Hacker News Reader

Comment by sebringj | Hacker News Reader

This is short-term pessimism, inline with the common dystopian future of sci-fi. It is really difficult to predict human behaviors and technology in the long term.

Humans have the ability to steer things to their advantage so interesting things can happen.

Just for kicks here's a more hopeful approach that I've heard...merge AI with the brain like Kurzweil was predicting. Imagine if everyone was a genius and telekinetic so then we get the neural internet from that. We would actually be less divided, more connected and more productive as a human race. Our current sense of self is too selfish and our intellect too limited. I'm sure it would be very difficult to hold on to beliefs devoid of reason in that case so everyone could make better decisions, not just about themselves but about humanity, etc. because they could perceive things at a grander scale.

View on HN

I think this is already happening. In fact most AI pessimists don't consider that "man + machine >> machine." Consider chess. Kasparov's defeat happened in mid-1990s. Machine is considered > man. Yet, in freestyle chess, where players are allowed to consult machines, we have achieved ever greater heights of performance. In chess today, any sufficiently advanced chess engine can be incorporated into a (man+chess-engine) player to perform better. Similar things will happen in Go, Poker, etc. The handicap of not using a computer to inform your moves exists to make things more competitive, it doesn't mean that we are worse off!

Taking this further, most technology we use daily (word processors, email, the Internet, etc.) are enhancements of the man+machine organism and enhance our performance as well as the 'machine''s.

Yah I think so too. Super Tangent warning: I recently saw the documentary on Koko the Gorilla and was amazed at her emotional intelligence and communication skills. Why not integrate AI into other animals besides humans?

One more tangent: The perception of time would be very different if we were highly integrated into machines right? What if you could perceive things at 1000 times the normal human?

> In chess today, any sufficiently advanced chess engine can be incorporated into a (man+chess-engine) player to perform better.

This worked between about 2008 and 2013; but those days are long one. Engines such as Stockfish 8 surpass both humans and human+engine "centaurs". That is, incorporating a human into the loop only makes things worse.

This is just an artifact of how chess engines work. They are really dumb and just search a huge number of moves into the future. Combined with human pattern recognition, they can be improved. Humans can see patterns that the chess engine doesn't.

However Go was beaten by a neural network trained on millions of games. It already has pattern recognition. At least on par with humans, and probably superior to them. In the famous AlphaGo match, it considered and made moves that looked stupid to humans... at first.

I really doubt human + machine will do any better than just machine at Go. Or anything NNs take over. Human brains can't possibly compete with a machine that can iterate over billions of training examples in a few hours.

> Just for kicks here's a more hopeful approach that I've heard...merge AI with the brain like Kurzweil was predicting

Shouldn't this be the ultimate goal anyway? I know Elon Musk sees the future where there is a merging of man and machine to enhance what he sees as man's limited upload bandwidth (talking, writing, gesticulating, etc.)

Yes, I hope it is. I just was writing it in a style of a possibility rather than a prediction as its the same thing to be very positive or very negative, at least in terms of making errors into the future.

"I'm sure it would be very difficult to hold on to beliefs devoid of reason in that case..."

No offense but I literally laughed out loud when I got to that part. Reminds me of what people said about the internet. Doesn't that comparison lend credence to the notion that greater access to information might often lead to a retreat into comforting echo chambers? That greater ease of communication might often increase the vectors for infection for unreason?

The difference being...perception right? Currently we have a huge amount of information but the same limits of comprehension. We didn't suddenly get new anatomy when the internet came online. Its apples and oranges. If you are able to understand things at great depth, you are able to understand information at great depth and know the difference between bunk and not. Myths and legend-type information becomes harder to rationalize as truth because the origins of it do not result in a high probability of accuracy in which you are highly capable at that point to determine in a sub-microsecond.

To be fair, we are still fairly early on with this stuff.

Yes, folks have their echo chambers. But alternatively, perhaps we simply teach the new generation better critical thinking skills, introduce uncomfortable viewpoints, and how to better not fall for tricks such as "fake news".

Just because we have problems with echo chambers now doesn't mean that we will in the future. Heck, in a way, the internet is teaching us about our echo chambers - it isn't a new phenomenom (families and community provided it for us), but having a choice in one's echo chamber really is.

Ideally, I agree with you. The other part of me sees humanity in a transition period of having all of this amazing information at its fingertips but not the smarts to use it, at least for the common person. We need a brain boost to keep up and that could just be the AI we merge with at some point or some genetic/biological enhancement or combination of those, otherwise, I just feel we peaked as a world and will just go up and down until that happens.

Reply on news.ycombinator.com