AI-Generated Code Speed Could Outpace Teams' Ability to Review and Recover
Engineer Evan X. Merz, who says he was laid off from Meta in May 2026, argues that AI coding tools can make software products less reliable. He says that in his final months he generated most of his code with AI, yet his output did not rise dramatically because he prioritised quality while colleagues merged fragile changes faster. He identifies recurring problems, including shallow tests, overly complex code, weaker developer understanding, and pull requests arriving faster than they can be reviewed. The main concern is that teams may accumulate codebases their own engineers do not fully grasp, complicating incident response. The article notes these are his observations rather than company-wide data, and that UK regulators currently lack AI-specific rules, so organisations remain accountable under existing obligations such as data protection law.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in