AI Rewards Curiosity

The people getting the most out of AI right now aren't always the most technical. They're the most curious.

The VTuber project started with a dumb question: could I run two AI personas with different voices and switch between them live? That led to voice synthesis latency (how do you hide 500ms of delay?), which led to multi-character performance design, which led to real VTuber practitioners like Vedal987 and CodeMiko, which led — somehow — to a four-persona psychological framework for understanding how identity works in performance. I started with a tech question and ended up in philosophy. I didn't plan the route. The curiosity pulled me through it.

The dominos thread was the same shape stretched over two years. It started with "can I build a browser game for my family?" and spiraled into game state representation, notation systems modeled on chess algebraic notation, tournament bracket logic, spinner layout algorithms, and eventually training AI models to play. Twenty-six conversations deep, and the question I was answering had nothing to do with the question I'd started with.

That's the shape: curiosity generates high-quality constraints. You don't ask for "an AI app." You ask for a weird, specific thing you actually care about. The weirdness is useful because it forces concrete tradeoffs, and concrete tradeoffs are where learning happens. A physics parody song forced me to understand anti-de Sitter space because surface understanding can't rhyme. A dominos game forced me to understand game state representation because the board has to render correctly. The constraint pulls you deeper than studying ever would.

This is why heavily optimized prompting can plateau. A perfect prompt often gets you a predictable answer. Curiosity gets you a better question, and better questions pull you into architecture, data, UX, and implementation details you didn't know to ask for.

My cousin Alex had the same experience. He went from zero programming knowledge to building websites with AI in days — not because the AI was a good teacher, but because his questions kept getting better. Each answer opened a door he hadn't known was there, and he kept walking through. The curiosity spiral works whether you're technical or not. The prerequisite isn't skill. It's engagement.

Curiosity is the real prompt engineering.