Bret Taylor

Tokens are a terrible unit for pricing AI. It's like measuring engineer productivity by lines of code written. An Apple engineer famously submitted a negative number after a big refactoring — his way of saying the metric was idiotic. Outcomes are what matter.

Waking up every morning asking 'What is the most impactful thing I can do today?' is an interesting question. Being able to answer it accurately is the hard part.

Every founder naturally defaults to their superpower as the solution to every problem. Engineer? It's an engineering fix. Designer? Time for a redesign. BD person? A partnership will save us. If your solution matches your background, it's at least 30% likely you chose it for comfort, not truth.

FriendFeed had a better product than Twitter, was more reliable, and innovated faster. We totally lost — for no reason related to product at all. Biz Stone was getting celebrities onto Twitter while we were polishing features. Distribution beat product.

When you get advice, don't just ask what to do — ask why. People extrapolate from very few experiences and present personal anecdotes as universal rules. Ask why three times and you can build a first-principles framework instead of just following a rule.

Agent is the new app. Software is going from helping an individual be slightly more productive to actually accomplishing a job autonomously. That's a fundamentally different value proposition — and it's measurable.