Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

money.cnn.com

Post reply on HN
Print view View original

71–80 of 98 posts

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#71

Post by icebraining »

hourislate wrote: post #59
JustAnotherPat wrote: post #13

I don't trust Zuckerberg on any issue in which Facebook has much to gain from one side. We all know how much money he'd like to make by having some AI programs follow us around 24/7.

One guy is trying to change the world and one guy is trying enslave it. When you have the likes of people like Stephen Hawking agreeing with Musk, why wouldn't you listen and be wary. Mark doesn't seem like a very intelligent person, just a very lucky person.

Musk was the guy that created a car that can track and report everywhere you go using a permanent connection to his mothership, plus films and uploads the surroundings and even has an camera inside. The only protection being his word that they won't read it unless you allow them.

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#72

Post by pesenti »

caio1982 wrote: post #56
pesenti wrote: post #53

Earlier quoted context omitted.

This is not about science fiction. It's about science today and in a near future and its consequence on policies and regulations.

I honestly believe some of Asimov's concerns regarding AIs are not science fiction but rather forward thinking and will eventually happen.

"eventually" is the key word here. Most experts in the field admit to have no clue how and when these predictions will ever come true. It makes discussing them an overly speculative exercise.

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#73

Post by convexfunction »

gonehome wrote: post #26

They're talking past each other because "AI" has become a useless term. Musk's comments are about a potential (maybe conscious) artificial super intelligence and the risks that could come from that if we enable one somehow without understanding the preconditions. Zuckerberg is just talking about "AI" in the sense that machine learning can solve a valuable class of problems. The super intelligence risk could be seriou…

> We also don't really understand how consciousness works either.

Sort of a nitpick, but people concerned about AI risk usually don't postulate that an AI needs to be """conscious""" to be dangerous; it just needs to be a thing that can apply a lot of optimization power in unexpected ways toward goals that aren't perfectly aligned with what humans generally want.

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#74

Post by drivelous »

I understand that Musk's ideas of doom and gloom concerning AI are far into the future (loceng drops the Andrew Ng quote that draws parallels to people worrying about the overpopulation of Mars) but even if that is the case isn't it still worth noting now?

My simple understanding of this all is that once we create AI that surpasses the intelligence of the human race, we as inferior beings will no longer be able to predict what they will do. If that's the case and the desire of AI (can AI have desire...?) runs contrary to human will, then there's no way to cut it that bodes well for the continued existence of the human race. And once that line is crossed, it's never going back.

Is that not a reasonable thing to be worried about even if it's 200 years away? Even On the Origin of Species was published only ~160 years ago.

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#75

Post by pesenti »

tarr11 wrote: post #57

Here is a pretty good article where many of the players are interviewed in depth. http://www.vanityfair.com/news/2017/03/elon-musk-billion-dol... “If you want a picture of A.I. gone wrong, don’t imagine marching humanoid robots with glowing red eyes. Imagine tiny invisible synthetic bacteria made of diamond, with tiny onboard computers, hiding inside your bloodstream and everyone else’s. And then, simultaneously, the…

That example has absolutely nothing to do with AI...

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#76

Post by hndamien »

rspeer wrote: post #68
danso wrote: post #50

As overloaded as the term "AI" has become, to the point where mainstream discussion is unlikely to be substantative, I'm glad a tech-business leader like Elon Musk is at least taking a public position of some skepticism. AI can be as empowering or as dangerous as we design and constrain it -- but to a layperson, AI and "algorithms" seem like inevitable magic, in the way that iPhones have steadily "improved" in hardwa…

Musk's view of AI relies on AI being magic, too. He's using the definition of "AI" that exists only in hyped-up promises about the future, and evaporates when applied to present technology. Musk dismisses AI experts who tell him that his claims bear no resemblance to real technology, because Musk isn't talking about the same AI as them. He's talking about sci-fi AI. There are no experts in sci-fi AI because, if there…

Don't underestimate the exponential.

