@SteveClough@metalhead.club My opinions, for what little they're worth: As you've implied by writing two separate questions, intelligence is distinct from sentience. Intelligence can be measured by visible performance. It can be limited to certain areas or tasks, just as it is in humans. AGI (if that turns out to be possible) is not a binary thing: an AI could be more intelligent than humans in some areas and less intelligent in others. Sentience — or consciousness — is harder to assess. I can't even prove that a dog is conscious, although I believe it is. A practical definition of consciousness is the ability to suffer. Can a dog suffer? I believe so. Can an amoeba? I think not. But I don't know where the boundary lies. There's no such thing as a soul, and therefore no machine has one. Behaviours often ascribed to a soul emerge from human biology and society. Bonus fourth opinion: I don't see why consciousness couldn't arise in silicon; I don't see any reason why it has to be carbon-based. But I don't think LLMs will ever get us to AGI or to conscious machines.