The Code Nobody Reads

Addy Osmani withdraws his advice to read every line of AI-written code and proposes automated review of every change, human review matched to risk, and a person approving every merge.

An engineer two weeks into a job at a large company wrote that their days went on approving AI-written code nobody had read. Addy Osmani, who writes the Elevate newsletter and helps build an AI tool that writes code, withdraws his advice of two years ago to read every line of AI-written code. He sets out what should replace that reading.

• Osmani proposes that AI agents review every pull request first. Small changes to less sensitive code then get a lighter human check, and core and sensitive code gets careful review by its owner.
• In his model a person still approves every merge and decides how closely to read it. He cites Anthropic, the maker of the Claude AI models, where an automated reviewer checks nearly every pull request but approves nothing.
• He argues that tests written by the same AI agent that wrote the code prove little. Checks become trustworthy when they come from an independent source, such as a specification a person wrote.
• He asks engineers to state in the pull request which parts they did not read, and asks leaders to measure incidents and rollbacks instead of merged pull requests.

An engineering team can decide in advance, for each part of its code, how much human review a change needs, instead of leaving it to each reviewer on the day. This is an inference and is not the author's claim; Osmani says he is not neutral, because he helps build an AI tool that writes code.

Derived by Working Surface from the article. Source line: Nearly every change gets automated review, a person still approves what merges, and it’s up to that person to decide how deep to go.

28 September 2026: when agents build, people keep the first screen, the review, the files agents read and the whole product.