and if that is indeed the point: that the difference is simply size, then what does that law look like? surely it would need to specify a size of the relevant neural network that is able to derive works
but that’s then just an arbitrary number because we just don’t know what it would be