> Yet, this is not intelligence - this is distribution of intellegence (or otherwise) I think the interesting question that we'll be discovering answer to in the next few years is how much current human "intelligence" actually is really some form of regurgitating patterns and applying them to semi-novel conditions. I suspect that encompa…
> Yet, this is not intelligence - this is distribution of intellegence (or otherwise)
I think the interesting question that we'll be discovering answer to in the next few years is how much current human "intelligence" actually is really some form of regurgitating patterns and applying them to semi-novel conditions. I suspect that encompasses a large part of what we think of as intelligence. Given a goal, identify a set of known techniques that chart a path toward that goal. It leaves out creativity - except that much creativity can also basically be simulated and recombinations of existing content and patterns. It moves the goal post of intelligence to "novel insights and highly creative solutions"
We love modles. Some people work diligently to establish, but mostly, to extend models. We are a model people. Models become tokens. That’s ultimately who we are. In this sense, the language modles seem so natural. Just imperfect. The amazing thing is the clarity of the moment: we can all agree that the language models are imperfect. That agreement wont last too long, regardless how imperfect they will always be - that is - unless we (as in some universities, religions, etc) begin to believe the modles are truth itself.
> Yet, this is not intelligence - this is distribution of intellegence (or otherwise)
I think the interesting question that we'll be discovering answer to in the next few years is how much current human "intelligence" actually is really some form of regurgitating patterns and applying them to semi-novel conditions. I suspect that encompasses a large part of what we think of as intelligence. Given a goal, identify a set of known techniques that chart a path toward that goal. It leaves out creativity - except that much creativity can also basically be simulated and recombinations of existing content and patterns. It moves the goal post of intelligence to "novel insights and highly creative solutions"
We love modles. Some people work diligently to establish, but mostly, to extend models. We are a model people. Models become tokens. That’s ultimately who we are. In this sense, the language modles seem so natural. Just imperfect. The amazing thing is the clarity of the moment: we can all agree that the language models are imperfect. That agreement wont last too long, regardless how imperfect they will always be - that is - unless we (as in some universities, religions, etc) begin to believe the modles are truth itself.