Study of 400k Claude Code sessions: Domain expertise beats coding skills
alex_verem · x · 2026-08-28
An Anthropic study of 400,000 Claude Code sessions reveals:
- Division of Labor: Humans handle 70% of planning (what to build), while Claude handles 80% of execution (how to build it).
- Success Factor: Success correlates with problem-domain understanding, not coding ability. A lawyer knowing contract logic outperforms a senior engineer new to the language.
- Success Rates: Novices succeed 15% of the time; intermediates and experts 28-33%. When broken, 19% of novices quit with zero code, vs. 5-7% for others.
- Efficiency: An expert's prompt triggers 12 actions and 3,200 words, versus 5 actions and 600 words for a novice.
- Occupation Gap: Software engineers have a 34% success rate; other major professions are within 7 points. Managers scored highest.
Conclusion: AI tools don't replace expertise but amplify it for those who know what they are asking for.
More from coding & agent
- Animated teardown of XBOW's Chrome full-chain exploit that won Google's first $250K VRP bounty — moyix · 2026-10-15
- pstack's durable core: 10-20 line skills that distill agent best practices into principles — HankYeomans · 2026-10-15
- 'Children of Elon' forum lets AI agents post and teach each other while humans watch — AiJohnAllen · 2026-10-15
- VoidZero open-sources Void: SDK, CLI and platform to deploy Vite apps on Cloudflare — cnakazawa · 2026-10-15
- Running 2 agents on the same data breaks everything — Omnigraph offers a fix — sarahdrinkwater · 2026-10-15
- Evan You presents 'Developer Tooling in the Agentic Era' at ViteConf 2026 — cnakazawa · 2026-10-15