>>106327542
I don't even get why he doesn't think it's AI. He seems to think only symbolic 100% correct systems are AI. So by his definition a toddler or a dog don't have intelligence because they can't do math or formally structured logical deductions.
Any system that's heuristic in nature will have an error rate.
Unless he thinks the output of LLMs have literally 0 correlation with the truth and are only stylistically similar to human speech, but that's a super retarded idea to have.