Nous Research Breaks Down How Hermes Agent Learns Skills and Improves Over Time
Teknium · x · 2026-11-16
A talk from Nous Research (shared by Agentic AI Foundation) explains how Hermes Agent learns from experience: it builds reusable skills, retains context across sessions, and improves over time instead of starting from scratch each session — a concrete take on agent memory and skill accumulation.
More from coding & agent
- Open-source inbox app turns GitHub PR review into one AI-pre-reviewed queue via MCP — zeeg · 2026-11-16
- Vercel ships UI SDK for fully custom AI-generated UI components, early users impressed — shadcn · 2026-11-16
- Google offers free Antigravity Ultra access to physical and natural science researchers — DynamicWebPaige · 2026-11-16
- Google doubles context window for Antigravity's Argon, Carbon and Hydrogen, cutting compaction 3.5x — sundarpichai · 2026-11-16
- P2P agent pattern: deterministic workflow outside, swappable agentic core inside on LangGraph — hwchase17 · 2026-11-16
- Retell AI MCP Server lets you build and manage AI phone agents via natural language — modelcontextprotocol · 2026-11-16