The 20-60-20 rule for AI coding: without the human 40%, you just ship slop
_jaydeepkarale · x · 2026-09-03
Responding to a question about what skill matters most when AI agents do 70–80% of development, the author proposes the 20-60-20 rule: 20% of effort writing context, 60% letting the LLM work, 20% reviewing its output. The 60% only pays off if the human does the other 40% well—otherwise the result is slop and tech debt.
More from coding & agent
- Cohere Labs releases ATE dataset of ~700K tools from public MCP servers — Cohere_Labs · 2026-09-03
- Cursor's harness beats VS Code, but hooking it to local Ollama models needs workarounds — walkingriver · 2026-09-03
- Baseten ships remote MCP server and skill, cutting coding agents' token usage by up to 57% — baseten · 2026-09-03
- davidad: rollout rubrics in verifiable domains should check whether the model ran the verifier — davidad · 2026-09-03
- Dev's Claude Code sales plugin sent 1,097 emails in 85 days: 0 paid — No_Job_9995 · 2026-09-03
- Ripple Edit: a vibe-coded ComfyUI extension that moves node groups to free up space — GeroldMeisinger · 2026-09-03