AI now writes code, designs interfaces, analyzes data, runs tests. What took a team five years ago, one person and a few agents can handle today. Execution has become cheap. Competence has been commoditized.
But there's something AI can't do: decide whether something should exist.
That requires taste.
Taste isn't aesthetics
People confuse taste with visual preference. But taste isn't about what you like — it's about what you've seen, what you've experienced, and what conclusions you've drawn from it.
A Linux kernel maintainer recently rejected a patch that AI-based review had flagged as flawless. Not because it was buggy — because it solved the problem in the wrong place. Technically correct, architecturally wrong. You don't learn that distinction from code. You learn it from years of consequence.
When designing Phora, this kind of judgment mattered most: not what I could build, but what was worth building.
Why it matters now
As long as execution was the bottleneck, taste was a luxury. Now that anyone can ship fast, the difference is in what you ship.
Steve Jobs said it in 1995: "Ultimately, it comes down to taste. It comes down to trying to expose yourself to the best things that humans have done and then try to bring those things into what you're doing."
What's changed: AI didn't create the gap between people with taste and people without it. It just made the gap impossible to ignore.
How taste develops
Taste isn't innate. It's built from experience and mistakes — you ship something, watch it fail, understand why, and decide differently next time.
After 16 years of projects, I don't decide faster because I got smarter. I decide faster because I've seen the consequences of enough bad decisions.
AI is a perfect executor. But the filter through which decisions pass — that's still human.
There's always a next level.
If you like what you see — whether you're building a product or a team — I'd love to hear about it.