Topics›Autonomous Agents with Reliable Outcomes›Autonomous Bug-Fix Loops at Scale: Linear's 300+ Fixes, Anthropic's 388 PRs
★HighAct: MediumFull content
Autonomous Bug-Fix Loops at Scale: Linear's 300+ Fixes, Anthropic's 388 PRs
The Pragmatic Engineer·September 3, 2026
New
Summary
Linear's Datadog/Sentry-triggered bug-autofix loop fixed 300+ bugs in 30 days; Anthropic's 'Claude maintains apps' Slack-based routine opened 388 PRs across repos with 180 merged (45% merge rate) via crash fuzzing, dedup, and dead-code removal — both still human-gated before merge.
Why this signal
Concrete, at-scale evidence for what 'reliable autonomous agent outcomes' looks like in production today, including the merge-rate ceiling that comes with keeping humans in the approval loop.
Original email
The Pragmatic Engineer · 9/3/2026
The Pulse: tech companies move to open AI models