Google DeepMind Founder Demis Hassabis: Three Truths about AI • Hacker News Board

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#121

Post by gowld » Tue, Oct 02, 2018, 7:33 PM UTC

munificent wrote: post #118

mindcrime wrote: post #75

Earlier quoted context omitted.

Our children will be astonished and mystified - and furious - to see all the things we got up to instead of dealing with this thing that is the only thing that matters. That's been happening since Generation X kids were old enough to understand what a mess our parents made of the world. Unfortunately we haven't done a very good job of making a difference. Frankly I'm a little disappointed in my cohort in that regard.…

I am really optimistic about the younger generation, actually. I'm a gen-Xer. Our mantra was "whatever". Instead of trying to solve problems, we just tried to detach from them. As long as we were self-aware and cynical about what we were being sold, we thought things would be fine. But that's obviously stupid in retrospect. People in power love cynicism — why bother defeating other people when they'll defeat themselv…

The woke generation before the Gen X was the....baby boomers. Remember civil rights and environmentalism? Perhaps the simplistic generational model is a poor fit for reality.

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#122

Post by gowld » Tue, Oct 02, 2018, 7:37 PM UTC

oh_sigh wrote: post #116

mrhappyunhappy wrote: post #94

Earlier quoted context omitted.

What’s the point of living if you can’t achieve your biological function?

If procreation was the sole important biological function, gay people would be much rarer.

Not necessarily.

https://www.sciencedaily.com/releases/2010/02/100204144551.h...

"Summary: Male homosexuality doesn't make complete sense from an evolutionary point of view. One possible explanation is what evolutionary psychologists call the "kin selection hypothesis." What that means is that homosexuality may convey an indirect benefit by enhancing the survival prospects of close relatives."

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#123

Post by oh_sigh » Tue, Oct 02, 2018, 7:41 PM UTC

gowld wrote: post #122

oh_sigh wrote: post #116

Earlier quoted context omitted.

If procreation was the sole important biological function, gay people would be much rarer.

Not necessarily. https://www.sciencedaily.com/releases/2010/02/100204144551.h... "Summary: Male homosexuality doesn't make complete sense from an evolutionary point of view. One possible explanation is what evolutionary psychologists call the "kin selection hypothesis." What that means is that homosexuality may convey an indirect benefit by enhancing the survival prospects of close relatives."

Right, but OP seemed to imply that an individual choosing to not procreate was against some kind of basic biological functioning.

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#124

Post by adverbly » Tue, Oct 02, 2018, 7:45 PM UTC

smallnamespace wrote: post #86

mindcrime wrote: post #75

Earlier quoted context omitted.

Our children will be astonished and mystified - and furious - to see all the things we got up to instead of dealing with this thing that is the only thing that matters. That's been happening since Generation X kids were old enough to understand what a mess our parents made of the world. Unfortunately we haven't done a very good job of making a difference. Frankly I'm a little disappointed in my cohort in that regard.…

To tackle big problems, you need the structures in place to organize people, lead them, figure out not just the grand goal but the sub-objectives to be met, and to reward those who contribute and punish cheaters, and also the money and resources to do all of those things above. But at this point you're no longer talking about getting a diffuse group of people together who believe in the same thing, but building somet…

Except that the structures we have in place have time horizons even shorter than a human lifespan. Stock market runs on quarterly reports, governments run on election periods of a couple years. The organizations we give the most power to just don't operate using long term incentives. And then democratic processes don't even let everyone who is currently alive vote, much less those who aren't born yet. It's a problem that largely requires unselfish thinking which makes it extremely difficult to incentivize.

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#125

Post by qbilius » Tue, Oct 02, 2018, 8:07 PM UTC

mindcrime wrote: post #93

I've been spending a lot of time lately digging into "outdated" 80's era approaches to AI. Some people think I'm wasting my time and ask "why aren't you doing Deep Learning?" This is the reason why: "You can think about deep learning as it currently is today as the equivalent in the brain to our sensory cortices: our visual cortex or auditory cortex. But, of course, true intelligence is a lot more than just that, you…

What would be a good starting point for digging into that 80s stuff, in your opinion?

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#126

Post by Notorious_BLT » Tue, Oct 02, 2018, 8:20 PM UTC

titzer wrote: post #99

ur-whale wrote: post #78

Earlier quoted context omitted.

That's because you have a working assumption there's something we can do about it. I don't believe there is: in that regard, humanity is as much in control of its own destiny as a mold colony growing on a slice of bread under a glass dome. To wit: we have been incapable of solving basic problems like ending hunger or war for our entire history. There is zero reason to believe we can actually do anything about somethi…

I was with you up until: > If a solution exists to global warming, it's expansion, nothing else. Musk is correct in that regard. And then I was like: yeah, this is exactly why we should not expand; we don't deserve it. But I'm an optimist. The universe is a whole hell of a lot smarter than we are, and it will properly contain and eliminate us.

> We don't deserve it

The universe doesn't care about your concept of 'deserving'. By the universe's definition, anything we can achieve is 'deserved'.

