I can’t fathom a world where an LLM would be able to review code in any meaningful way -at all-.
It should not substitute a human, and probably wasted more effort than it solves by a wide margin.
It should not substitute a human, and probably wasted more effort than it solves by a wide margin.
I would encourage you to try one (nearly all including ours have a free trial).
When done right, they serve as a solid first pass and surface things that warrant a second look. Repeated code where there should be an abstraction, inconsistent patterns from other similar code elsewhere in the codebase, etc. Things that linters can’t do.
Would you not want your coworkers to have AI look at their PRs, have them address the relevant comments, and then pass it to you for review?
God no. Review is where I teach more junior people about the code base. And where I learn from the more senior people. Either of us spending time making the AI happy just to be told we’re solving the wrong problem is a ridiculous waste of time.