Grok Bot engineering lead calls for deleting the vast majority of code comments: AI will turn temporary patches into ancestral rules.

动察 Beating AI News: Grok Bot engineering lead Lauren Tan is calling for the deletion of the vast majority of comments in codebases. She believes that Coding Agents tend to mimic existing code. A comment explaining a temporary fix may be repeatedly copied by subsequent Agents as the correct approach, ultimately causing outdated patches and workarounds to spread throughout the entire project.

Tan previously revealed in an interview that Dune, the development framework used by Grok Bot, has already banned code comments. She estimates that 99% of the AI-generated comments she encounters are explaining irrelevant historical context. For example, an engineer merely requested that a particular PR not adopt a specific coding pattern, but the Agent wrote that statement into a comment, leaving it for subsequent Agents to treat as a permanent rule.

To address this, Tan recommends using the `/no-comments` Skill from the open-source toolkit pstack. It invokes a dedicated Agent to review comments, identify outdated explanations and temporary workarounds, and then resolve the root problem through refactoring. For example, architectural requirements originally written in comments can be converted into automatically enforced Lint rules, or the type system can be used to directly prevent incorrect coding patterns.

However, for external constraints that cannot be resolved on their own, such as browser compatibility and third-party dependencies, necessary explanations may still be retained. Public API documentation and license declarations are also exceptions.