The big picture

The new 0 → 1 is one local user to infinite users

FoxDog AI · 5 min read

For decades, the hard part of a startup was building the thing. "Zero to one" meant turning an idea into working software — and that took a team, months, and money. The build was the moat.

AI broke that. With Claude Code or Codex, one motivated person goes from idea to a working app in a weekend. The build is no longer the bottleneck. So where did the hard part go?

The build collapsed. Distribution didn't.

Building got an order of magnitude cheaper and faster. But the App Store gauntlet, the developer accounts, the review process, and the eternal question of "how do I get anyone to actually download this?" — none of that got easier. If anything it got harder: more apps, more competition, stricter store rules.

The result is a flood of apps that work perfectly… on exactly one machine. The graveyard isn't full of bad code. It's full of good apps that never shipped, or shipped and never found a single user.

1 → ∞ is a different game than 0 → 1

"Zero to one" — can you build it — is now a solved-ish problem for anyone with an agent and a weekend. The new frontier is one to infinity: going from the single local user who built the app (you, on localhost) to real users in the world.

And the skills don't transfer. Prompting an agent to build a feature has nothing to do with creating an App Store Connect record, passing review, or earning your first 100 users. The people who win won't be the ones who can build — everyone can build now. They'll be the ones who can ship, learn, and iterate fastest.

What this means for you

• Don't over-invest in building the "perfect" app. Build the smallest real version and get it in front of users.
• Treat distribution as the actual product work — because it is now.
• Let real feedback, not your assumptions, decide what to build next. The fastest path to a good app is real users, fast.