'Most people run Claude at 30%': the surgical revision prompt that beats 'make it better'
JafarNajafov · x · 2026-10-26
Continuing his Anthropic prompting series, JafarNajafov argues lazy prompts leave Claude at 30% of its potential. Technique 8: instead of "improve this," ask the model to "identify the single weakest sentence and rewrite only that one, then explain why it was weakest" — surgical targets beat generic polishing.
More from coding & agent
- Dev hot take: if you're still using PRs, you're behind — TAbrodi · 2026-10-26
- Blogger ditches Codex for Claude to rebuild personal blog after 'hallucinating' fixes — sujingshen · 2026-10-26
- Cloudflare dashboard adds one-click "Onboard your agent" setup prompt — fekdaoui · 2026-10-26
- Nebula: a shared self-learning memory layer for Claude Code and Codex — BoysenberrySad3641 · 2026-10-26
- Postiz-cheap: a free open-source alternative to the $99/mo social scheduler — matchaman11 · 2026-10-26
- AWS re:Invent 2026 dedicates Security Day to AI: 100+ sessions on securing autonomous agents — DavidLinthicum · 2026-10-26