If your org has engineers LLM dumping into your MRs/PRs, make it policy that code reviews require a live-review session where the engineer who wrote the code walks the other engineers through the code. This will rapidly fix your problem as the 'author' of the code who cannot explain his/her code will not want to be grilled this way ever again.
Accountability is the real answer. If you don't enable individual and team accountability, then you are part of the problem.
This is fantastic. What a terrific combination of the creative presentation and the clear exposition of information. You've hit on a very nice aesthetic and a stunningly clear articulation of the underlying mechanics.
The problem this solves: The ChatGPT desktop app can't normally "see" Xcode's compiler warnings or errors when helping you debug your project.
The solution:
1. Put the posted run script in your Xcode target's build phase.
2. Point an Xcode editor to the new BuildIssues.md in your project.
3. Point ChatGPT desktop to Xcode context.
Now you'll be able to direct ChatGPT to support fixing the issues at will and it sees the errors/warnings
Accountability is the real answer. If you don't enable individual and team accountability, then you are part of the problem.