Growing behavior trees as agent harness: an agent that writes its own reflexes
CatAstro_Piyush · x · 2026-11-19
Sihan Xu publishes a 15-minute essay (with code) arguing that future agents won't sit waiting for user turns — they'll act in the world and keep responding to what happens around them. Two core ideas: (1) behavior trees may be the best harness design, letting routine situations be handled without waking the model; (2) an agent can grow its own tree by learning new reactions from lived tasks, so the reflex tree starts at one node and gets denser over time. Framed via Silver & Sutton's 'Era of Experience' argument that agents need long-running experience streams.
More from coding & agent
- ArkenOS launches Optimus-1, billed as the first swarm harness for autonomous research — CatAstro_Piyush · 2026-11-19
- Claude Code 2.1.328 teardown: prompt files down 19%, system prompt share near 80% — ClaudeCodeLog · 2026-11-19
- Claude Code 2.1.328 ships 144 changes: auto mode for Opus/Sonnet 4.6 skips permission prompts — ClaudeCodeLog · 2026-11-19
- Claude Code 2.1.328 adds Auto permission mode and MCP hook failure blocking — ClaudeCodeLog · 2026-11-19
- T3 Code hits 31.7T tokens a day, burning over $10M in API spend within 24 hours — 0xkarasy · 2026-11-19
- Ruby 4.0.10 released with critical security fixes for Regexp and case conversion — kieranklaassen · 2026-11-19