Raw LLM Responses

​​@noctis-mbrOR, as an alternative, you could look at other more credible researchers than Hinton or just use a little basic sense. Hinton's reasoning is highly motivated, but his claims are patently absurd. There are numerous real research papers from Oxford, Apple, and other AI researchers that explicitly debunk everything he says about LLM reasoning capabilities, they are there for you to Google. "The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity." "Chain-of-Thought Is Not Explainability." "Is Chain-of-Thought Reasoning of LLMs a Mirage?" LLMs do a pretty good imitation of human reasoning with problems that are pretty simple and that they've basically seen in their training datasets, not so good if you make the problem complicated even if the reasoning steps remain the same. They're also easy to confuse if you just add some irrelevant detail to the question. Plus his tendency to reduce all of human experience to equivalent to all of our produced textual data without realizing that we have nerve endings all along our skin, and eyes, and ears, a sense of balance, a latent unconscious awareness of the positions of parts of our bodies relative to other parts. All of this sensory input is both constant and processed by our brains in real time. No LLM is capable of processing the amount of raw data our brains do in real time. That's just not a thing, and it's completely obvious that that's not a thing if you just sit down and think about it for just a few seconds.