3 ms·

So wait, doesn't this also indirectly imply that we will never be able to decipher neural networks of sufficient size for the general case?