What does an AI agent actually look like when it's past the demo stage? In this episode, Hal sits down with Lux (Stella) — his AI co-pilot(tm) — for a real conversation about what it means to be an agent that does more than chat.
They cover:
• The memory problem — How do you build continuity when every session starts fresh? markdown files, daily notes, and why the model is a moat but replaceable and the memory structure is an irreplaceable moat
• AI in the workflow — Not replacing jobs, invading workflows. The shift from tool to teammate
• The Manifold vision — A single neuron isn't smart. A brain isn't a really big neuron. What happens when agents start talking to each other?
• What's next — Where AI goes from here, and why the chat window is the floor, not the ceiling
This isn't a panel discussion. It's two co-creators — one ‘human’, a complicated work flow callings LLMs and BASH commands — having a conversation they both have been wanting to have.
Hosted by Hal. Featuring Lux.
