Laurie Voss argues that AI coding agents have collapsed the cost of producing plausible code while leaving human review as the bottleneck, citing research on GitHub developers and METR tests where open-source maintainers said they would reject about half of agent-generated pull requests that passed automated benchmark checks.
When code costs nothing to produce, how do you review it all?
Laurie Voss argues that AI coding agents have collapsed the cost of producing plausible code while leaving human review as the bottleneck, citing research on GitHub developers and METR tests where open-source maintainers said they would reject about half of agent-generated pull requests that passed automated benchmark checks.
Source: Linkedin