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#77

Post by convexfunction »

pesenti wrote: post #42

A majority of the experts in AI would side with Zuckerberg on this one. See for example: http://blogs.discovermagazine.com/d-brief/2017/07/18/artific... . It's not that we shouldn't worry about AI, it's that the issue Musk is raising - AI as an existential threat -, is distracting from the real issues.

You may be interested in http://slatestarcodex.com/2015/05/22/ai-researchers-on-ai-ri..., or more directly addressing your claim, http://slatestarcodex.com/2017/06/08/ssc-journal-club-ai-tim...:

> 50% of experts think there’s at least a ten percent chance of above-human-level AI coming within the next ten years.

(ten years!? sheesh)

> And 40% of experts think that there’s a better-than-even chance that, once we get above-human level AI, it will “explode” to suddenly become vastly more intelligent than humans.

> And 70% of experts think that Stuart Russell makes a pretty good point when he says that without a lot of research into AI goal alignment, AIs will probably have their goals so misaligned with humans that they could become dangerous and hostile.

> ... 5% think this is “among the most important problems in the field”. Only a tenth – 1.4% – think it’s “much more valuable” than other things they could be working on.

Not totally clear which of those Zuckerburg wouldn't agree with, but this isn't exactly a fringe concern anymore, even if most people think it's not worth thinking about yet.

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#78

Post by circlefavshape »

convexfunction wrote: post #73
gonehome wrote: post #26

They're talking past each other because "AI" has become a useless term. Musk's comments are about a potential (maybe conscious) artificial super intelligence and the risks that could come from that if we enable one somehow without understanding the preconditions. Zuckerberg is just talking about "AI" in the sense that machine learning can solve a valuable class of problems. The super intelligence risk could be seriou…

> We also don't really understand how consciousness works either. Sort of a nitpick, but people concerned about AI risk usually don't postulate that an AI needs to be """conscious""" to be dangerous; it just needs to be a thing that can apply a lot of optimization power in unexpected ways toward goals that aren't perfectly aligned with what humans generally want.

Oh! Good insight!

... but we already have that, right? Undesirable outcomes from rulesets that govern the behaviour of a large number of (conscious or unconscious) actors are a pretty standard feature of civilisation

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#79

Post by hndamien »

AndrewKemendo wrote: post #49
arcanus wrote: post #37

Earlier quoted context omitted.

Good point. Another noted expert in "AI" is Andrew Ng (formerly Google and Baidu) who stated that, "Fearing a rise of killer robots is like worrying about overpopulation on Mars." ...For some reason Elon is really hung up on gradient-based descent methods terminating all of humanity. We just aren't close to that possibility yet. It is not even in sight.

The problem is that both are potentially right. The majority of the ML field still ignores entirely the AGI discipline - and in general rightly so as there are enough narrow problems to solve to fill a lifetime. That's the perspective of Lecun/Ng etc... because they are scientists. The others Bostrom et al. are philosophers, so they approach it from a different perspective. There are valuable reasons to think about i…

Correct. There were plenty of economists that said Bitcoin would fail, and philosophers that thought otherwise. The jury is still out, but I know who is winning.

https://twitter.com/damiendonnelly/status/91766074021904386

Re: Elon Musk says Mark Zuckerberg's understanding of AI is 'limited'

#80

Post by convexfunction »

circlefavshape wrote: post #78
convexfunction wrote: post #73

Earlier quoted context omitted.

> We also don't really understand how consciousness works either. Sort of a nitpick, but people concerned about AI risk usually don't postulate that an AI needs to be """conscious""" to be dangerous; it just needs to be a thing that can apply a lot of optimization power in unexpected ways toward goals that aren't perfectly aligned with what humans generally want.

Oh! Good insight! ... but we already have that, right? Undesirable outcomes from rulesets that govern the behaviour of a large number of (conscious or unconscious) actors are a pretty standard feature of civilisation

Yes, we do have that already, and it's kind of odd that people who see no point in thinking hard about risks to humans from this particular type of inhuman optimization process would readily agree that other types of inhuman optimization processes (with much better understood risk profiles and limitations) can be very harmful. :)
Post reply on HN