The recent rise of LLMs created an interesting and seldom discussed (until recently) niche - the use of AI for therapy. There are shades of gray to this, and they share a common thread. AI can be used to make numerous areas of your life easier: work, decision making, navigating emotional situations, or even social situations. All of this leads to one outcome: the outsourcing of your mental effort. What should be happening within, causing neural changes, learning, and building resilience, has now been outsourced to the oracle that is ChatGPT (or Claude, or Gemini, or...).

Not every product needs to chase venture scale. Some of the best tools are tiny, opinionated, and built by one person who cares a little too much. We live in an era that celebrates platforms. The default ambition is to build something that serves millions, raises rounds, and eventually becomes infrastructure. But there's another path—one that's quieter, more personal, and arguably more impactful per user. **Small software** is software *built* for a `specific person` with a specific problem. It doesn't need onboarding flows or analytics dashboards. It doesn't need to "scale." It just needs to work really well for the people it was made for. ```typescript 12345 ``` I think about the tools I reach for every day—the ones that feel like they were designed by someone who uses them. There's a texture to that kind of software. The defaults are right. The edges are smooth. You can feel the care in every interaction. The best small software comes from a place of personal frustration. You build it because nothing else exists that does exactly what you need. And then you share it, not because you want to build a company, but because you suspect other people might feel the same friction. ${'/neuron1-sim.gif'} There's freedom in building small. You can make opinionated choices without worrying about alienating a market segment. You can ship fast because there's no stakeholder review. You can change direction overnight because the only roadmap is in your head. The internet has made distribution free, and AI has made development faster. The barriers to building small, personal software have never been lower. We're entering an age where a single developer can build something that genuinely changes how a small group of people live their lives. That's not a consolation prize. [That's the whole point.](https://google.com)

For years I waited to be good before sharing anything. Lately I've been experimenting with the opposite approach: publishing half-formed thoughts and letting them breathe. There's a particular kind of paralysis that comes from wanting to be taken seriously. You draft something, read it back, and think: "Who am I to say this?" So you close the tab and tell yourself you'll publish it when it's ready. It's never ready. I spent most of my twenties in this loop. I had notebooks full of ideas, half-written blog posts, projects that never saw the light of day. Not because they were bad, but because they weren't perfect. And perfection, I've learned, is just fear wearing a nice outfit. Learning in public means publishing before you're comfortable. It means writing about things you're still figuring out. It means being wrong sometimes and letting people see it. The surprising thing is that people respond to honesty more than expertise. When you share your process—the confusion, the breakthroughs, the dead ends—it resonates in a way that polished thought leadership never does. People don't want to read another person who has it all figured out. They want to read someone who's figuring it out in real time. There's also a selfish reason to learn in public: it makes you learn faster. When you know someone might read what you write, you think more carefully. You check your assumptions. You organize your thoughts in a way that casual note-taking never demands. I'm not suggesting you tweet every half-baked idea. But I am suggesting that the bar for "ready to share" is much lower than you think. The internet is big. Most things you publish will be read by a handful of people. And those people will be grateful you shared. So here I am, learning in public again. Some of these essays will age well. Some won't. That's the whole point.

Every tool that saves us five minutes also takes something invisible in return. I've been trying to notice what I'm trading away when life gets too smooth. Convenience is the organizing principle of modern technology. Every product pitch, every feature update, every redesign promises to make something easier, faster, more seamless. And they deliver. Life in 2026 is astonishingly convenient compared to even a decade ago. But convenience has a cost, and it's one we rarely account for because it's paid in a currency we don't track: engagement with the world. When you order food through an app, you lose the conversation with the person at the counter. When you automate your finances, you lose the monthly ritual of looking at where your money went. When GPS handles every drive, you lose the mental map of your city that used to live in your head. None of these losses are catastrophic on their own. But they accumulate. Slowly, without noticing, you can find yourself living in a world that requires very little of you. And a life that requires nothing of you starts to feel like it has nothing to offer you. I've been running small experiments. Cooking without a recipe. Driving without navigation to places I sort of know. Writing with the door closed before asking anyone for feedback. The friction is real, and sometimes I fail. But the engagement is real too. I'm not a luddite. I love technology. I build technology. But I think we need a more honest conversation about what we're optimizing for. Speed and ease are means, not ends. The end is a life that feels like yours—one you're actively participating in, not just consuming. The most convenient life isn't necessarily the best one. Sometimes the best life has a little grit in it.

The internet used to feel like an empty field where anyone could set up a tent. Now it's a shopping mall with security guards, but there are still side doors if you know where to look. I built my first website when I was fourteen. It was terrible—Comic Sans, animated GIFs, a visitor counter that I'm pretty sure only counted my own visits. But it was mine. I didn't ask anyone if I could make it. I didn't need a platform's approval or an algorithm's blessing. I just made a thing and put it on the internet. That feeling—of building something and sharing it without gatekeepers—is what drew me to technology in the first place. And somewhere along the way, we lost it. The modern internet is organized around permission. You need a platform to reach an audience. You need an algorithm to be seen. You need to play by rules you didn't write and can't read, rules that change without notice and optimize for engagement rather than quality. But here's what I've noticed: the permissionless internet isn't dead. It's just less visible. People are building personal websites again. They're starting newsletters that bypass social media entirely. They're creating software and distributing it directly to users. The infrastructure for building without permission has never been better—it's just that most people have forgotten it exists. Building without permission means accepting a smaller audience. It means your work won't go viral because there's no algorithm to amplify it. It means growing slowly, through word of mouth and genuine connection. That sounds like a downside until you realize what you get in return: creative freedom, ownership of your work, and a direct relationship with the people who care about what you make. I think we're at the beginning of a quiet rebellion. Not loud, not organized, not branded. Just individual people deciding to build things on their own terms and share them with the world directly. This website is part of that. No analytics. No tracking. No algorithm. Just words on a page, waiting for the right person to find them.