The concept of stacked pull requests allows developers to break changes into smaller, dependent updates, making code reviews more manageable and improving development speed significantly.

90% of the code for some products is now written by AI, leading to unexpected bottlenecks in other areas like the merge queue, which had to be completely rearchitected due to the increase in pull requests.

Our engineering team of 15 uses about five Devins each to help them code and move faster, with Devins merging several hundred pull requests into production every month.

Using deep research integration with GitHub, I can analyze pull requests, identify potential improvements, spot bugs, and suggest fixes in minutes.

Developers are experiencing a different reality with AI tools like GitHub Copilot, which significantly boosts their productivity.

A team of 'Devins' is already producing 25% of Cognition's pull requests, and they are on track to hit 50% by the end of the year.

Sweep automatically creates pull requests against your code base based on the issues you file, and it's crucial to clearly communicate this exciting functionality on the website to avoid losing potential users.

Sweep's messaging suggests it helps developers focus on shipping code by managing technical debt, allowing more time for actual coding.

Currently, about a quarter of our pull requests are committed by Devins, and we expect this to exceed 50% by the end of the year.