So much fun talking to @vtahowe about building long-running agents, managing…

January 30, 2026

So much fun talking to @vtahowe about building long-running agents, managing state and context, and the (very large) new surface area of AI security and user interface things we all need to start thinking hard about. https://t.co/wtpoS9QHAL

Insecure Agents Podcast@insecureagents

Ep. 22 with @kwindla is LIVE

We discussed all the hard agent engineering problems and what separates a POC agent from an enterprise agent

> security
> latency
> non-blocking context compaction
> agent memory
> stateful agents