Every note, filterable — by growth stage or language.
Cloudflare Wallets can link an agent payment to an account and constrain how it spends. A separate question remains: was an in-bounds purchase actually wanted?
"AI-native game" has one honest definition: pull the model out and see if play survives. Passing that test says nothing about whether the result has any taste.
I said taste was judgment renamed. Trying to pull the two words apart, I found I could — and the difference is what the moat argument keeps hiding.
Everyone repeats that taste is the moat in the AI era, but almost nobody says what taste is in a given job. It's domain judgment, renamed.
When several coding agents work at once, the hard part is deciding what each one should change, and in what order.
Coding agents need more than context; they need codebases that make the right place to work clear.
When agents make patches cheap, review has to move from accepting code to absorbing knowledge.
AI agents make individual output cheaper before they make organizations better at deciding what should exist.
The risky coding agent isn't the one that ignores you. It's the one that helps too much.
Most enterprise AI value is lost not to model limits but to the organisation around the model. Coordination, not intelligence, is the binding constraint.
AI breaks the day-rate convention. The question is no longer whether consultants use AI but whether firms can stop selling effort and start underwriting results.
AI is unbundling what consulting really sold. The pyramid is becoming a barbell — generic analysis squeezed, judgment-heavy work repriced upward.
No notes at this stage yet.