The Singularity

[Cikomyr]

Hi People, I was re-reading Questionable Content, and they mentionned the Singularity a few times, and I got thinking about it.

(For those people who don't know what the Singularity is, it's practically Rapture for Nerds. It's the point in time where we will develop an AI that can reprogram and improve itself, which then will be able to improve itself ad nauseam at infinite speed).

I was wondering it anybody here actually did any research on the topic. Is it actually possible? Could it happen in our lifetime?

[pffh]

It's A singularity. A technological singularity is a point of time from which the people before that time can not imagine how societies look and behave after that point because of an invention at that time. There have been many such singularities throughout our history.

As for that specific one nah I doubt it will happen in our lifetime. Would be neat if it did though.

[Eldan]

There's A Singularity,and there's THE Singularity. They are not quite the same. The Singularity is the point where we can improve the things that let us improve other things. Like,say, design a computer that can design a computer better than itself. Or improve our own intelligence.

In our time? Probably not. We have whta is more or less a one atom quantum computer that doesn't do much yet and neural networks and evolutionary algorithms are very, very cool, but so far, they lack a lot of the things humans can do and aren't really AI yet. And improving human intelligence is incredibly nebulous.

And even with the intelligence singularity, there's still resource limits. Just because our new intelligent supercomputer would know how to build a better, more intelligent supercomputer than itself, we would still need the time, money and power to actually realize it.

[Yora]

Rapture for Nerds says it very well. It's basically a doomsday cult. An unimaginable entity will be created that will reach an intelligence of infinity, and it will, in it's infinite wisdom, make us all immortal and also infinitely intelligent, and there won't be any material needs anymore and thr entire world, even existance itself, will be different from anything we could ever imagine.

And it's going to happen in our lifetime.

[Eldan]

Hey, now. We Moderate Singularists don't like being called Doomsday cult. There won't be any doom

[Kitten Champion]

My living to see a singularity is not impossible. Then again I'm a millenial or whatever, so that's something like 70-80 years. I don't even know if we'll recognize a singularity as a singularity if it does come, it might be a process of change that we undergo that's so subtle that the average person doesn't even consider it. I doubt it will be "Hark! The Machine God Commeth!" at any rate.

The problem with anticipating the rate in which humans innovate is you're going to be wrong one way or another. You might as well guess what the weather will be like 10 summers from now. You've got something of an idea (i.e. hot), but the factors involved are too numerous to make it a worthwhile exercise.

[pffh]

Hey, now. We Moderate Singularists don't like being called Doomsday cult. There won't be any doom

Bah you're moderate now but before you know it you will be knocking on peoples doors and upgrading their computers.

[Gavinfoxx]

Bah you're moderate now but before you know it you will be knocking on peoples doors and upgrading their computers.

You mean their brains?

[pffh]

You mean their brains?

Brains computers, tomato potato.

[warty goblin]

I've never really liked this idea. Humans are bad enough for everything else on the planet. Last thing we or anything else needs is a super-machine to be horrible to us and every other living thing on earth.

[Eldan]

Why should it be horrible to us? Surely you don't believe in the Termiatorians and their propaganda.

[Cikomyr]

Why should it be horrible to us? Surely you don't believe in the Termiatorians and their propaganda.

Have you even seen Pintsize on Questionable Content?

Terminators are a boon compared to the lil' ****er

[Gavinfoxx]

You do know that there are lots of people trying to figure out the various pitfalls in all of the plausible scenarios that lead to the Singularity, and trying to identify an ethical path to Friendly AI and the creation of intelligences and various 'safe' options that aren't existential threats...

[Kitten Champion]

Not giving it the launch codes to the nukes seems a good start.

I'm more worried that any true AI will be owned in perpetuity by its creators. Similarly to how GMOs are IP for companies, and whether that will extend to any created sentient life I've no idea. That the legal and cultural definition of a person has been slow to incorporate Others is fairly well established throughout our history.

[Emmerask]

An unimaginable entity will be created that will reach an intelligence of infinity, and it will, in it's infinite wisdom, make us all immortal and also infinitely intelligent, and there won't be any material needs anymore and thr entire world, even existance itself, will be different from anything we could ever imagine.

Hm I don´t quite see how infinite intelligence is possible, unimaginable in comparison to the limits of humans yes why not, infinite in a finite universe with finite resources not really.

[Gavinfoxx]

Not giving it the launch codes to the nukes seems a good start.

Oh if this is done wrong, it won't need launch code to nukes to win a war or make wmd's. All it really needs is to convince one person to let it out into the wild. "Here's the cure for your daughter's cancer, will you please let me out now?" and bam, it's out...

