7 ms·
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 p
by hcarvalhoalves 2y 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.