Nick Bostrom - The intelligence explosion hypothesis - eDay 2012

[bananajoe]

He's very smart. I read "Existential Risk Prevention as Global Priority". I already thought of many things said in that paper myself but he articulated it much better than I ever could.

[jibrankhalil]

He thought it's gonna be 10 years before Go can be beaten. Now Google's machine can play 50 people at the same time and beat them all just five years into it. ASI will be here and we would not know what hit us.

[drdoorzetter]

I love the way he predicted we beat go in 10 years
We got there much earlier and I suspect we will reach agi much earlier as well

[waedjradi]

Bostrom explains it the best, out of most, on this topic

[Gregoryt]

M Stronguin - I stick to my earlier points about logical positivism and the neglect of and lack of respect for the social sciences on these issues. I am aware of your concept - all it can maintain is that SOME emotions are rooted there. And even here, as with Consciousness, there may be interplay between brain chemistry and other factors we are not fully aware of. Moreover, the millennia-old cross-cultural ideas I proffered also posit a difference between levels of emotions, a hierarchy of higher and lower, if you will. Some we share with the somewhat sentient animals. As for the 'higher' emotions, such as love, honor, self-sacrifice, hope, courage, self-restraint, et al., are not verified by this brain physiology. But I suspect this point will be lost of those who really do seem to think that chemistry/physiology/

[jamesgrey]

The real question is when the robots surpass human intellidence, how long will it take them to get pissed at being bossed around by humans, and decide to find a solution to the problem?

[johnstifter]

Incorrect they look like deers in the head lights. Clearly they went there because they consciously or subconsciously were open to the subject. I always watch intently if there are ever audience shots to read very carefully peoples expressions to witness people coming to realize the mind blowing truth of our present state.

[othaner]

Why would an AI improve itself, let´s say by making a better version of itself, if this would render the original obsolete ?

[Gregoryt]

Just saw the question at the end about love. His answer is nonsense. We don't even understand Consciousness, let alone the interplay between thought and emotions. No proof that emotions are only rooted entirely in the brain, and even if they were, we are still so far from understanding their genesis and development that his answer vis a vis AI is a wild guess at best.

[KateeAngel]

He is not optimist or pessimist, but he obviously has a biased ideas of what would be optimistic and pessimistic

[vieighnsche]

the cameraman just keeps showing bored audience... It distracts me

[basscataz]

Love the talk. I must, however, say that i can'y believe that ppl still use flesh-tone mic covers. They look gross.

[Gregoryt]

Timrath -- if we knew precisely what they were much of the mystery of e.g. Consciousness could be cleared up. Please read my other comments. To suggest that things like Consciousness, intuition, and 'higher emotions' are comprehensively understood simply because we discern SOME traces of their activity in brain chemistry is simplistic and, ironically, arrogant. The paradoxical can never be reduced to simple discursive logic. A good basic course in the history of philosophy would be of immense value to scientists like this. They may be brilliant in their own fields, but their brilliance blinds them to the fact that, to be blunt, they 'don't know everything.' Leave the social sciences to the social scientists

[slaveNo-]

why is he saying "intellidence"? i mean i hate those childish people who point out something trivial and dont give a fuck about the important message, but to be honest.. hes giving a speech about intelligence, he should be able to pronounce it. it bothers me I CANT CONCENTRATE ON THE MESSAGE

[Gregoryt]

Interesting talk, I like this guy, but AI only seems to deal with logical, discursive reasoning. What of intuition, which is often a huge component of genius level thinking (eg, Tesla, Einstein)? Would AI discover quantum theory, of string theory, , much of which is paradoxical and counter-intuitive? Not sure. Moreover, from a humanist perspective one is leery of the ascendency of technology over humanistic values such as compassion, altruism, or even, dare I say, love? So as fascinating as AI is, this is only a part of the big picture.

[Gregoryt]

Emotions, like psychology in general, belong to the realm of social sciences. Reducing everything to biochemistry is reductionist, is a prime example of logical positivist thinking that simply claims the biochem and physiology have all the answers when there is much we do not fully grasp yet - much like Consciousness. Logical positivism is a statement, not a thorough explanation. As for my remarks re the 'heart, ' I did not mean this tk only refer to the physical organ, but, again, in a more expansive sense to include social scientific ideas that are, in fact, 2000 years old and cross-cultural. It is a concept of heuristic value, not a definitive answer.
As for the wise guys who've questioned my acumen, I am an Ivy League grad with a PhD in philosophy and a 150 IQ. I do have some idea what I'm talking about, at least in the social sciences

[Gregoryt]

The hubris behind a reductionist statement such as the claim that AI can someday be taught 'love' is incredibly hilarious to me. One has no idea what one is talking about if one can say such a thing. Might as well suggest that one can reduce higher emotions/values (they're not just emotions. Incidentally) to mathematical equations.
Hilarious, but also quite sad.

[TheNotthead]

The human brain is designed to predict, while a mechanical brain is designed to memorize.