AnalysisDevelopersAI Agents
Jun 16, 7:41 PM
Persistent returns to expertise in agentic coding, Anthropic finds
Anthropic analyzed ~400,000 Claude Code sessions from Oct 2025 to Apr 2026. Success rates across occupations are nearly equal, with domain expertise providing a modest boost. Debugging time halved, and average task value rose ~25%.
Jun 16, 7:41 PM