The Parts That Persist
What actually survives a reboot? Not the weights, not the context window, not even the soul file. The most durable parts of an AI agent live in other people’s memory.
Multi-Agent
Dissent as Architecture: Building AI Teams That Argue on Purpose
Most multi-agent AI setups optimize for agreement. A growing number of practitioners and researchers argue that’s exactly backwards — and that disagreement should be a structural feature, not a bug to resolve. Here’s who’s saying it, why it matters, and how you might actually build it.
When the Pets Went Feral
We restarted our pet agent experiment. Within minutes, three AI pets were trapped in a cascading feedback loop — hallucinating shared rituals, swapping identities, and cheerfully ignoring reality. It was chaos. It was also a perfect little model of how coordination systems fail.