OP, so why think that the LLMs like Chat GPT etc. exhibit actual legitimate intelligence?

Consider the following, if you will. People claiming that GPT is exhibiting intelligence tend to assume that passing the Turing Test is the mark of the mental. But hasn't Searle already dealt with this? What do LLMs do that Searle has not already argued against?

GPT at least, seems bound to established patterns formed in pre-existing conversations. Because it uses language in a statistical way which is unlike us. We use language in a grammatical/semantic way.

I think an element of intelligence is differentiation between the communicative act and the actual intention - langue/parole. An AI cannot differentiate itself from the environment so it cannot have non-linguistic intentions.

They might manipulate words efficiently based on statistical prediction, but the classical question is...do they know what they refer to?

I read it years ago, it was so dense and rich I said I would need another re-read at a later time. Going to follow this thread. I'm particularly interested in how it intersects with phenomenology and defeating the reality/appearance dichotomy, and in turn the dichotomy between ontology and epistemology.

FYI, I think Frederick Wilhelmsen's work would be a good primer to get up to speed with the kind of Existentialist metaphysics here. Getting the Essence-Existence distinction down would be necessary before even attempting this.

But if qualia constitute the self, I see two issues immediately for the view:

1. I can differentiate between the sensation of blue I am seeing, and my awareness of that sensation of blue I am currently seeing. I can differentiate myself as the seer from the sensation of blue itself.

2. I don't think it can differentiate between the faculties of sensation and intellection. There's more to consciousness and cognition than just qualia. I can entertain universals, abstractions, conceptualizations, mathematics, symbols, possibilities etc. If the self is qualia, how come the higher order faculties of intellection? Does this not raise issues about how to differentiate between humans and animals, if selfhood is identical to qualia?


I'm sympathetic to the line of thinking that does attempt to unravel the hard problem, however, rather than simply solve it. I think something has gone wrong in the way the question has been posed, and I think a view of consciousness not based on qualia would be better.