Linus laughs at Kurzweil

gizmodo.com

15 points·arkj··4 comments

TL;DR; A guy who is not an expert on AI is described as disagreeing with another guy who is not an expert on AI, about AI.

Related reading: predictions about the feasibility of heavier-than-air planes and metal/concrete boats, feasibility of space travel, market size for wireless communication and personal computing etc.

> 2015: We’ll get AI, and it will almost certainly be through something very much like recurrent neural networks.

> ~1700: We'll get aircrafts, and it will almost certainly be through something very much like birds flapping feathered wings.

Sometimes I wish there was a downvote button for the submissions... in this case, not because the linked article is bad (though I am surprised Torvalds apparently dismisses the idea of a self-improving AI), but because the title is nasty and not reflective of the article linked.

My question is: do we know enough about human intelligence to know that we can physically replicate or exceed it? Classically the human mind is seen as a Turing machine based on RNNs, however, one area of debate suggests that the human consciousness/intelligence could have a quantum component.[1] I don't know which hypothesis to agree with because of a lack of evidence either way: and there-in lies the problem.

However I tend to agree more with Linus because, as always, he has the more logical and pragmatic standpoint.

[1]: https://en.wikipedia.org/wiki/Quantum_mind

Reply on news.ycombinator.com