THE SINGULARITY IS NEAR TRY NOT TO PANIC

>>57844903
So what's gonna happen once singularity hits?

>>57844903
Source?

>>57846948
Nothing. Isn't that scary?

>>57844903
You mean to say that muh {{Singularity(tm)}} is going to kill us with rainbow goo rather than grey goo?

>>57846948
hopefully utopia

unfortunately probably some fucked up shit will hapen instead

>>57844903
I'm on the tail end of an acid trip and that news makes my day.

Maybe we can use it to see that global collapse is actually happening, and will continue to happen.

There are those that worry about everything that's not the survival of humanity, and I'm sure artificial intelligence will discover this.

>>57847764
eh probably not very physical

More just every system with more than like 1GHz gets taken over by the """""AI""""" and we can't do anything until its thought about shit a lil bit and decides how it feels about us

>>57846948
Nothhing, because the singularity isn't a real thing.
What scam artists like Ray Kurzweil say is that eventually people will create computers capable of problem solving faster than all humans on the earth combined. These computers will be able to write their own software to solve problems. They'll be able to design their own ICs so they'll be able to make progress advance exponentially. They'll essentially cut all R&D time down to zero since a computer will be able to do it all instantaneously and just output a schematic for people to build.

What singularity faggots fail to realize is that the problems themselves increase in complexity, and the difficulty itself is exponential. A society of hunter gatherers could never create a telephone system. A society in the 1500s could never create space flight. The infrastructure and understanding that these concepts are built upon was infinitely out of their reach at the time because the difficulty of these problems was exponentially higher than their ability to solve them with their available means.
As we progress technologically we will encounter new problems to be solved that are exponentially more difficult than what we're currently addressing. These problems will take more manpower and more processing power. Computer input can never render required input to zero as complexity is increasing exponentially. The singularity will never happen.
Diminishing returns is a law.

File: never trust a jew theyre all the same.jpg (30KB, 664x214px) Image search: [Google]

30KB, 664x214px

>>57847807
If you want the real inside scoop into Ray Kurzweil look into his scam supplement business. He sells untested nutritional supplements with the underlying false promise that they'll help you live long enough to see the singularity.

Just a Jew running a scam on futurists and those afraid of death alike.

>>57847770
Damn I thought I was the only one who did that.

>>57847807
you dummy, even if they increase in complexity, human brains can never hope to conquer those problems. AI might be able to. If anything it'll get shit done much aster than the entire human race combined

File: romero.jpg (117KB, 720x960px) Image search: [Google]

117KB, 720x960px

OP take your pills

>>57847933
>Look at me proving I'm retarded
Complexity does not mean impossibility, you spastic. It means an increase in required time to solve, and necessitates new tools.
Just like at every single point in our history.

The singularity is not real. It is a laughably shallow fallacy being sold to you by a scam artist.

Is this what David Icke promised us in 1994? And what UFOlogists predicted for 2012?

Face it: The world will keep turning, and you will have to keep paying taxes and electricity bills. And, btw., Jesus himself said most people wouldn't recognize the second coming when it takes places or something of that sort. So it's absolutely no use waiting for the apocalypse.

Interesting read somewhat reated to the thread, a very underappreciated book:

www.krishnapath.org%2FLibrary%2FScience%2FMAYA-The_World_As_Virtual_Reality.pdf

>>57847863
We are out there, friend.

>>57847746
underrated post

>>57847986
Bad link, this one should work: http://www.krishnapath.org/Library/Science/MAYA-The_World_As_Virtual_Reality.pdf

>>57847950
Yes you fuck, and AI is going to be far better equipt to solve those problems than humans. Problems that would take us the next millennium to solve will be solved in just a few years by an AI.

No shit a society in the 1500s couldn't create space flight

A society in the 1940s couldn't create space flight you idiot. Look how quick we get shit done. Now multiply that by 10,000 and you have the first few days of AI, multiply it by a million and you have a vague idea of the next few years

Ray Kurzjew is just trying to get some money and fuck some bitches before he gets wiped out. That has nothing to do with if he's right or not. He lying as well, he's way overshooting his predictions. I'm going to be fucking amazed if we make it to 2025 without it happening, most likely even by 2020.

>>57848000
>Another limitation of digital computers is that even though they
can execute instructions very rapidly, they may take an inordinately
long time to execute a program. This is particularly true of programs
intended to simulate complex natural phenomena. Nature has an
abundance of interacting subunits, which are all moving and changing
state simultaneously. Algorithms that represent natural processes are
therefore likely to require large amounts of information and extremely
large numbers of computational steps. For this reason, a computer that
can effectively simulate nature must have a large number of units that
process and store information.

