The error in your thinking is that you assume the essence of human cognition cannot ever be reducible to a algorithmic process that our current transformer mode

The error in your thinking is that you assume the essence of human cognition cannot ever be reducible to a algorithmic process that our current transformer models are approximating. Which may be the case... but we don't know for certain yet, so your certainty of the negative is also not warranted. I can say the same your fears of machine intelligence is clouding your ability to objectively assess evidence. You can design a novel problem and see for yourself the reasoning and logical deductions an LLM will make to solve it, like many have already done. > If you asked GPT directly if it were intelligent or sentient it would not agree with you either If you think this class of questions is appropriate to gauge reasoning ability, I don't know what to tell you.