How Cognition Uses Devin to Build Devin

The Cognition team on running Devin against their own codebase, reached through whichever surface someone is already in: the web app, Slack, Linear or Jira tickets, the CLI and the API. The work they point it at is the high-volume iterative kind. Bug fixes and edge cases, test coverage, CI failures, lint errors and CVE remediation, PR review, documentation upkeep, codebase questions, and audits that run daily to catch design system violations from the previous 24 hours. Ask Devin gets used constantly to scope a piece of work before a session starts.

The number they lead with is 659 Devin PRs merged into their own codebase in the week before publication, against 154 in their best week of 2025. The claim that goes furthest is about who gets to contribute: anyone is able to contribute regardless of their technical expertise or role in the company, they don’t need to understand and set up Git or any command line tools.

Read it with the obvious caveat. This is Cognition describing how well Cognition’s product works on Cognition’s code, the numbers are self-reported, and merged PR counts measure throughput rather than value. The idea that survives the discount is the recursive one: a vendor whose fastest route to a better agent is pointing the current agent at itself.