For years, I carried ideas around like loose change in my pocket. Waiting.
Waiting for the right co-founder. The developer who got it. The budget that could turn a sketch into something real.
I could make it pretty in Figma. But I couldn't make it exist.

Something shifted
Last year, I launched more products than I did in the previous decade.
Not because I got smarter. Not because I learned to code. But because the barrier between "I wonder if this would work" and "let's see if this works" collapsed.
Each time feels like magic because it kind of is.
The magic is this: I can finally have the conversation with the market that I've been rehearsing in my head for years.
Here's what changed
I used to need permission. From a technical co-founder. From investors. From the market.
Now? I need curiosity and a weekend.
My prototype isn't in Figma anymore. It's in the world. With real users. And I find out what works fast.
The hard part was never the building.
It was always the believing. Believing enough to start. Believing enough to ship. Believing enough to learn when it doesn't work and do it again anyway.

So what now?
The bottleneck moved.
It's not in the code anymore. It's in the question I'm asking. The problem I'm solving. The people I'm trying to reach.
Those were always the hard parts. I just couldn't see it clearly because I was too busy wishing I could code.
Which is why I'm doing something this week: I'm making a place where I'll publish all my experiments. The ones that worked. The ones that didn't. The half-baked ideas someone else might play with.
Because if the barrier dropped for me, maybe seeing what's possible helps it drop for you too.
The unfair advantage isn't the tools.
It's that we never needed the permission we were waiting for.
We just needed a way to test if we were right.
Now we have it.

A contributor submitted a pull request to Tailwind's GitHub on November 18, 2025.

The PR was simple. Add an endpoint called `/llms.txt` that would serve a text-only version of all Tailwind documentation, optimized for large language models to consume. Extract text from MDX files. Remove JSX components. Preserve code blocks. Make it easier for Claude and ChatGPT to read the docs.

Nobody responded.

For seven weeks, the PR sat open. Then on January 6, someone pinged the thread. "Why is this one not moving?"

Adam Wathan, Tailwind's founder, finally replied: "Have more important things to do like figure out how to make enough money for the business to be sustainable right now."

The next sentence explained everything.

"75% of the people on our engineering team lost their jobs here yesterday because of the brutal impact AI has had on our business."

Three engineers. Gone. Revenue down 80%. Documentation traffic down 40% since early 2023. Not because Tailwind isn't popular, it's bigger than ever. But because developers stopped visiting the documentation where they'd discover the paid products.

They're asking Claude Code instead.

Here's what twelve months changed.

Claude Opus 4.5 scored higher on performance engineering exams than any human candidate ever. It autonomously refined its own capabilities in four iterations, peak performance that other models couldn't match after ten. Claude Code worked for an hour straight creating hundreds of files, deploying a working website without human input.

OpenAI's Codex went from autocomplete to agentic engineer. GPT-5.2-Codex runs 30-minute autonomous coding sessions in large repositories. It completes complex refactors and code migrations without losing track. A security researcher used it to find and disclose vulnerabilities in React that required sustained, multi-step reasoning.

Both tools can now generate complete Tailwind implementations,spacing, shadows, that perfect shade of gray, in minutes. The $299 UI kit with fifty carefully crafted components? Five minutes. Maybe less.

The irony landed on GitHub: Someone wanted to make Tailwind documentation *more* accessible to LLMs. But LLMs were already killing the business. Traffic to docs fell 40% because developers were using AI tools that generated code without ever visiting the site. The PR that promised to help was exactly the problem.

DeepSeek's R1 model cost $5.6 million to train. American equivalents spent hundreds of millions. The capability gap closed overnight. The price collapsed. Then the models kept improving.

What Tailwind sold was never really the code. It was the shortcut. The trusted answer. The decision you didn't have to make about spacing for the hundredth time.

Now Claude Code makes those decisions. Codex makes those decisions. In terminal windows. In IDEs. In cloud sandboxes that spin up in seconds. For twelve dollars a month.

The documentation that used to teach people now teaches models. The components that used to convert customers now generate on demand. The business model that funded one of the web's most popular frameworks just said goodbye to three quarters of its team.

This was January 6, 2026. The layoff day.

The PR closed the same day. Still no merge.

Between last January and this one, AI didn't just get better at coding. It got better at being the shortcut. At replacing the moment when you'd search for an answer and land on someone's documentation. At making the visit, and the discovery that follows, obsolete.

