And many people with comparable backgrounds have similar opinions to him, they just don't talk about it. Yes, his interviews to the general public have been more than a little alarmist. But if you listen to his longer talks, much of what he says is accurate, or at least generally agreed upon in the computer science community. He is taking those points to their logical extreme, and that extreme isn't particularly comforting.

At the moment I don't see any evidence, theoretical or empirical, to refute any of the more technical statements that he has been making (perhaps you can find some?). The same cannot be said for certain prominent members of the neuroscience community.