Urgent.News

What's breaking now, across thousands of outlets.

AI

Harvard Says Most Organizations are Using AI Write More Code but Don't Deliver More. Here's How to Fix It.

We gave developers coding agents but kept the old way of working. Then we were surprised when nothing got faster. AI helps us write more code. So why aren't we shipping more software? A new paper from Harvard tries to answer that question. Fiona Chen and James Stratton looked at data from 655 companies and more than 720,000 workers. They found that AI tools help people write a lot more code. But…

A new Harvard study examines how organizations are utilizing AI to write more code, but struggle to deliver more software. Researchers Fiona Chen and James Stratton analyzed data from over 655 companies and more than 720,000 workers. While AI tools helped developers write significantly more code, the amount of finished software remained nearly unchanged.

The reason behind this phenomenon lies in the code review process. After companies adopt AI agents, code waits longer for review, takes longer to be reviewed, and reviewers request more changes. In essence, speeding up one step of the process creates a bottleneck at the next step.

The study has garnered significant attention, with some interpreting it as evidence that AI is overhyped. However, the authors argue that the paper actually sheds light on where bottlenecks occur within organizations, rather than the limitations of AI itself. One explanation for this bottleneck is the lack of adaptation in the way companies work. Most teams follow traditional practices such as assigning work via tickets, planning in two-week sprints, and reviewing every change, regardless of its complexity.

To address this issue, the authors propose adopting a more flexible approach to work management. They suggest that companies should revise their processes to accommodate the increased speed of AI-generated code, rather than simply adding more friction in the form of additional review queues and approvals. By making small adjustments to existing workflows, organizations can harness the benefits of AI-generated code while minimizing the negative impact on software delivery.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

More from Friday 20 November →