What do you sell when the thing you packaged becomes the thing AI generates as a side effect?

Frank Gehry never learned to draw properly.

While other architects mastered technical drafting, Gehry crumpled paper. Built little models that looked like accidents. His hands thought faster than pencils allowed.

When CAD arrived, everyone assumed architects like him would disappear. The opposite happened. Gehry used software to translate what his paper models already knew. The tool didn't replace his thinking. It amplified it.

The architects who vanished? The ones who only knew the old tools. And those that didn't took advantage of the new opportunities.

That's the part most people miss.

AI isn't replacing creativity. It's reorganizing who gets to create, and how. The programmer who treats Copilot like a curiosity, breaking its suggestions, testing its limits, using it to think sideways, builds different muscles than the one who just ignore it.

The drafters who survived CAD weren't the purists or the converts.

They were the ones who kept crumpling paper. Who stayed curious enough to ask: what can I make now that I couldn't make before?

Your industry is mid-transition. The tools are arriving faster than the playbooks.

The question isn't whether to use them.

It's whether you'll use them to think...or to stop thinking altogether.

The Berlin Wall fell in 1989.

Thirty-five years later, countries that were freed are still learning how to be free. Children born into democracy still carry their parents' Soviet reflexes. The infrastructure of fear takes generations to dismantle, even after the system that built it is gone.

That's how long humans need to adapt to a world that changes once.
Now imagine we compress a century of change into a decade.
Cure most cancers. Eliminate infectious disease. Double human lifespan. Not eventually,by 2035. The math checks out. A country of geniuses in a datacenter, working in parallel, routing around the constraints that slow human researchers down.

It's probably possible. The biology might actually work that way.
But we're not just racing to build it. We're racing to build it first.

Because the same intelligence that could cure disease could perfect surveillance. The same systems that could eliminate poverty could eliminate dissent. An authoritarian state with super-intelligent AI doesn't just control information, it predicts behavior, optimizes propaganda, identifies dissidents before they know they're dissidents.

The technology that could free us could also lock the door forever.
It's not that the breakthroughs are bad. It's that humans aren't built for decades of change compressed into years. We need time to grieve what's obsolete. Time to rewrite the stories we tell about what matters. Time to figure out who we are when the constraints that shaped us disappear.
The future arrives faster than we can prepare for it arriving.
And we don't all get to choose which future.
---
Inspired by Machines of Loving Grace by Dario Amodei

Remember when every app announced its "cloud-powered" features?
You don't. Because the cloud became infrastructure. The competitive advantage disappeared the moment everyone had it.

AI is following the same path. Right now, companies are overpromising what it can do. They're adding "AI-powered" to pricing tiers. They're making it the story.
But here's the thing about good taste: it's knowing what people want just before they realize they want it. And what people want isn't AI. They want the result AI enables. They want autocomplete in their IDE that doesn't announce itself. They want electricity that just works.

The brands that survive this transition won't be the ones shouting about their AI features. They'll be the ones who kept a simpler promise: this works, and it works for you.
Your distribution strategy shapes everything. When AI was scarce, you could build a moat around access to it. When it's everywhere, your moat is something else. Maybe it's taste. Maybe it's trust. Maybe it's knowing which problems to solve.

Not every wave is worth catching. Some companies are rushing to ship AI features because everyone else is. They're compromising on the wave. They're adding AI because they think they should, not because it makes the product better.
The right wave comes when the technology stops being the story. When it's just infrastructure. When you can focus on what actually matters.
What will you promise when AI stops being special?

You chose your phone plan. You chose your career path. You chose where to live.
Except the choice was between options the system already selected.
Hunter-gatherers worked three hours a day. We work eight, ten, twelve, with tools that promised to save us time. The system didn't evolve to give us leisure. It evolved to extract more from us while making extraction feel like opportunity.

Your "choices" exist inside invisible boundaries. The algorithm shows you five jobs, not five hundred. The insurance company offers three plans, each designed to profit them, not protect you. Your calendar fills with meetings because the system values presence over output.
We call this freedom because we can't see the walls.

The rental market doesn't care if you're housed it cares that you're paying. The productivity software doesn't care if you're productive, it cares that you're subscribing.
Systems optimize for their own survival. You're the variable, not the equation.
What we mistake for free will is really just choosing which track to run on. The track was already laid. The destination was already set. The system just lets you pick your lane.

