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.
A short interactive field guide to reliable AI assistance and the agent harness around a model.

The Harness Is the Asset is an interactive web presentation about what makes AI assistance dependable. Its main idea is that the system around a model, the harness, matters as much as the model itself.
It walks through six scenes that take three to five minutes to read, each with optional explainers. It started as a talk given at a machine learning meet-up in Perth and is aimed at people who want AI to do repeatable, checked work.
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.