Rather than prompting GPT into implementing a solution, can we prompt it to try to preemptively find issues with the codebase or missing-functionality?
Also, do we know what languages GPT-4 "understands" at a sufficient level? What knowledge does it have of post-2021 language features, like in C23?
Also, do we know what languages GPT-4 "understands" at a sufficient level? What knowledge does it have of post-2021 language features, like in C23?