Never go full AI and hand off a diff you haven't reviewed
retained text
Every few months someone will join my team and get really into such and such harness or whatever and start outsourcing their brain to AI and think they're being a 100x engineer, but don't realize how shit their output is because they're not even looking at what their agents are vomiting out. It's embarrassing and not viewed well by others. Don't be this guy. [PracHub](https://prachub.com/?utm_source=reddit&utm_campaign=andy) has a cache-code review exercise for practising how to check generated changes against invariants and tests. Some tips: * if you're using AI to send Slack messages you've gone too far * everyone already knows AI exists. If you think AI is giving you some productivity hack that nobody else has, you're probably missing something * if you're creating a tool that "lets your agents coordinate with each other" you've gone too far * do not make other people review a diff that you haven't reviewed yourself
document fetched 2026-10-30 00:28 · canonical https://reddit.com/r/OpenAI/comments/1xijfh5/never_go_full_ai_and_hand_off_a_diff_you_havent