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

Hawking: AI could end human race

bbc.com

Post reply on HN
Print view View original

111–118 of 118 posts

Re: Hawking: AI could end human race

#111

Post by sailingparrot »

xienze wrote: post #101
sailingparrot wrote: post #67

Earlier quoted context omitted.

70% of hacker news readers develop software for a living, that doesn't make us in any way competent to judge that. I would even argue that may make us less competent, because we use our day to day development routine as a reference for that, while in fact it's very highly probable that the first true AI will have nothing to do with the technologies and paradigm you use everyday. I highly doubt it will coded in Java a…

> Moreover who told you a dangerous AI needed to be based on human brain? All these doomsday AI scenarios are predicated on a machine becoming "self aware", having independent thought, the ability to handle self-directed learning, etc. Sounds a lot like human intelligence.

Yes, it's a possibility, and I based the rest of my argumentation on a human brain model scenario.

Re: Hawking: AI could end human race

#112

Post by rbdn »

shittyanalogy wrote: post #93
sailingparrot wrote: post #88

Earlier quoted context omitted.

Thank you for the compliment. Stephen Hawking partakes in research, with Intel, mostly on speech recognition, to help him being able to better express himself. Elon Musk funds several startups in the field of advanced AI.

You don't get it. Science relies on published peer reviewed, repeatable, testable evidence. Not the authority of famous, rich, or even smart people. You can accept authority as much as you personally want, but the reason society progresses technologically, and even towards an artificial intelligence, is because of published, peer reviewed, testable, repeatable evidence. If what they fear is backed by such evidence th…

Publishing and peer-review is one part of science, but you are missing an equally important one, which is coming up with (hopefully testable) hypotheses.

But how can you have something testable about something that does not exist yet? You would have to have evidence that an actual, intelligent program does something hostile, which is obviously impossible at the time since AGI has not been invented yet.

However, we do have evidence of naturally occuring hostile intelligence. If an AGI is vastly superior in some regards (for example through much tighter integration in knowledge and control systems), it could in fact pose a danger. The article even clearly states a hypothesis: "Humans, who are limited by slow biological evolution, couldn't compete and would be superseded."

We can’t really assign a probability of occurrence to scenarios like these, thus we have to include it in the list of events that could potentially wipe us out and which we may be able to prevent. Since people are actively working on AGI, it should probably be somewhere at the top where nuclear warfare, global warming and comet impacts are listed.

Re: Hawking: AI could end human race

#113

Post by hcarvalhoalves »

ceejayoz wrote: post #16
hcarvalhoalves wrote: post #7

We fear what we don't understand, so I guess all those testimonials reflect their understanding of strong AI.

That's a fairly useless platitude. We fear plenty of things we understand.

Why read it as "we fear everything we don't understand"? Of course I'm talking about irrational fears. Strong AI not only doesn't yet exist, there's not even agreement on what it is.

And in the topic of useless platitudes, "X could end the human race" is also one.

Re: Hawking: AI could end human race

#115

Post by leichtgewicht »

https://www.youtube.com/watch?v=_ZG8HBuDjgc#t=1985 (Douglas Adams talking, fairly humorously, about extinction)

"ending the human race" is such an interesting thing. Put aside that bone-headed nazis will tell you that humans have more than one race: We do extinct other species in incredible amounts (ask any greenpeace flyer). If we were extinguished by our own invention seems sort of fair.

But by the off-chance that an AI figures out that we are actually worth being, then there might be an realistic chance that it will actually help us survive our own mistake.

A bit of a gamble but if curiosity kills the cat then we are lucky to be alive anyways :)

Re: Hawking: AI could end human race

#116

Post by DougN7 »

Honest question: do any lowish-level programmers (C, C++, etc) have this fear? I pick this group because they are close to the metal so to speak. I'm in that group and don't have this fear. To me a computer is just a super fancy pocket watch (or Turing machine) with no chance of becoming self aware.

Re: Hawking: AI could end human race

#117

Post by pluma »

butwhy wrote: post #42
pluma wrote: post #18

Earlier quoted context omitted.

Intelligence does not protect against stupidity, sadly. I'm all for siding with Musk, but authority does not guarantee correctness.

If you want to make big claims like that, you better be able to back them up. So go ahead, cough up a few hyperlinks to some conclusive and concrete evidence and research that absolutely rules out the potential of humanity-harming AI in this context. Otherwise, you can quit your trolling.

How is the null hypothesis that AI is harmful?

Re: Hawking: AI could end human race

#118

Post by pluma »

marvin wrote: post #47
pluma wrote: post #15

There isn't a market for strong AI yet, anyway. We don't want machines that think like humans. We want machines that can perform certain tasks with an error rate less than or equal to that of humans (adjusted for other economical factors like price of labour). While strong AI is a great plot device for science fiction, it's still fiction. Sure, there are attempts in academia to replicate human intelligence, but even…

I disagree. If you could run a machine that performed equally well in a real-world environment as a human, at a lower cost than it costs to feed and pay a human laborer, it would be a no-brainer to invest in thousands of these machines to perform various consulting services that are bought and sold over the internet. And here I've only touched the slice of the world economy that exclusively consists of information-in…

Why do you think we need general AIs to replace human labour?

I'm struggling to find any cost-intensive labour that would require anything beyond a small group of specialist AIs to replace human labourers.

Also, these discussions tend to assume that we (by which I mean those making these decisions) want to replace every profession with AIs. We don't. Even if we fully automate certain forms of labour, we still want human oversight, if only so we have someone to blame in case something goes wrong.

Post reply on HN