While I’m still very much worried about AI use for large amounts of code (e.g., beyond the normal line-or-two code-complete tools) for a few reasons, I’m experimenting with using an AI code review bot: https://www.coderabbit.ai/
I’m looking for feedback, but I think this use case seems okay:
- it’s not writing code, it’s adding comments and doing some level of static analysis
- any changes it suggests are marked accordingly and have to be manually approved
- it’s helping with code review because at the moment, it’s a big need (i.e., I’m the main one doing review and I can’t review my own code)
- it suggests areas to consider or test for manual inspection
I’m curious to know other thoughts (e.g., is this a slippery slope or does it seem okay to you too).