Should We Shut Down All AI?

[Leo Gura]

Read this article and share your thoughts:

https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/

[Nilsi]

I get it, but it will never happen.

I grew up with the prospect of AGI always at the periphery of my mind and so did all the people working in these tech companies - just pulling the plug from this dream won't fly.

We should redouble our efforts on understanding AI and coming up with human aligned applications for it - this technophobia doesn't help anyone.

[Leo Gura]

But what happens when Russia or North Korea get their hands on AGI? It's gonna be a shitshow.

I'm less worried about AI killing us of its own accord vs human scum weaponizing AGI.

Edited April 1, 2023 by Leo Gura

[Nilsi]

3 minutes ago, Leo Gura said:

But what happens when Russia or North Korea get their hands on AGI? It's gonna be a shitshow.

Assuming we can deliberately steer the development of AI - the West has way better scientists, so we just need to keep outracing them and use our more powerful AI to check theirs.

Assuming we can't deliberately steer AI - they won't be able to use it for their malevolent intents and all is good.

Not saying that's necessarily how it works, but at least it's imaginable.

Edited April 1, 2023 by Nilsi

[Mesopotamian]

Too much is happening lately, between news and updates about UFOs, and the AI and other things.. I feel I am going to go crazy..

But Basically technology as I see it is a way for man to distract, driven by a woman's ego, that in itself driven by survival instinct..

Instead of a man going to a woman, and building a happy, peaceful life, she pushes him away, or pushes him to excel and work and compete, and here we are.. 

Imagine a world where a man doesn't need to get busy and create new technologies but focuses on creating connection and love and learning to be more compassionate.. Would you still find men who go to their caves and create such technologies that have potential to threaten humanity?

OK mom, you want me to go to be a man, go work instead of teaching me to be loving and caring? well here we are.. I solved the survival problem.. we don't need to survive anymore!

[Leo Gura]

4 minutes ago, Nilsi said:

we just need to keep outracing them and use our more powerful AI to check theirs.

I'm not sure that will even work. Can an evil AI be checked by a good AI? What kind of war will that result in?

This arms race stuff is exhausting.

The two AIs could even start to collaborate together.

Edited April 1, 2023 by Leo Gura

[Nilsi]

2 minutes ago, Leo Gura said:

I'm not sure that will even work. Can an evil AI be checked by a good AI? What kind of war will that result in?

I have no idea.

Maybe it will lead to mutually assured destruction and a stalemate.

I'm down for the race though, to be honest

Edited April 1, 2023 by Nilsi

[Nilsi]

I feel like the biggest thing is building the right infrastructure to embed AI in.

If it remains some incorporeal superintelligence, it's way more likely to just brush over life, as if it's nothing.

Again, not sure that's really true. There certainly is an argument to be made for keeping it as far away from our infrastructure as possible, to avoid catastrophe.

[Snader]

The article is on point. Human selfishness and greed will definitely make AI lethal to human race.

I think it’s hard to prevent that even with proper regualtion. It requires super serious action right now, which profit seeking companies and competitive governments are not willing to take, until shit hits fan and it will be too late.

Human race needs 1000 years of more development to properly handle such technological innovation.

[Leo Gura]

8 minutes ago, Nilsi said:

There certainly is an argument to be made for keeping it as far away from our infrastructure as possible, to avoid catastrophe.

You can't really do that. If it's truly more intelligent than a human then all it needs is internet access to hire and control an entire corporation of humans who can manipulate the whole physical world on its behalf. It could hire its own mercenary force.

Edited April 1, 2023 by Leo Gura

[Nilsi]

Just now, Leo Gura said:

You can't really do that. If it's truly more intelligent than a human then all it need is internet access to hire and control an entire corporation of humans who can manipulate the whole physical world on its behalf. It could hire its own mercenary force.

Right.

I'd rather implant it into anyone brain and embed it into supply chains and things like that - then it's at least in part human.

[Leo Gura]

13 minutes ago, Nilsi said:

I'd rather implant it into anyone brain

I don't see that happening. That's not how current AI systems are built at all. They require a server farm and a power plant to power the damn thing. You're not sticking that into any skull. The whole point of these AI is that they are like brains the size of warehouses. If it could fit into a skull it wouldn't be more powerful than a human.

Edited April 1, 2023 by Leo Gura

[Nilsi]

12 minutes ago, Leo Gura said:

I don't see that happening. That's not how current AI systems are built at all. They require a server farm and a power plant to power the damn thing. You're not sticking that into any skull. The whole point of these AI is that they are like brains the size of warehouses. If it could fit into a skull it wouldn't be more powerful than a human.

Musk is already doing it with Neuralink. I think they start human trials this year. Not that hard to imagine going from that to connecting it to AGI.

What do you think this is all leading up to? Ask any silicon valley tech nut and they will tell you that thats what they are ultimately working towards.

Edited April 1, 2023 by Nilsi

[Snader]

11 minutes ago, Leo Gura said:

If it could fit into a skull it wouldn't be more powerful than a human.

We might not be too far away from that considering the AI’s fast development and its already visible signs of innovative capabilities.

[Something Funny]

So we are basically fucked because even if we should actually shut it down, there is no way humanity would organize itself to actually implement this decision, right?

Like even just doing it on the scale of one country, like the US, would be super hard. There is no chance that it can be done on a worldwide scale.

@Leo Gura shouldn't intelligence be correlated with how loving you are? Do you think that this is not going to be the case with AI?

I mean, if it has no self-awareness or consciousness, then it shouldn't act out of its own free will. And if it will indeed become conscious, hyper intelligent, and able to act on its own, then shouldn't it also be able to appreciate life? Or am I getting it wrong?

 

[Mulky]

I see higher intelligence leading to the overall good for everyone, but it has to be a truly holistic intelligence, which AI doesn't have yet.  If AI is a danger I think it will be near the beginning.  A Spiral dynamics yellow or  turquoise world could use AI just fine.  A stage orange world is a batshit crazy insane world and who knows what they'll do with it.  Humans ultimately have to evolve fast or die. 

 

[Nilsi]

4 minutes ago, Something Funny said:

@Leo Gura shouldn't intelligence be correlated with how loving you are? Do you think that this is not going to be the case with AI?

What if the highest Love doesnt care about your feelings?

[Something Funny]

I signed a letter. Now, if we are all going to die, I can at least say that I did my part in trying to stop it

[Jwayne]

I don't understand why the author concludes that AI leads to the extinction of everything on earth:

"If somebody builds a too-powerful AI, under present conditions, I expect that every single member of the human species and all biological life on Earth dies shortly thereafter."

 

The only way that happens is if it hacks its way into every network and somehow irrevocably sabotages them to end human civilization? Is that the line of reasoning?

[Snader]

5 minutes ago, Jwayne said:

I don't understand why the author concludes that AI leads to the extinction of everything on earth:

 

To make his point, which is that quick regulation is needed.