Skip to content

Posts

  1. An assistant working at my desk

    I use Dot and Orca every day. How an AI assistant with computer use orders groceries, claims refunds and sends daily updates, and what I still wish it did.

  2. Build simple first

    Too many features and you can’t tell what’s good or what makes users stay. Why AI makes bloat easy, why to start with an MVP, and how to earn complexity.

  3. One team, one set of skills

    Most teams get instruction files wrong because nobody shares skills. Why AI agents need a pilot’s checklist, three skills to write first, and why rules decay.

  4. Adapt or die

    AI can do most of the coding, so an experienced engineer’s job is deciding what to build. Why you’re the architect, AI is the contractor, and where to start.

  5. Someone else dug the hole

    Is AI a bubble? Probably. Why we can’t back out of it now, what a pop would do to the economy, and what has to go right so the bubble doesn’t have to pop.