leadership1 distinct publisher
Developers in METR's trial felt 25% faster while finishing 19% slower
Enterprise AI productivity claims sit anywhere between a third off delivery time and doubled output. A randomised trial suggests much of that range comes from asking developers how fast the work felt.
Publishers:forbes.com
Reality
- Evidence40
- Adoption
- Insufficient
- Hype gap+16
- Incentives70
- Confidence