Video by Deep Mind Founder Demis Hassabis

I was amazing to see an ostensibly "really, really smart guy" peddling this kind of shallow, self-contradictory arguments.

Similarly, he later states his real approach is a hybrid of engineering and biological simulation. But if the code for intelligence is so sparse, that would say that this approach wouldn't work - you'd need to extract the one, true intelligence form from DNA or something.

And just as much, we don't even know enough to be able to say that human intelligence rather than merely mammal intelligence is what we need to have. And humorously enough, he presents an algorithm derived from a mouse-brain, again contradicting his "human intelligence is sparse 'cause it only happened once" argument.

And I suspect we could on and on. I recall a lot of neuroscientists are critical of any contemporary "this is how the brain works" pronouncements.