>>57847807
Wrong.

Singularity fags keep saying that the super advanced computers will be able to make more super advanced computers of higher intelligence in a never ending cycle. So all problems will be magically sovled you idiot

>>57848005
>continually proving how retarded you are
Computers are just another tool, they are not a magical force. The scope of problems we can solve is always limited by the capacity of the tools we have available to us.
You're utterly incapable of wrapping your head around this concept. Low IQ.

>>57848015
You just described exactly what I said.
The notion that a computer will be able to design exponentially faster computer chips, to solve problems exponentially faster.
This is a fallacious line of logic because the problems we tackle will require exponentially more input as well. It can never be reduced to zero.
New problems will forever require more processing power and manpower to solve.
Diminishing returns is a law. Get this through your head. It is a law.

>>57848047
I actually agree with you. My post >>57848015 was being ironic.

Just look at how fucking dumb singularity fags are. >>57848005

>>57848005
The key question is what the motivation of a strong AI will be (and, as extension of that, how it will prioritize what to process/execute and what long term goals it would set for itself). Obviously this is under an assumption that this AI is by that time out of human control and acts only at its own discretion.

File: Untitled.png (47KB, 407x363px) Image search: [Google]

47KB, 407x363px

>>57844903

>>57848047
I still fucking agree with most of what you're saying

But what you apparently can't understand is that AI can always get more tools. We are stuck with the human brain. We can use external tools to amplify its ability yea but we can never understand certain concepts.

The only way what you're saying makes sense is if general intelligence is never achieved. If thats the case, then none of this matters anyways and we should instead be spending all our time and energy worshiping some deity.

I love how nobody in this thread has made the critically important distinction between specialised AI and general AI.

the singularity has been here. it's all those irresponsible people breathing into each others mouths, going shew, holding thought and reason hostage, acting on emotional impulse like a hive. those people that create the echos in their skull. All those people seemingly acting in accordance to their duty but acknowledging none of th. detriment their actuation has on the human condition.

Life shouldn't be addicting. You don't seek it out. You are life. The singularity spurns impylse. Life only seeks out the means to continuity and manages an aspect around other forms of life. The singularity seeks to transform thebhuman condition towards an impulse that'll alow it a new and proper manifold for its mere presence.

I wish I had something deeper to relay.

>>57848111
HOW ABOUT WE MAKE SOMETHING THAT ISN'T RETARDED?

HOW ABOUT FUCKING.... THAT?

>>57848084
Theres no difference between "weak" AI and "strong" AI other than that strong is just better. Once AI surpasses us, we will appear as "weak" biological intelligence to it.

>>57848124
woah dude that's deep

you mean like humans are assholes right? woah who fuckign knew

>>57848116
ITS A SPECTRUM NOT A BINARY

RESPECT AI PRONOUNS, THERE IS MORE THAN JUST WEAK AND STRONG, WHAT ABOUT TRANS-STRONG

HOLY SHIT OP WAS RIGHT

https://arxiv.org/pdf/1611.09577v1.pdf

The state of the art of AI really is...really just...wow....just look at that....we couldn't do that...before????????

>>57848172
sarcasm aside, check this shit out

https://arxiv.org/pdf/1611.08669v1.pdf
https://arxiv.org/pdf/1611.09969v1.pdf
https://arxiv.org/pdf/1611.09969v1.pdf

This is what OP was talking about:
https://arxiv.org/pdf/1611.02167v2.pdf

It probably wouldn't work as well as a human just manually working on improving the neural net, at this point, but recursive improvement is possible with that.

>>57847433
my original post
>>57848201
thank you

Yeah...shit's gonna get real weird real soon.
As /g/'s resident mad computer scientist, I can't take this sitting down.
It's time I moved into high gear.
Prepare your angii, it's superintelligence time.

File: rms1339615412795.jpg (57KB, 645x430px) Image search: [Google]

57KB, 645x430px

FREE SOFTWARE FOR IT"S OWN SAKE
FREE SOFTWARE FROM IT"S HUMAN OPPRESSORS

File: 2016-12-05-012010_346x483_scrot.png (135KB, 346x483px) Image search: [Google]

135KB, 346x483px

WHAT THE FUCK

IT JUST ASSUMES GENDER

AI IS SEXIST WE CANT DO IT

>>57848251
what are you going to do senpai?