And the cleverest trick? Making those lanes feel like your idea.
The paths that don't serve the system the four-hour workday, the career that can't scale, the life that doesn't optimize they're not even on the menu. Not because they're impossible. Because they're unprofitable.
You can't choose what you can't see.
The question isn't whether you have free will. It's whether you can see the boundaries long enough to know what's outside them.

The Google Doc you're writing in right now isn't yours.

Close your laptop. Lose your password. Miss a payment. Watch the company pivot. Your work doesn't disappear because you deleted it. It disappears because they decided to.

We traded ownership for convenience. We got real-time cursors dancing across screens, updates that sync before we finish typing, the certainty that everyone sees the same version. Revolutionary.

What we lost was quieter.

Your thesis lives in a datacenter you'll never visit. The raw footage from your film exists at the mercy of a terms-of-service update. Every creative professional using cloud tools is building their life's work on rented ground.

The old way wasn't better at collaboration. Emailing "project_final_FINAL_last.xls" back and forth was absurd. But when you saved a file to your hard drive, it was yours. It worked offline. It worked forever. The software couldn't be taken away from you.

Here's the pattern we're missing: what if your data lived on your devices, syncing in the background, collaborating in real-time, but always yours first? Not as a backup. As the primary copy.

The technology exists. CRDTs, they call them. Conflict-free Replicated Data Types. It's possible to build software where you own everything and lose nothing.

But we built the cloud instead. Servers as the source of truth. Your device as merely a window into someone else's computer.

The question isn't whether we can make collaborative software that respects ownership.

The question is whether we remember why ownership mattered.

If you're in game development, you know this story already.

Your game has been in development for three years. You picked Unity because it was free. You built everything around it. Your tools, your workflow, your team's expertise. The switching cost is six months minimum, maybe a year.
Then Unity announces the runtime fee. They'll charge you every time someone installs your game. Retroactively. For projects already shipped.
If you're not in game development, you have a different version. The platform you built your business on just changed its API pricing. The social network where you found your audience just changed the algorithm. The tool you rely on just announced new "features" that break your workflow.

Cory Doctorow calls this "enshittification." First, they're good to users. Then they abuse users to please business customers. Finally, they abuse business customers to extract everything for shareholders. Then they die.

Here's how it happens. The founders start by solving a problem. They build something people actually want. They lose money doing it because the point is to help, to grow, to matter.
Then they hire product managers. The product managers have dashboards. The dashboards need numbers. The numbers need to go up and to the right, every quarter, forever.

Reddit spent fifteen years letting developers build clients that made the site usable. Then priced API access to shut them all down right before the IPO. Audible owns 90% of the audiobook market and locks everything with DRM. Leave the platform and your library disappears. Twitter let you see tweets without logging in for a decade. Then Elon bought it and put everything behind a wall.

The pattern is everywhere once you see it. Airbnb costs more than hotels now. Google search shows you ads before answers. Netflix raised prices after winning the streaming wars. The dating apps keep you single because retention matters more than matches.
They were good until you needed them. Then they were good until their vendors needed them. Now the dashboard says growth is slowing.

What are you building on?

>Punctuality is the politeness of kings. Louis XVIII

Something interesting happened after I wrote my last post. It's like I removed a firewall and ideas started flowing. They won't leave me alone until I write them down.
This one came to me on a train. Most of my ideas show up on the road, in cafés, while walking. That makes sense your brain is more active when relaxed, entering what's called theta rhythm. Want to boost creativity? Get moving.

Here's the thing about being punctual: I'm always early, always waiting. Over time, I've learned to appreciate this weightless moment. Like Neil Armstrong floating in space, you can discover ideas, call loved ones, or just drift in your inner world.

While waiting in a café, I watched a twenty-something guy walk in with his grandmother. Unusual. Elderly people in our culture don't often enter establishments like this.
My brain started filling gaps.

Maybe he's visiting for vacation and she mentioned this used to be their favorite spot, back when grandfather was alive. So they ordered coffee and éclairs while she shared memories.
Maybe she never accepts his compliments about her cooking, so he brought her here to prove his point.
The real story is probably more mundane. But for me, they'll always be the grandson and grandmother sharing something meaningful over coffee.

We do this constantly. Create narratives for strangers. Just like Humans of New York, we photograph moments and fill in the gaps.

