3 ms·
That's what you might expect, but it requires another breakthrough IMO. From what I understand of LLMs, they can't be incrementally improved to add logical reas
by nprateem 4y ago
That's what you might expect, but it requires another breakthrough IMO. From what I understand of LLMs, they can't be incrementally improved to add logical reasoning since they're just guessing the next word. It's impressive and fine in many cases, but there are many where it's not enough.
- HDThoreaun 4y agoThey already have some form of reasoning though. GPT4 can solve novel problems. It's certainly not incredible at it, but the incremental logical reasoning improvements have already begun.
- antibasilisk 4y agoThey do logically reason, that's the whole point.
- nprateem 4y agoThey claim to. But there's no inherent understanding. Otherwise they could do maths and reason correctly about code. It's just probabilities.
- antibasilisk 4y agoBut they do often seem to reason correctly about code.