|
|
|
|
|
|
by hcarvalhoalves
760 days ago
|
|
|
I’m not sure the Halting Problem is a compelling argument theoretical reason why AI can’t program. I do see a compelling argument on the limits of LLMs being probabilistic (rather than logic/symbolic), as I’ve seen on GPT/Pilot coming up with code that looks syntactically valid but calls completely made-up APIs. |
|
|