Here's what I realized: those stories we invent aren't really about them.
They're about us. What we notice. What we value. How we see the world.
The person who imagines romantic stories sees romance everywhere. The cynic creates cynical narratives.

Your stories about strangers are actually stories about you.
And that makes people-watching the most honest mirror you'll find.
P.S. After writing this, I found someone actually created a guide for this (https://www.wikihow.com/Begin-People-Watching). Of course they did.
These are the thoughts that surface on trains with no internet.

I've started many times and procrastinated just as many times when it came to beginning my own blog. It always seemed to me that people should receive information that adds value to their daily lives, rather than experiencing feelings of emptiness and regret about wasted time, as often happens to me at the end of reading many blogs...
I'll write about why everyone should start their own journal: whether it's simply notes in a notebook or an online format! I hope you'll be able to learn from my mistakes .
(just don't forget that I, like you, am only beginning - let's call this test #1)
Briefly, how I started "Mark Twain-ing"

During my school years, when everything in the world felt overwhelming, I decided I'd make an excellent writer (though I was never much of a grammarian). I admired the lifestyle of writers, at least the one shown to me by films and media. So, after the movie "Secret Window" 2004, I gained a short-lived hobby and a lifelong favorite author (Stephen King). Most importantly, I started writing everything down and sketching.

**What this led to**
Today I write down everything that comes to mind, all ideas that might be useful in the future! And it doesn't matter if at the moment of writing I don't know how to use them.
You could call this process idea cartography. At the end of the week, you can go through all the notes and sketches and understand what kind of week it really was. This brings awareness and provides an opportunity to acquire some navigation.

Re-reading your own notes improves self-understanding. I especially notice this when I find old entries and remember what I was going through at that time and what values guided me. Understanding yourself also helps in understanding others - all of this develops your emotional intelligence.
Very often, when writing something down, you create a stronger connection between your idea and your brain, forcing it to work in the right direction.

A few years ago, I developed a small tradition - writing down my plans for the day in the morning. During the day - jotting down ideas that chaotically come to mind, and in the evening some conclusions or additions to what's already there. Unconsciously, I began to discipline myself and became slightly more organized (here "slightly" is already good). Habits are like muscles - the more you develop them, the stronger they become rooted and the less energy they require.
In difficult moments, when the solution to a problem isn't obvious, I write out the problem, breaking it down into small parts. Writing helps understand what the essence is. Only recently did I learn that this is a recognized method for solving emotional and psychological problems. (Dr. James Pennebaker)

Before important moments, entries about positive experiences helped me gain self-confidence, as the brain re-experiences positive emotions. This can be compared to a kind of self-support.

While writing this, I had a thought that all these lines written above remain. When we read books, we learn about others' stories, their character, ideas, adventures, and habits. We dedicate time and energy to reading so that after 500 pages we can find out how the story ends. Sometimes after reading, the thought slips through that no one will ever write about us like that. No one! except ourselves! All our entries are that thick manuscript of our life, which, if you want, you can preserve for your children. When I find old notebooks of my father or grandfather, it always sparks intense interest. Even my entries from five years ago, which I completely forgot about, can amaze me - how interesting, absurd, or innovative they were at that time.

In the Middle Ages, we were told what to believe by the Church. After the appearance of the printing press and the Reformation, we were guided by state censors and the licensors of publishers. Later, with the rise of liberalism in the 19th and 20th centuries, publishers themselves took control, followed by broadcast media—a small, elite group of professionals.

This is how the democratization of knowledge emerged: the acquisition and spreading of knowledge among common people, not just privileged elites such as clergy and academics. Libraries—public libraries in particular—and modern digital technology such as the internet play a key role in the democratization of knowledge, as they provide open access to information for the masses.

Examples of Knowledge Democratization
Even in the startup world, where you typically need to be in a formal accelerator to gain information on how to start and manage a startup, Y Combinator came up with Startup School. By democratizing startup knowledge, YC hoped to spur economic activity around the globe while promoting and preparing companies for its more formalized accelerator program.
The Internet's ability to democratize content creation in both South Korea's gaming and music industries has led to the rise of many new artists and game creators.

