"If we redefine intelligence to include things generative models can do, then they are intelligent." Seems like a good argument, but it is not.
The other argument made, "Planes are like birds, in that they can fly, without using the same mechanism. Generative models are more like human intelligence, but with a different mechanism.", is indeed sound. If generative models were more like aeroplanes and not paper planes.
The rest of the essay is placating fears that arise as if these arguments were true in substack-core language, but the premise fails to establish that these models are "intelligent".
The other argument made, "Planes are like birds, in that they can fly, without using the same mechanism. Generative models are more like human intelligence, but with a different mechanism.", is indeed sound. If generative models were more like aeroplanes and not paper planes.
The rest of the essay is placating fears that arise as if these arguments were true in substack-core language, but the premise fails to establish that these models are "intelligent".
reply