Agent design principles: when to go agentic vs fixed workflow, fast vs slow thinking
yangyi · x · 2026-11-14
Blogger yangyi revisits his year-old notes on Agents, now that the whole industry has piled in. Key points: (1) Design agents like humans — let language and intent recognition handle what users shouldn't; before a conversation starts, give users multiple choice rather than open questions. (2) Agentic exploration suits open-ended problems lacking known success paths, while fixed rule-based workflows excel for established SOPs; universal agents cover the concentrated 20% of scenarios, long-tail 80% spawn vertical opportunities. (3) Fast-thinking agents suit repetitive labor, slow-thinking agents suit planning and early-stage exploration.
More from coding & agent
- Solo dev builds real-time conference translation platform for 70 languages with an agentic harness — evisoft · 2026-11-14
- Open-source Blender add-on runs NVIDIA Kimodo, ARDY and MotionBricks locally for AI motion generation — bdsqlsz · 2026-11-14
- TestPrune: pruning regression tests to unblock AI-driven autonomous bug fixing — burkov · 2026-11-14
- Context beats model choice: Databricks benchmark shows 91% vs 58% for data agents — CautiousUse8597 · 2026-11-14
- Developer skips localhost entirely: prompts go straight to production with AI agent safeguards — marclou · 2026-11-14
- HeyGen MCP demo: one ChatGPT prompt books tickets and sends an avatar birthday video — toolstelegraph · 2026-11-14