One neuron at a time
by Caleb
Push single neurons in small open models and turn them into villains, flatterers or pirates. Try the recorded replies, then run it yourself.
Articles and design principles for multi-agent systems, with detailed walkthroughs.

Orchestrated AI Team is a public collection of technical articles on multi-agent AI systems. It describes a setup where one orchestrator agent delegates work to specialist agents, each on a model suited to its job.
It holds a manifesto of design principles and a detailed architecture walkthrough. It is for engineers building orchestrated AI applications.
by Caleb
Push single neurons in small open models and turn them into villains, flatterers or pirates. Try the recorded replies, then run it yourself.
by Jeremy
Capability-based security, scope-bounded planning and out-of-band audit for autonomous LLM agents.
by Chris
A guided tour of Dave, a personal AI agent: what it can do, why each feature exists and how it is put together.