Anthropic analyzes 207k research sessions: data analysis dominates, formal sciences use Claude most
plopesresearch · x · 2026-11-19
Anthropic published "AI for scientific discovery: How researchers use Claude," analyzing 207,000 Claude and Claude Code research sessions with privacy-preserving methods:
- Fields: Relative to researcher counts, formal sciences (CS, math, statistics) use Claude most; health and lab-based sciences least. CS alone accounts for 32.7% of Claude Code sessions
- Tasks: Data analysis dominates — about a third of Claude sessions and nearly half of Claude Code sessions — followed by writing. Yet scientists rank idea generation and data gathering as most important, while using AI least for them
- Capability alignment: Tasks in the top 20% of theoretical AI suitability get 2–4x more usage
- Specialized tools help: On Claude Science, idea generation's usage share is roughly double that of regular Claude
More from AGI Musings
- AI Prof David Krueger Launches Evitable to Fight the 'Race to Replace Humanity' — DavidSKrueger · 2026-11-19
- Why the AI era will force society to redefine 'work' itself — Paimaamu · 2026-11-19
- Grady Booch names frontier models' shared pattern: the 'Fallible Chaperone' — Grady_Booch · 2026-11-19
- WSJ charts: youth unemployment climbing worldwide, AI may be squeezing young software devs — zacharynado · 2026-11-19
- Brian Armstrong: teams of specialized open-weight models can beat frontier models — markjeffrey · 2026-11-19
- Normalizing Anthropic's Claude usage data by field size: CS leads by a wide margin — plopesresearch · 2026-11-19