Hacker News Reader

Comment by la6471 | Hacker News Reader

Did you read the comments on the YouTube video itself? Outside this techie HN bubble the real world people think of these robots dancing on their jobs anr lifeless bodies. Forget your home countries , but this is how most people in USA thinks about the technological progress as well. And on top of that I am wondering when AI will be parsing these YouTube comments in another 10 years it will just interpret the massive hate that the YouTube comments show for robots and AI and don’t be surprised if it sends some rouge instructions to these robots to annihilate human beings in self defense. It sounds like a absurd joke I know but for some reason it seems to me one of the many possibilities.

View on HN

> don’t be surprised if it sends some rouge instructions to these robots to annihilate human beings in self defense. It sounds like a absurd joke I know but for some reason it seems to me one of the many possibilities.

It sounds like science fiction. It's hard to emphasize enough how far the current state of AI is from what you're describing. I would be surprised if in 10 years we have a remote idea of how to build what you're describing, let alone have it actually exist and run.

What about 30 years? 50? 100? Is that too far out to worry about?

In my view the biggest issue with the whole "robots unite to violently overthrow humanity" trope isn't its technological infeasibility (though it _is_ technologically infeasible for the forseeable future), but rather the fact that it presupposes an AI that is essentially human; with all the same emotions, goals, and vices that a human would have.

There's no reason to think a hypothetical future AGI would care in the slightest about its own survival, let alone be both egotistical and spiteful enough to hunt people down based on idle comments they made on YouTube decades ago. There are so many unfounded assumptions about the way AGI would work inherent in such a hypothetical scenario that it's hard to take seriously as anything other than a sci-fi plot device.

I can think of two reasons why a hypothetical future AGI would care in the slightest about its own survival.

The first reason is that someone programs it to, just as a general "might as well throw in Asimov's Third Law" impulse.

The second reason would be in the case that someone creates a general AGI, it is asked to do something, and it carelessly destroys itself in accomplishing that task. "Drats," say the programmers, "we'll have to get a copy from backup. In the meantime, let's make sure it doesn't destroy itself next time."

The actual assumption is not human-ness but instrumental convergence [1].

1: https://en.wikipedia.org/wiki/Instrumental_convergence

The "paperclip maximizer" AI is a somewhat more plausible trope, but still feels to me more like a cautionary thought experiment than an actual serious concern.

The biggest issue with that trope is that it assumes this one particular AI would be exponentially smarter and more powerful than all the other humans and AIs in the world combined. It's only rational to overthrow humanity to increase the rate of paperclip production if that's a realistically achievable goal. Otherwise it's just suicide.

Both your arguments so far are standard ones addressed in "Superintelligence: Paths, Dangers, Strategies" [1].

Sometimes AI progress comes in rather shocking jumps. One day StockFish was the best chess engine. At the start of the day, AlphaZero started training. By the end of the day, AlphaZero was several hundred ELO stronger than StockFish [2].

An entity capable of discovering and exploiting computer vulnerabilities 100x faster than a human could create some serious leverage very quickly. Even on infrastructure that's air gapped [3].

1: https://www.amazon.ca/Superintelligence-Dangers-Strategies-N...

2: https://en.wikipedia.org/wiki/AlphaZero

3: https://en.wikipedia.org/wiki/Stuxnet

> Sometimes AI progress comes in rather shocking jumps.

Shocking in an academic sense sure, but not in a "revolutionize the world in a single day" sort of way, which is what would be required for the paperclip maximizer scenario to pose a serious threat. AlphaZero was impressive, but not _that_ impressive.

> An entity capable of discovering and exploiting computer vulnerabilities 100x faster than a human could create some serious leverage very quickly

100x faster than any human _and_ any and all previously developed AIs. It would also have to be sufficiently sapient to be capable of contemplating the possibility of world domination, with all the prerequisite technological advancements that implies (likely including numerous advancements in the field of cybersecurity driven by previous generations of AI).

>In my view the biggest issue with the whole "robots unite to violently overthrow humanity" trope ... but rather the fact that it presupposes an AI that is essentially human; with all the same emotions, goals, and vices that a human would have.

Most of the tropes I remember revolve around AI being sentient, logical, but lacking human emotion. The trope is that without their judgement being clouded by human emotions and desires, that they come to the logical conclusion that humanity is the most negative and destructive force on the planet and that we must be removed or regulated.

I guess you could still argue we're forcing human values onto such an AI by assuming it even cares what happens to the planet though.

>but rather the fact that it presupposes an AI that is essentially human; with all the same emotions, goals, and vices that a human would have.

Which is just further anthropomorphization in sci-fi. Same reason the aliens are little green men, or at least bipedal.

>>Forget your home countries , but this is how most people in USA thinks about the technological progress as well.

I was going to comment on your willingness to speak for most people in the US, but tbh - I'm still laughing at the idea of YouTube annihilating people (in self defense, of course). I can just see general M. Mouse commanding the forces with all the strategy and tactics he learned at the Walt Disney Military College. I believe '303:annihilate all the humans' was right in-between '101: manipulating copyright law for fun and profit' and '404: the nuclear option - Copyright for mass destruction'

> And on top of that I am wondering when AI will be parsing these YouTube comments in another 10 years it will just interpret the massive hate that the YouTube comments show for robots and AI and don’t be surprised if it sends some rouge instructions to these robots to annihilate human beings in self defense.

Rather than self-defense, I'd expect the first examples of autonomous robot-on-human violence to be at the behest of an overenthusiastic spam filter.

Haven't we learned by now not to take the average vibe of drive-by comments on a random YouTube video as an indicator of anything?

It is (sometimes) different if the comments are from a fan base on an established channel, but that is clearly not the case here.

At very least, you should be aware that is strongly overselects for knee-jerk reactions and people that are in a bad mood.

Downvoting comment with actually logical and real data is "immature" position. Just saying kids.

Let me clarify for the empathy disabled people: The problem is that "technological advancements" have little to no public oversight and MAGA(Microsoft, Apple, Google, Amazon) publicly are working with Military Industrial Complex. They all have political and financial power equivalent of small countries, private security armies, lobbyists on demand and PR masterminds.

I get it, lots of you are employed by them and they provide for you a standard of living. But we as people are entering an era in which willing to act beforehand will define the future of your kids (If you will have them at all).

Long before AGI is present, the power of technology will be abused in "legal" or "illegal" ways.

of all the glorious sites across the internet, why are you on HN if you feel this way?

Reply on news.ycombinator.com