Harvard study reveals AI coding assistants are clogging up code reviews instead of speeding up work.
Original headline: Harvard researchers find AI coding agents shifted the bottleneck to code review — complexdiscovery.com
More in Programming and developer news
A Harvard working paper analyzed data from 718 firms that adopted AI coding agents. While engineers using these tools produced about 30% more code, the actual amount of finished, approved work didn't increase. The researchers found the bottleneck in development had simply shifted downstream: getting code through human review took 49% longer.
The study shows that the promise of AI speeding up software delivery hits a wall at the human checkpoint. For development teams, this means AI's productivity gains might be offset by slower, more cumbersome review cycles. The key takeaway is that measuring AI's success requires looking beyond raw code output to focus on approved work and reviewer efficiency.
Read the original at complexdiscovery.com
Satirical commentary on real tech news, summarised with AI. Read the original source for the facts. How ZeroDaily works