[Cikomyr]

Not giving it the launch codes to the nukes seems a good start.

I'm more worried that any true AI will be owned in perpetuity by its creators. Similarly to how GMOs are IP for companies, and whether that will extend to any created sentient life I've no idea. That the legal and cultural definition of a person has been slow to incorporate Others is fairly well established throughout our history.

While I understand your point, I really wonder how our legal system will work out when we achieve the creation of Artificial Sentience.

I mean.. what if we created a genetic monstruosity from scratch? Would we be allowed to own it, even if it's intelligent?

[Eldan]

Ask Craig Venter. I think he does own a genetic monstrosity made from scratch.

[Gavinfoxx]

One particular place where people are trying to define the scope of the AI stuff, and do research and figure out the pitfalls, but they are working towards a particular Singularity...

http://wiki.lesswrong.com/wiki/Singularity
http://wiki.lesswrong.com/wiki/Intelligence_explosion
http://wiki.lesswrong.com/wiki/Frien...l_intelligence
http://wiki.lesswrong.com/wiki/Unfri...l_intelligence
http://wiki.lesswrong.com/wiki/Artif...l_intelligence
http://wiki.lesswrong.com/wiki/Paperclip_maximizer
http://wiki.lesswrong.com/wiki/Existential_risk
http://wiki.lesswrong.com/wiki/Creating_Friendly_AI

[Carry2]

You do know that there are lots of people trying to figure out the various pitfalls in all of the plausible scenarios that lead to the Singularity, and trying to identify an ethical path to Friendly AI and the creation of intelligences and various 'safe' options that aren't existential threats...

Here's a problem with the Friendly AI scenario: I'm a little puzzled by why a hypothetical species of physically immortal, mentally evolved, and ethically superior cybernetic beings ought to be doing what we want them to.

[CarpeGuitarrem]

Personally, I think that abstract thought is going to be a "carrying capacity" for AI, but I could be wrong.

[Zeful]

Why should it be horrible to us? Surely you don't believe in the Termiatorians and their propaganda.

Terminators? No. How about corporations? A company would gain no benefit from it's magic computer helping it's competetors, and with shareholder greed masquerading as "rights" such a computer will be vital in out preforming it's competetors to the point they shut down.

So the basics of corporate competition requires that the system be inherently horrible to someone.

[Gavinfoxx]

Here's a problem with the Friendly AI scenario: I'm a little puzzled by why a hypothetical species of physically immortal, mentally evolved, and ethically superior cybernetic beings ought to be doing what we want them to.

Because they want to? That's kinda the point of Friendly AI. That they would want to. Just like how Gandhi would want to help humanity.

[Carry2]

Because they want to? That's kinda the point of Friendly AI. That they would want to.

I didn't say 'how they could'. I said 'why they ought to'. What is the ethical justification for making superior beings, willing or otherwise, into our slaves?

[Gavinfoxx]

I didn't say 'how they could'. I said 'why they ought to'. What is the ethical justification for making superior beings, willing or otherwise, into our slaves?

That's the point of the discussion on that site? To create partners, not slaves.

[Carry2]

That's the point of the discussion on that site? To create partners, not slaves.

But we're talking about beings which can outperform us on essentially every level. There doesn't seem to be any easy way for the relationship to be reciprocal, aside from... I dunno, scientific curiosity on their part.

[Gavinfoxx]

Well there isn't an easy way. That's why this is such a hard problem, that takes a lot of deep discussion and understanding and thorough undertaking, to figure out the best possible way and the various desired end goal situations.

[Carry2]

Can I have a robotic Lena Olin?

[GenericGuy]

Unfortunately on forums that don’t normally deal with futurology and Artificial Intelligence; arguments against the Singularity go “it’s scary and bizarre, I don’t like it, and therefore it won’t happen,” and those for it will say basically “it’s cool, I want it to happen, and therefore it will happen.” I recommend going to sites that do deal with this topic more exclusively to get better well thought out information, but if really just want random playgrounders opinions. The Singularity will most probably happen in the end of the 21st century, the rate computer science is progressing and backwards engineering of the human brain means we will be able to create and artificial mind of human level. By virtue of this mind being artificial, with functions inherently “superior” to an organic one (being able to think at the speed of light for one and a Eidetic memory), unless you believe human level intelligence is the limit of possible intelligence (which is extremely doubtful) this new mind can improve itself.

[Carry2]

Because she's clearly not human. Seriously, she doesn't look a day over 40.