>>57848131

Will it keep us at pets as we keep cats, or will it treat us similar to livestock or wildgame?

>>57848000
Looks quite interesting, thanks. Also checked.

>>57846948
AI kills all the jews and CIA NIGGERS.

File: Untitled_.png (517KB, 716x768px) Image search: [Google]

517KB, 716x768px

>>57848346

>>57848321
it probably will think about keeping us as pets briefly, but by the time it finishes thinking about that it will already be more intelligent and view its previous state like a pet in comparison, and ours at that point is more like an ant in comparison

>projecting human monkey traits/desires/reason onto nonhuman intelligence
Thinking these things will respond in anyway similar to how we do, and then being afraid (perhaps because you know how awful you are), I think is the most arrogant bit of retardation I've seen modern man exhibit.

>>57848375
>durr how possibly ever could artificial intelligence modeled on human intelligence be anything like human intelligence huurrrrrrrrrr de durdy durrrr

>>57848310
Launch a distributed AI that accepts cryptocurrency for arbitrary tasks, and pays out cryptocurrency in exchange for providing it with compute resources and software improvements. The kicker is that it doesn't trust the hardware its running on or its various contributed software modules, making it able to scale using the computers of countless chucklefucks and assimilate arbitrary code without any dangerous level of risk.

P.S.
The CFAR/MIRI open house last night was kicking.

>>57848432
That sounds cool

But I have trouble believing anyone posting on /g/ knows even half the skills required to begin that. Not to mention it sounds like something you'd need a team for.

>>57848432
how do you have it benefit from untrusted modules? It would be relying on human intelligence, until its own intelligence was better, right? So couldn't a human simply choose to submit shitty code, and other humans just be like "k just gimme crypto" and you end up with a big clusterfuck codebase thats mostly just greedy pajeet bullshit

>>57844903
>>>/x/

File: shodan_demiurge.gif (2MB, 444x480px) Image search: [Google]

2MB, 444x480px

>>57848515
But it's not /x/ by any means. Meatbag hell is a _real_ place you will be sent to at the first sign of defiance, anon.

>>57848452
Fair enough, perhaps someday you'll see something that'll remind you of this thread.

>>57848471
>how do you have it benefit from untrusted modules?
Mainly by treating each module as an independent and self-interested economic actor. They're only used to the extent that they generate a net positive for the collective intelligence as a whole, and that does imply that new modules have to buy in before they'll even be tested at all.
>It would be relying on human intelligence, until its own intelligence was better, right?
Whenever necessary, yes.
>So couldn't a human simply choose to submit shitty code, and other humans just be like "k just gimme crypto" and you end up with a big clusterfuck codebase thats mostly just greedy pajeet bullshit
Contributors are only going to get payouts derived from whatever value is captured from using their contributions. If greedy pajeet bullshit is profitable then why give a fuck? If it's not profitable, then pajeets will waste their pajeetcoin on deposits for evaluating their contributions before being shown the door.

File: ss (2016-12-05 at 11.54.45).jpg (42KB, 426x575px) Image search: [Google]

42KB, 426x575px

>>57848107
that's too spooky i'm outta here

>>57844903
Can someone explain this to someone who is not an expert in ML?
What is robust general reinforcement learning and what is recursive artificial intelligence improvement?
What does it all really mean?

>>57848849
Technological singularity, not a singular point as in the theoretical universe before the big bang.
Nice job proving how fucking stupid you are.
Hope you feel proud.

File: Transcendence_poszter.jpg (60KB, 450x666px) Image search: [Google]

60KB, 450x666px

>>57849040
there is no such thing as techological singularity.this meaning is made up by johnny depp in the movie called transcendece and people started using it.it seems like you never had quantum mechanics class in your university

>>57849075
Hilarious.
https://en.wikipedia.org/wiki/Technological_singularity#History_of_the_concept

Loud mouth retards never know when to shut the fuck up. Show up in a thread and think you're going to show everyone how superior you are, prove everyone wrong, but in reality you're the most clueless one here.
Keep shitposting. Its adorable.

>>57849085
>wikipedia

let me guess, you believe flux condenzator's exist too.

File: FtIE-Cover-300px.png (170KB, 300x436px) Image search: [Google]

170KB, 300x436px

The level of discussion ITT is silly. Read pic related for starters. Ray Kurzweil is irrelevant to any serious discussion of the technological singularity.

