willcb: in-context learning for personal continual learning needs neurosymbolic harnesses
willcb · x · 2026-09-19
Continuing his thread on personal continual learning, willcb argues that in-context learning works for real-time personal continual learning if and only if the community leans into unconventional neurosymbolic program harnesses — building on his earlier point that today's best approach combines harness engineering, long-horizon RL for memory and multi-agents, and online RL over task flywheels built from live interaction traces.
More from coding & agent
- Mnemos engine builds synaptic-like agent memory via resonant engrams — RileyRalmuto · 2026-09-19
- Codex's built-in browser now supports extensions and cookie import — vista8 · 2026-09-19
- Jev's structured-output model impresses, Yegge argues fences beat sandboxes for agents — njyx · 2026-09-19
- Archify, a 67k-star open-source skill, turns one-line system descriptions into interactive diagrams — alex_verem · 2026-09-19
- Generating a 40s motion graphic video with GLM writing code, no video model — 9r4n4y · 2026-09-19
- UnCorreoTemporal MCP server lets AI agents grab OTP codes and verification links with zero human help — modelcontextprotocol · 2026-09-19