3 ms·
If you ask me to solve increasingly dififcult Tower of Hanoi problems, I don't expect to be good at it. Neither would I expect a fellow human to be. So based on
by salviati 1y ago
If you ask me to solve increasingly dififcult Tower of Hanoi problems, I don't expect to be good at it. Neither would I expect a fellow human to be. So based on this should we question our intelligence?
I heard about that paper through an "AI explained" video [0], so I might be biased, but I agree with that video that the Apple paper is "meh" at best: it points out LLM limitations that are hardly a surprise.
[0] https://www.youtube.com/watch?v=wPBD6wTap7g https://www.youtube.com/watch?v=wPBD6wTap7g
- vincnetas 1y agoProbably the difference between you and AI is that you would acknowledge that it's too difficult for you, and not to bullshit your way through.
- saithound 1y agoThat's _exactly_ what the LLM did: the article's authors decided to count that as a failure.
- vincnetas 1y agoHm was reading only TFA not the research paper. But TFA mentions this : Perhaps the most unsettling finding is what failure looks like. Even when models are completely wrong, they sound persuasive. The reasoning is fluent, the explanations are structured, and the conclusions are confidently delivered. But the logic doesn’t hold.
- rcarmo 1y agoThat sounds a lot like a salesperson. And yes, there is a human tendency to twist reasoning to make the written word look polished, and I don’t think LLM training has fixed that bias.