Archive

Anthropic repriced the developer subsidy to chase Fortune 10 contracts. Cline open-sourced the exit the same day. Two bets that don't end in the same place.

Frontier models are producing offensive exploits and dissolving startup moats simultaneously. The two bets don't end in the same place.

Anthropic contains cyber capability as liability. OpenAI monetizes it as moat. The CVE in production decides which bet wins.

The capability layer shipped this week. The trust layer didn't. The gap compounds every time another runtime hits mass deployment.

Two parallel bets on AI — leverage versus headcount offset — and the math on which one compounds isn't close.