| ▲ | walthamstow 3 hours ago | |
Using a different lab's model to review the code to the one that wrote it. Claude model plans/codes, OpenAI model reviews, or vice versa, or DeepSeek or GLM or whatever. | ||
| ▲ | a3nz3g a few seconds ago | parent [-] | |
I also did it this way. I think using the same lab's model review the code also helps. I generally start a new Claude Code and a Codex let them to review the code focus on 3 things: 1. bug, 2. cut corner, 3. code style. This generally works for be, I think a solid spec is the key for model to review. And i generally ask them to follow the spec and if spec does not say then follow first priciple. Also if there are user experience change need to notice me. How about your patten? | ||