On AI

AI gets plenty of criticism. But for all its flaws, it's leveling the playing field in a way I haven't seen before.

The criticism that matters

The "AI will take our jobs" discourse doesn't bother me much. Technology has always pushed people to grow, adapt, and find new ways to be valuable. That's preferable to stagnation.

What does concern me is generative AI trained on the work of other creators. There's something uncomfortable about models learning from millions of artists, writers, and developers—often without consent or compensation—and then competing with them. That's a real problem worth solving, not dismissing.

But the tool itself? It's just a tool. How we build it, train it, and use it—that's where the ethics live.

The barrier is lower

You can now just ship things. It has never been more true.

I studied engineering, started as a designer, and for years barely wrote any code. Then AI changed that. I can now actually build what I want, prototype fast, contribute to code at work, and turn ideas into real products.

The traditional path used to be: learn for years, then maybe build something. Now it's: have an idea, start building, learn as you go. The tool fills the gaps while you're still figuring things out.

I don't have to spend months learning a language before I can use it. I focus on building, not on memorizing syntax. I focus on systems design, not on low-level framework details. Code is no longer exclusive to engineers. Even though I'm a computer science engineer by education, I never really wrote code—because I liked to design and build products. Now I can do both.

Building vs maintaining

Here's something I've learned: building and maintaining are completely different modes.

When you're building, you don't care where you go. Your changes are destructive and that's fine. You can literally change the entire trajectory of a project. You're exploring, not preserving.

Maintaining is different. You need to preserve functionality. You need to understand how things work so you don't break them while fixing bugs. You can't just ask AI to "fix this bug" if you don't understand the architecture or system design. The fundamentals still matter—especially when you're responsible for keeping something alive.

AI is incredible for building. For maintaining, it's a powerful assistant, but you still need to understand what you're working with.

Taste is the differentiator

AI doesn't replace taste, judgment, or understanding what makes something good. Those still matter—maybe more than before. What it does is remove the years of syntax memorization and boilerplate writing that kept people from even starting.

But here's the thing: taste is inherently human. It's hard to define, impossible to measure objectively. You can't rate a design the way you can validate code. Code is deterministic—it either works or it doesn't. Taste isn't.

For AI to have taste, it would need an inherent feedback loop and a definition of "good" to validate its output against. But you can't define that. What's good depends on context, culture, audience, and a thousand subtle things that shift constantly.

This is why taste is the real differentiator now. AI can execute, but it can't decide what's worth executing. That's still on us.

Faster cycles mean better products

When you can go from idea to working prototype in hours instead of weeks, you learn faster. You test assumptions sooner. You catch bad ideas before investing months into them.

At work, this changes how we build. Features that would've needed careful planning and specialist developers can now be sketched out, tested with users, and iterated on before committing to a full implementation. The feedback loop is tighter.

The fear is real

I'll be honest: I'm scared.

Not of AI taking my job, but of how much I have to change myself to adapt. The pace of change is relentless. The tools evolve faster than I can master them. I'm constantly learning, unlearning, and relearning.

And it's not just me—people way smarter than me are struggling too. That's both comforting and terrifying.

Adapting to AI isn't a one-time thing. It's a continuous process of staying relevant, staying curious, and accepting that the ground will keep shifting under your feet.

What this means

The gap between "I wish this existed" and "here's a working version" is smaller than it's ever been. That's powerful for anyone who wants to make things but didn't have the traditional path to get there.

I just love making stuff, and AI lets me do more of it. The cost is constant adaptation. The reward is actually building the things I imagine.

That tradeoff works for me.