Of course, I have no original thoughts. Something is not created out of nothing. That would be Astrology, perhaps :).
But the difference between a human and an LLM is that humans can go out in the world and test their hypothesis. Literally every second is an interaction with a feedback loop. Even typing this response to you right now. LLMs currently have to wait for the next 6-month retraining cycle. I am not saying that AGI cannot be created. In theory it can be but we are definitely milking the crap out of a local maximum we've currently found which is definitely not the final answer.
PS Also, when I said "it can spot mistakes," I probably gave it too much credit. In one case, it presented several potential issues, and I happened to notice that one of them was a problem. In many cases, the LLM suggests issues that are either hypothetical or nonexistent.
But the difference between a human and an LLM is that humans can go out in the world and test their hypothesis. Literally every second is an interaction with a feedback loop. Even typing this response to you right now. LLMs currently have to wait for the next 6-month retraining cycle. I am not saying that AGI cannot be created. In theory it can be but we are definitely milking the crap out of a local maximum we've currently found which is definitely not the final answer.
PS Also, when I said "it can spot mistakes," I probably gave it too much credit. In one case, it presented several potential issues, and I happened to notice that one of them was a problem. In many cases, the LLM suggests issues that are either hypothetical or nonexistent.