Jerry Liu: define an eval and hillclimb—"eval driven development" solves most tasks
hwchase17 · x · 2026-10-09
LlamaIndex founder Jerry Liu argues that almost any task can be solved by defining an eval and hillclimbing over it, rather than hand-crafting deterministic or agentic workflows. Data providers' job is defining evals for all economic activity; yours becomes pointing frontier intelligence at the right problem and defining what good looks like. Agent application interfaces will evolve accordingly—complex processes still need explicit workflow builders, but most tasks compress into goals plus eval instructions. LangChain's Harrison Chase dubbed it "eval driven development."
More from coding & agent
- gemini-cli PR adds canCreateSymlinks() check to skip failing Windows symlink tests — supunyasanthaofficial · 2026-10-09
- Hot-reload Lua on a virtual ESP32 and vibe-code hardware live from Claude — genmon · 2026-10-09
- Agent running tests in a repo copy still hit the live database via env vars — Stunning-Sherbet1853 · 2026-10-09
- Dev Uses Agents to Wireframe UX for DB Architecture Decisions, Cutting a Week-Long Loop to Instant — genmon · 2026-10-09
- Scaling Anthropic's AI-native SDLC playbook into an agentic software factory — Pavan_Belagatti · 2026-10-09
- Satori's next version adds 3D transforms, calc(), min/max/clamp() and more CSS — shuding · 2026-10-09