[Eisenstein]

So there is a theoretical framework which can be tested against to achieve AGI and according to that framework it is either not possible or extremely unlikely because of physical laws?

Can you share that? It sounds groundbreaking!

[Terr_]

The people who claim we'll have sentient AI soon are the ones making the extraordinary claims. Let them furnish the extraordinary evidence.

[Eisenstein]

So, I think people in this thread, including me, have been talking past each other a bit. I do not claim that sentient AI will emerge. I am arguing that the person who is saying that it can't happen for a specific reason is not considering that the reason they are stating implicitly is that nothing can be greater than the sum of its parts.

Describing how an LLM operates and how it was trained does not preclude the LLM from ever being intelligent, and it almost certainly will not become intelligent, but you cannot say that it didn't for the reasons the person I am arguing with is saying, which is that intelligence can not come from something that works statistically on a large corpus of data written by people.

A thing can be more than the sum of its parts. You can take the English alphabet, which is 26 letters, and arrange those letters along with some punctuation to make an original novel. If you don't agree that means that you can get something greater than what defines it components, then you would have to agree that there are no original novels because they are composed of letters which were already defined.

So in that way, the model is not unable to think because it is composed of thoughts already written. That is not the limiting factor.