https://video.ut0pia.org/videos/watch/cd640abd-4d20-4f50-9ca2-dd2782f59d26 The abstraction is three things: state, a synchronous reducer that derives state from events, and an after-append hook for side effects. The split matters: when your program restarts after 100 events, you want to catch up state without replaying LLM requests. Everything that happens (streaming chunks, tool calls, errors, circuit breaker triggers) is an event in the log. The interesting part is deployment. Jonas demos "dynamic worker configured," an event whose payload is a JavaScript string containing a processor. Append it to any stream and that stream becomes an AI agent without server or dependencies. The broader implication: processors from different authors on different servers can compose against the same stream, and a safety checker can inject context in a 200ms window before an LLM request without blocking the agent if it doesn't make it. Speaker info: https://x.com/jonas, https://www.linkedin.com/in/jonashuckestein, https://github.com/jonastemplestein Fri, 15 May 2026 19:02:00 GMT https://validator.w3.org/feed/docs/rss2.html PeerTube - https://video.ut0pia.org https://video.ut0pia.org/lazy-static/avatars/0287a09a-aae7-4840-9843-b416426e7046.webp https://video.ut0pia.org/videos/watch/cd640abd-4d20-4f50-9ca2-dd2782f59d26 All rights reserved, unless otherwise specified in the terms specified at https://video.ut0pia.org/about and potential licenses granted by each content's rightholder.