Anthropic Study: AI Coding Tools Cut Comprehension 17%, No Significant Speed Gain

Anthropic published a randomized controlled trial of 52 junior engineers learning Python's Trio library. The AI-assisted group scored 50% on comprehension vs. 67% for the manual group — a 17-point gap concentrated in debugging. The AI group finished ~2 minutes faster, but the difference wasn't statistically significant. The critical split: developers who used AI for conceptual inquiry scored 65%+, while those who delegated code generation scored below 40%. The how matters more than the whether.