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#127

Post by mindcrime » Tue, Oct 02, 2018, 9:04 PM UTC

qbilius wrote: post #125

mindcrime wrote: post #93

I've been spending a lot of time lately digging into "outdated" 80's era approaches to AI. Some people think I'm wasting my time and ask "why aren't you doing Deep Learning?" This is the reason why: "You can think about deep learning as it currently is today as the equivalent in the brain to our sensory cortices: our visual cortex or auditory cortex. But, of course, true intelligence is a lot more than just that, you…

What would be a good starting point for digging into that 80s stuff, in your opinion?

Good question. There's a LOT of material out there, and it could take a lifetime to parse through it all. So my own approach has been to skim a lot and latch onto whatever superficially seems interesting at first blush, do a deeper dive, and decide whether or not to keep digging. As it happens, one of the first things I latched onto was abductive inference, and an approach to automating same called "parsimonious covering theory". I've been working on my own implementation of PCT as a way to learn that.

I also think there are some interesting ideas in Minsky's The Society of Mind approach, so I've been going through some of his stuff. That and the old "Blackboard Architecture" model, which I see as having a tenuous sort of connection.

And while it's more "90's" than "80's", I've also been reading a lot of older literature on Multi-Agent Systems lately. There's been a lot of talk lately about "Multi-Agent Reinforcement Learning" so I'm interested in that entire space.

Really, there's a thin, tenuous thread connecting most of my interests in this area (with the PCT stuff being sort of un-connected w/r/t the rest)... MAS's, Blackboard Architecture, Society of Mind, all deal with the idea of multiple "things" (call them modules, agents, components, whatever) working together in some fashion. I suspect there's "some there there" in terms of unifying all of these various AI approaches and achieving something useful. I could be wrong, but this is where I've been directing a lot of energy lately.

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#128

Post by mindcrime » Tue, Oct 02, 2018, 9:10 PM UTC

YeGoblynQueenne wrote: post #100

mindcrime wrote: post #93

I've been spending a lot of time lately digging into "outdated" 80's era approaches to AI. Some people think I'm wasting my time and ask "why aren't you doing Deep Learning?" This is the reason why: "You can think about deep learning as it currently is today as the equivalent in the brain to our sensory cortices: our visual cortex or auditory cortex. But, of course, true intelligence is a lot more than just that, you…

That particular comment is very likely traceable to some work undertaken by DeepMind in learning symbolic representations using deep learning, described here: Learning explanatory rules from noisy data https://deepmind.com/blog/learning-explanatory-rules-noisy-d... From where I'm standing (studying symbolic machine learning algorithms) the obvious next step in machine learning and AI is the combination of sub-symboli…

From where I'm standing (studying symbolic machine learning algorithms) the obvious next step in machine learning and AI is the combination of sub-symbolic, with symbolic approaches. It's just, nobody has quite figured out how to make it all work yet. There are reasons for this, not all technical, but it seems to me it's the easiest "next big thing" to at least try.

I 100% agree, FWIW. I believe that figuring out how to properly integrate those different regimes will be a Big Deal.

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#129

Post by DennisP » Tue, Oct 02, 2018, 9:34 PM UTC

gowld wrote: post #120

DennisP wrote: post #64

Earlier quoted context omitted.

A while back I read Six Degrees by Mark Lynas, who read 3000 peer-reviewed papers on the effects of climate change and summarized them, one chapter per degree, with extensive references. My takeaway from that book was that three degrees was pretty terrible but life would go on, but at four degrees modern civilization would have a hard time surviving. Go all the way to six and there won't be many humans left on the pl…

How is that possible? From the Amazon blurb, it seems that the book talks about what would be destroyed by weather change, but nothing about what would be created . If temperature increases, some current cold areas would be more habitable. If sea levels rise, some inland area would become coastal. These changes would happen over generations, so populations would move. There would be tumult in human and animal life, w…

We're currently at one degree above preindustrial, and the world looks a lot like he described for that level.

The book was published in 2009 but most of the news since has not been encouraging. A recent book that covers similar ground, also with a lot of references, is Unprecedented by David Ray Griffin; the conclusions are similar. If you want more detail on the geologic evidence, see Storms of My Grandchildren by James Hansen.

You're surely right that there will be things that improve. However, we know from geology and paleontology what's happened in previous major warming periods: mass extinction, and a huge loss of biomass, with most of the survivors clustered around the poles. That doesn't bode well for us.

Re: Google DeepMind Founder Demis Hassabis: Three Truths about AI

#130

Post by sometime » Tue, Oct 02, 2018, 9:55 PM UTC

TeMPOraL wrote: post #6

paganel wrote: post #2

To be honest, I can't really understand how come there are still people who, when confronted to a world that has been irremediably changed by humans' use of technology (for worst, they say) still believe that an improved technology used by the same humans will make things better. I'm looking at these quotes from the article: > The reason I say that is that if you look at the challenges that confront society: climate…

Frankly, I can't really understand how come people believe what you just wrote. Yes, technology created many problems, but we can't solve them by doing less technology, without reintroducing problems we've already solved. We can either not solve them, or try to solve it with more progress (likely introducing new problems as well). In other words: we dug a hole for ourselves with technology, but the only way we can di…

The problem is that more technology also creates even more dangers that we would need to work on the same time. Whether we invest more or less in tech does not seem to affect the outcome, but how much of it is done with foresight and additional safety research.