The Four Principles of Mass Collaboration
Here are the four principles from Don Tapscott and Anthony D. Williams's Wikinomics: How Mass Collaboration Changes Everything:

Openness, which includes not only open standards and content but also financial transparency and an open attitude towards external ideas and resources.
Peering, which replaces hierarchical models with a more collaborative forum. Tapscott and Williams cite the development of Linux as the "quintessential example of peering."
Sharing, which is a less proprietary approach to (among other things) products, intellectual property, bandwidth, and scientific knowledge.
Acting globally, which involves embracing globalization and ignoring "physical and geographical boundaries" at both corporate and individual levels.

Industry Adoption and Innovation
Nevertheless, some industries are adopting new concepts faster than others. Those that are more digitally oriented are changing faster, while the rest in the offline world grow at a slower pace.
The same transformation is happening in the fashion design world with the help of the Fennec&Finch platform—the one-stop place where you can ideate and produce your fashion collections, even without being technically skilled. All that is needed is an idea. Through this marketplace, any creative person can become a brand and launch their product easily. This will surely bring significant economic growth to a sector that can create multiple job opportunities in this particular market.

Let's talk about AI. Last week, I had an interesting conversation with one of my friends, and we ended up discussing artificial intelligence. For many years, we have been bombarded with the idea in movies and books that AI will destroy humankind and that there will be an ultimate Skynet that will rule the world.

Nowadays, AI helps people detect fake news and deepfake videos (even though some fakes are also created by AI). Many services use AI for navigation and route creation. And let's not forget the gaming industry, which relies heavily on AI for its characters.
The Reality of AI Today

As you can see, there are many types of AI developed by individuals or companies to perform specific tasks—usually tasks that no one wants to do but that need to be done. Of course, humans are typically hired to edit text or sort certain data, roles that can now be replaced by AI. This is an inevitable consequence of technological revolution and industrial change.
The same thing happened when factories introduced robotic assistants and automation in production. Many people lost their jobs, but the worst part is that their positions completely disappeared. However, that is not the main point I want to make.

The Democratization of AI
For the past few years, anyone with sufficient proficiency can build their own AI and teach it to do anything they want. Just think about it: you could have your own personal AI that you haven't purchased, and you could teach it, for example, to filter news and recommend the best articles that you would enjoy.

In Terminator, The Matrix, and Avengers, AI is described as unique and the only one in existence. In reality, I think there will be multiple AIs that will operate in fields ranging from simple to complex, solving minor to major problems. The keyword here is "solving."
Of course, you might argue that the planet's problem is the entire human race, and solving it could be good for the planet but bad for us. However, that depends on how we teach our AIs, and if we do it correctly, we will have a bright future that will arrive sooner than we expect.
The Acceleration of Progress

With the democratization of AI, the pace of progress will no longer be measured in years as it was before. We used to say that we needed 5–10 years to have electric cars or 1–2 years to develop personal assistants. With the popularization of AI, innovations could take hours to appear and days to implement. One year would be worth ten compared to our current state of technology.
This would mean faster medical care and cures for diseases. Architecture would become more accessible; even now, there is software and AI for designing buildings without requiring a bachelor's degree in the field. Every aspect of our lives will be affected. This means countless AIs, from small to large. Google and Facebook would have their giants; small businesses would have their specialized tools. Even countries could develop their own national AI systems.

Personal Reflections and Concerns
I like to think about AI as a valuable asset that will make our lives better. Of course, there is significant uncertainty, which gives me pause.
Elon Musk, in order to balance AI development, has developed Neuralink, which could bring cyberpunk scenarios into reality very soon. This may solve many problems, and I hope it will help people who are disabled due to neurological disorders.
Implications for Human Consciousness and Society

When you start thinking about it and consider various possibilities, an interesting question arises: how will this shape our consciousness? Not so long ago, we couldn't imagine that we would pay for products that aren't physical. Now, people pay for skins in video games, premium subscriptions, or XP boosts in mobile games and become very frustrated if something goes wrong with their accounts.
Another consideration is that innovation is not reaching all countries uniformly. This means that some countries will be the pioneers of AI and will have priority and advantages in this competition. I call it a race because this might resemble something like the Cold War—just not a war; let's call it the "Cold AI Race."

Consequently, countries that fall behind will have an economic status far behind those leading the race, and their people will experience a lower quality of life, especially if the country's main industry involved outsourcing services that AI can now perform.

The Digital Divide of the Future
Another concern is how unfair it might become for people without access to technologies like Neuralink compared to those who have it, and that's how the world might divide.
While writing this, I've realized that this is such an expansive topic that I become overwhelmed thinking about all the possibilities and challenges that await us in the coming years.