The future of intelligence – Demis Hassabis [video]
youtube.comI see a fallacy in their arguments, even at that level of understanding: you can't be sure you are measuring all the variables when comparing/diffing "both brains". I am not tired of mentioning the book "Biophysics of Computation: Information Processing in Single Neurons"[1]. There is a parallel HN thread about fMRI issues that adds more food for thought [2].
Not saying than other topics he is talking about are not fascinating, just saying that making arguments about AGI is a stretch. Super intelligent people such as Newton also worked on alchemy, and he thought it was possible at that time.
It is also odd, even for basic computer scientists, when he talks about Turing machines limitations and putting too much faith in their thinking capabilities. It's impossible to not mention the halting problem there. Happy he mention later the possibility of quantum effects in the brain though.