AI is the most powerful amplifier ever built. What are you amplifying?

A founder I know built four projects in six months with Claude Code. A crypto startup, a sync tool, a trading bot, a knowledge graph. Fast, functional, shipped. By month six he was burned out — not from overwork, but from producing a lot without any of it feeding what actually mattered.

AI didn’t cause this. It accelerated it.

Speed without direction is just faster drift

Most of what gets written about AI is about leverage: how to do more, ship faster, build with fewer people. That’s real — the leverage is genuine. But it frames AI as a neutral tool, and it isn’t neutral. It’s an amplifier. And amplifiers don’t care what signal they’re boosting.

If you know where you’re going, AI gets you there faster. If you don’t, it accelerates the confusion. The problem isn’t the speed — it’s that speed makes the confusion harder to notice until you’re six months in with four unfinished projects and nothing to show for it.

The question almost nobody asks first

Before “how do I use AI?” the useful question is “what am I actually trying to build?”

Not the product. Deeper than that. What problem matters enough that you’d keep working on it when it gets hard? What outcome, a year from now, would make the work feel worth it? What’s the thing you’d build even if AI hadn’t made it faster?

Most founders skip this step. It feels obvious. It rarely is.

The strange attractor

There’s a concept in chaos theory: even in chaotic systems, there’s a hidden center of gravity — a “strange attractor” that the system keeps orbiting. Apparent randomness with a hidden structure.

The equivalent for founders and engineers is your set of values, obsessions, and the kind of work that energises rather than depletes you. Know it, and you can navigate AI-enabled speed without losing your direction. Don’t know it, and every new capability — every model release, every tool update, every “you could build X in a weekend” — is another chance to drift.

What this means practically

AI didn’t change the fundamental problem of knowing what to build and why. It made the cost of ignoring that problem much higher — because now you can build the wrong thing very, very fast.

The founders who’ll get the most out of AI aren’t the ones who use it most aggressively. They’re the ones who use it in service of a direction they’ve already thought through.

Speed is a feature. Direction is the prerequisite.