Loop AI

Full write-up: loopai.com/blog/agi-pilled-o…

Most companies use AI one person at a time. Everyone has their own Claude Code or Cursor, their own prompts, and their own lessons that never leave their laptop. At Loop we started from first principles: once AGI is here, how will a company work? Then we set up our org to work that way today. A few months ago, we moved all our work to web agents in public Slack channels, all sharing one company brain. When one person works with the agent on a part of the system, the agent becomes an expert on it. After that, anyone else can contribute to that part without needing the one person who knew it. Tribal knowledge used to be a bottleneck, and now it builds up in one place. This only works top down. Leaders hold the most context: KPIs, priorities, and the reasons behind our calls. So we had to write ours down first. We wrote up how we set it up, what broke, and what it costs. Link below.