HackerTube 中文?

Does AI Actually Boost Developer Productivity? (100k Devs Study) - Yegor Denisov-Blanch, Stanford

AI EngineerYouTube ↗

Further reading 1

Will the leaderboard winner solve your problem? →A public score can shortlist models. A new version of your own task can reveal what the score misses.This video helps answer · Which sample produced the score, and what work did it miss?

Caption notes

Denisov-Blanch explains why commit counts make a poor proxy for AI coding gains. Tasks vary in size and fixing newly introduced defects adds commits too, motivating comparisons across task complexity and codebase maturity.

Public machine transcript (may contain transcription errors) ↗
Original passages · 2
delivering more commits does not necessarily mean more productivity
Does AI Actually Boost Developer Productivity? (100k Devs Study) - Yegor Denisov-Blanch, Stanford · ▶ 5:07–5:14
task complexity, code base maturity, language popularity
Does AI Actually Boost Developer Productivity? (100k Devs Study) - Yegor Denisov-Blanch, Stanford · ▶ 17:37–17:44

Hacker News comments

View original on HN ↗