary is a dollar not spent on extending runway, hiring someone, or doing something that pushes the company forward. Founders should ask themselves: am I creating the best conditions for the company’s survival, or am I protecting my own comfort first? There’s no universal answer. Some founders can pay themselves well and still build enduring companies. Others discover too late that their own compensation shortened the very runway they needed to find product-market fit.

"Punching above your weight" is a boxing metaphor that refers to performing better than what would be expected given your size, resources, or experience level. In boxing, weight classes exist to ensure fair competition. When a boxer competes successfully against someone in a heavier weight class, they're literally "punching above their weight."

In business and startup contexts, it means achieving results that surpass what would be expected given your company's size, funding level, or stage. For example:

• A small team acquiring users at rates similar to much larger, better-funded competitors

• A startup with a limited marketing budget creating viral content that rivals established brands

• A startup developing sophisticated marketing strategies typically seen in mature organizations

When investors say they want to see founders "punching above their weight," they're looking for evidence that you can do more with less and outperform expectations, suggesting you'll be even more effective with additional resources.

The startup world has previously seen small teams do massive things. Instagram had only twelve people when Facebook acquired them for one billion dollars. WhatsApp was bought for $20bn and they had less than a hundred employees. This underdog effect has always existed in startups, but AI takes it to a whole new level.

Think about the impact of AI. Before LLMs, a startup's output scaled linearly with headcount. Now, in 2025, AI gives every founder the power to punch above their weight, making it the most disruptive force in the history of technology. A solo founder can now feel capable of building things that once required 10 engineers.

I've spent the summer deep in AI coding, trying things out to see what AI coding is all about. A year ago, all that felt more like an unrealistic future, that AI would be able to write code that is production-quality.

Here are three key things I’ve observed while building with AI.

1) A new kind of magical pair

No longer do you need to have a team of developers. One developer can run with an idea for a tech startup and get it off the ground on their own, with the help of their magical companion, an AI agent. For example, Claude Code can do so many things: it can design and implement websites, mobile apps, games, marketing campaigns, debugging code, etc. The best combo is for the developer to be a Swiss Army Knife product manager, who knows about business opportunity, markets, ideas, novelty, which aren't the strength of the coding agent. AI can take ownership, AI can come up with ingenious ways of implementing a feature, and it can run all sorts of tests.

The more hours I spent in Claude Code, the more I unlocked its power. Every session felt like playing chess with infinite strategies, and each move taught me new ways to master the tool.

Success with AI starts with a mindset shift. The founder must treat AI as a collaborator that works alongside them, not as a disposable tool for one-off prompts.

2) Orchestration

What shifted my perspective was seeing how developers can design orchestrations that AI carries out flawlessly. The speed and quality of new updates to these tools have been staggering. All this has made me a firm believer in AI coding for production-ready apps and games.

I've become accustomed to having multiple Claude Code instances running and building different parts of the app. The UI and UX is built by one, while the second Claude Code instance is working on game logic, and the third one is building a level editor for the game.

A note on vibe coding: you cannot one-shot a production-grade app. These tools must be learned. As with any creative software, skill comes from practice, not from simply winging it. Success in AI coding requires a thoughtful understanding of system components and their interactions.

3) The codebase question: will it be a hassle to maintain?

One of the biggest risks with AI-generated code is ending up with a codebase that becomes difficult to maintain. Quick wins and one-shot prototypes can impress your friends and make you look like a genius. But without structure and consistency the code quickly turns into technical debt.

To avoid this, you still need the fundamentals of software development: clean architecture, documentation, and regular refactoring. AI speeds things up but does not eliminate the need for discipline. The real breakthroughs come when AI subagents are given defined roles, clear rules, and relevant context. In areas like refactoring, they can work wonders. With so many tools emerging right now to tackle these challenges, these are truly exciting times.

Claude Code is basically getting a new version every day. EVERY, SINGLE, DAY with something new and improved.

Final word

Will early stage VC funding be necessary? Yes, there will still be a need for funding, as AI won't be able to scale your business and the TikTok ads you'll want to buy will still have an upfront cost associated with them.

But do founders really need to raise millions before finding product-market fit? Unless you just want to pay yourself a corporate salary for two years, why not start with personal savings or a few hundred K from angels and get moving?

This summer I built four mobile game MVPs using Claude Code. No big team, no heavy coding. Just ideas, iteration, and testing with real users. Along the way I learned what actually works when you want to get from concept to prototype fast.

Now I’m putting that experience into a premium course for founders who want to move quickly without hiring developers. We’ll cover setting up Claude for mobile development, building your first app in 48 hours, testing with users before you invest heavily, and knowing when to kill ideas versus double down. It’ll include my proven prompts, MVP templates, weekly live debugging sessions, and a private builder community. If you’re interested, join the waitlist here: AI Game Coding Course Waitlist.