Linear: by October, 53% of PRs were reviewed by an agent and the author alone
jiayuan_jy · x · 2026-11-17
Linear's engineering post "Code review at Linear" explains how they rebuilt review for the agent era, optimizing for "judicious use of human attention." In January, almost every PR had a second human reviewer; by October, 53% were reviewed by an agent plus the author alone. The old two-human rule would absorb all the speed agents bring without necessarily improving quality, while relaxing review meant shipping with less confidence. Quoting author jiayuan: at Multica, 100% of code is written and reviewed by agents. Linear admits these practices may prove brittle as agents improve.
More from coding & agent
- Notable AI builder says workflow has shifted to cloud agents with unlimited Claude tokens — jiayuan_jy · 2026-11-17
- Jane Street's yminsky: agents are bad at software design — but does design still matter? — charliermarsh · 2026-11-17
- Five-stage agent routine that stopped an LLM deleting my repo — with the Brief prompt to steal — Iiifebri · 2026-11-17
- PlugIt turns any OpenAPI spec into a curated, read-only-by-default MCP server in a Go binary — Low_Strawberry2484 · 2026-11-17
- Testing a Langfuse release gate: one missing exported score correctly blocks the passing receipt — MongooseWonderful198 · 2026-11-17
- Amp vs Claude Code hands-on: user finds Amp gets more done even as $20-tier Fable 5.2 access ends — bytebot · 2026-11-17