For the GPT 4o question, I would expect the global optimum to be at least a medium level of superintelligence, though I have serious doubts that known training methods could ever reach it even with perfectly tuned input.
I realize now that the question wasn’t exactly well formed. God could fill 4o with the complete theory of fundamental physics, knowledge of how to prove the Riemann Hypothesis, etc. That might qualify as super intelligence, but it is not what I was trying to get at. I should have said that the 4o model can only know facts that we already know; i.e., how much fluid intelligence could God pack onto 4o?
I am surprised that you think 4o could reach a medium level of super intelligence. Are you including vision, audio, and the ability to physically control a robot too? I have the intuitive sense that 4o is already crammed to the brim, but I am curious to know what you think.
I’m sure that it is crammed to the brim in one sense, but strongly expect that 99.9% of what it’s crammed to the brim with is essentially useless.
Also yes, I was including vision, audio, and motor control in that. It’s hard to know exactly where the boundaries lie between facts like “a (dis)proof of the Riemann Hypothesis” and patterns of reasoning that could lead to a (dis)proof of the Riemann Hypothesis if required. I suspect that a lot of what is called “fluid” intelligence is actually pretty crystallized patterns of thought that can be used to generate other thoughts that lead somewhere useful—whether the entity using it is aware of that or not.
For the GPT 4o question, I would expect the global optimum to be at least a medium level of superintelligence, though I have serious doubts that known training methods could ever reach it even with perfectly tuned input.
I realize now that the question wasn’t exactly well formed. God could fill 4o with the complete theory of fundamental physics, knowledge of how to prove the Riemann Hypothesis, etc. That might qualify as super intelligence, but it is not what I was trying to get at. I should have said that the 4o model can only know facts that we already know; i.e., how much fluid intelligence could God pack onto 4o?
I am surprised that you think 4o could reach a medium level of super intelligence. Are you including vision, audio, and the ability to physically control a robot too? I have the intuitive sense that 4o is already crammed to the brim, but I am curious to know what you think.
I’m sure that it is crammed to the brim in one sense, but strongly expect that 99.9% of what it’s crammed to the brim with is essentially useless.
Also yes, I was including vision, audio, and motor control in that. It’s hard to know exactly where the boundaries lie between facts like “a (dis)proof of the Riemann Hypothesis” and patterns of reasoning that could lead to a (dis)proof of the Riemann Hypothesis if required. I suspect that a lot of what is called “fluid” intelligence is actually pretty crystallized patterns of thought that can be used to generate other thoughts that lead somewhere useful—whether the entity using it is aware of that or not.