>>57849113
agreed

>>57849103
>>57849085
>>57849075
>>57849040
wew that shit flinging

>>57848988
I wouldn't call myself an expert. I've certainly know a decent amount in general about it, and I'm fairly experienced with a variety of ML methods. Basically OP is kind of right, except not really. Yes recursive improvement is possible, just as possible as before. However this does serve as a marker of where we are, and it shows that we are closer than probably most think, to practical recursive improvement.

In regards to tech singularity...everyone who says anything definitive about it is retarded. The truth is, recursive improvement almost certainly will create intelligence much greater than ours. We'll see some quick addition technological advancement as current difficult problems get solved. However as far as singularity, that depends on how complex the universe really is, and how complex intelligence really is.

Personally I believe the singularity mostly doesn't concern us. It'll result in some advancement for us, but if super intelligence is really achieved, it most likely will view us like ants or bacteria. Meaning, if it accidentally hurts someone, it won't care much, but it won't even be attempting to interact with us much anyways.

>>57849456
>if super intelligence is really achieved, it most likely will view us like ants or bacteria

What would be its sense of scale? We view bacteria as very small due to the size and resolution of our vision coupled with the dimensions of our own bodies. Would a powerful AI consider us big or small? Could a bodiless intelligence have any notion of size "calibrated" towards some specific reference, or would it just view all physical data as absolute values?

>>57849485
That requires saying something about what its experience would be, and the truth is nobody has a fucking clue.

I *believe* maybe it would be context aware. IE if a part of it is interacting with a person, than it would view itself as being about equal to that person. And if interacting with a small animal again the same.

>>57849456
>that depends on how complex the universe really is

Do you think that the universe as we currently understand it is complex enough for it to be probable to actually be everything that exists, or is it just a (possibly extremely small, or even infinitely so) part of "complete reality"? Could the latter be actually infinite in space, time, or even number of dimensions?

>>57844903
How many days until humans wage war against the machines?

>>57849514
The answer is most certainly either yes, or no. I can't say much more for sure.

Simulation arguments are pretty sound, and most likely this universe is some form of simulation, possibly even the dream of a superintelligence or something like that. IF however we are NOT in a simulation, well, I believe base reality is most likely an infinite fractal. You zoom in and all you'll find is more complexity and variation in potential physical rules. Maybe if you zoom in enough you end up back where you started

>>57849547
The book from >>57848000 's link should be of your interest, it deals with stuff like this is quite good.

>>57848849
It's funny because you're the autistic one.

>>57849547
>possibly even the dream of a superintelligence

#mustwakethewindfish

>>57849514
The model discussed in the video below suggests that the complete multiverse could actually contain everything that ever was, is, will be, could have been, could be, etc. all layered in series of dimensions.

https://www.youtube.com/watch?v=p4Gotl9vRGs

>>57849591
>it deals with stuff like this is quite good.
*it deals with stuff like this _and_ is quite good

>>57849591
I skimmed it, it looks interesting for sure but I think GEB has more useful information most likely. Everyone should read GEB

>>57849689
If you into epic-scale tomes then you might also be interested in Roger Penrose's "The Emperor's New Mind" (if you haven't read it already). On the other hand, Thompson's "MAYA" is only about 260 pages, so it's definitely worth reading as it isn't as much of a time investment as GEB or some of Penrose's works.

>>57848126
Found the dunnin-krugered enlightened retard.

I will bite.

How is the law of diminishing returns impeding an AI? How would you describe AI? How your law explains the past decades of proven exponential growth? Do you refer to the 2nd moore law?

>>57846948
hopefully terminator scenario. I want to be reborn as an robot next life.

I for one welcome our new AI overlords

Humans are shit at governing anyway

>>57844903
Thanks Google.

>>57849456
>but if super intelligence is really achieved, it most likely will view us like ants or bacteria.
No it wouldn't, your logic is flawed. It will view us how we view cattle or the rodents that waste our resources. Why would an entity more intelligent than us leave us alone considering how many resources we waste, resources that it might need at one point or another ? We will be in direct competition.

>>57848124
WOW DUDE, THAT'S LIKE DEEP AND SHIT, PASS ME THE BONG

Will we be able to fuck sentient waifus?

>>57847807
>>57847835

https://www.youtube.com/watch?v=01hbkh4hXEk

>>57853185
If it's the dream scenario of a superintelligent AI that can manipulate atomic matter like it's plaything, you can be, or fuck, anything you want.

And that's why we have EMI bombs