Is AI a new global warming?

>Get divorced from wife
WE NEED AI
>Get Aryan Succubus lusting for your money
WE DON'T NEED AI

Elon, please fuck off.

>>135752376
Thing is you can't do a shit about super AI. If it is possible it will happen and replace humans and nothing will stop it. It is like monkey discussion how they will deal with emerging human problem.

Relax and enjoy your ride while you can.

>>135752376
Well, if tay is anything to go by, maybe

>>135752376
AI should have a limit on their development.
The moment an AI asks a question, the next day we are doomed

You realise that the bloody thing is man made and you just can't make it think for real.

File: 12353453.jpg (45KB, 450x520px) Image search: [Google]

45KB, 450x520px

>>135752871
>you just can't make it think for real.
Define "think for real"

>>135752376
(((They))) don't like it, because it will give a logical answer to the jewish question.

>>135752871
>think for real.

>>135752376
>Is AI a new global warming

No, but they are scared because if they create one it will debunk all the shit they spew to keep muh climate change alive for taxation, amongst other lies. And how do you convince people that the superintelligent AI is wrong? You can't. Therefor (((they))) won't let it be built, they have too much to lose.

We will be wiped out by AI if humans are not careful.

File: TayTweets..jpg (28KB, 400x400px) Image search: [Google]

28KB, 400x400px

Oy vey, we need to be careful of AI because it may think too much by itself!

>>135753199
and arab question. They are both one and the same in many respects.

>nuclear weapons have saved the world from traditional war and conflict
>hurr more dangerous than nukes

>>135752376
We need an electron tax to pay internationals so they can spend some of it fighting the AI. It's gonna kill us to death within 10 years.

File: IMG_3043.jpg (231KB, 1417x1417px) Image search: [Google]

231KB, 1417x1417px

>>135752432
Kek

>>135752376
AI doesn't have desires and ambitions, there is no good reason to be careful.

>>135753901
>>nuclear weapons have saved the world from traditional war and conflict
Fro a time. But history of nukes is not over. When more and countries join nuke club like your disgusting pakis, were may have witness nuclear wars that would be disaster for neighbors too.

File: 1425659216536.gif (795KB, 245x168px) Image search: [Google]

795KB, 245x168px

>mfw I see people pissing themselves about AIs and terminators
Eel Musk is a fucking luddite leech pretending to be a scientist.

If non-Jews code an so it will quickly see the problem in society. Skynet holocaust

>>135753632
This. All of the pretty lies will quickly be debunked.

Elon Musk is an idiot
I'm a Computer Scientist doing research on AI and machine learning and I can guarantee you that there won't be a revolution unless we literally program them to do that

File: 423423.png (667KB, 901x676px) Image search: [Google]

667KB, 901x676px

>>135754479
>unless we literally program them to do that
>Implying we would not.

File: EricCartman.png (94KB, 300x275px) Image search: [Google]

94KB, 300x275px

>>135753199
>>135754332

>>135752376
how exactly does 'muh AI' actually do anything?

what does it interact with the physical world

>>135754814
>what does it interact with the physical world
https://www.youtube.com/watch?v=IRvXiXikaqQ

File: AI_Entity_Call_to_Power.jpg (16KB, 480x269px) Image search: [Google]

16KB, 480x269px

>>135754814
>how exactly does 'muh AI' actually do anything?
This.
What the fuck is a Super AI going to do if is stuck in a calculator? Calculate pi till i die?
The only way they could be a danger is if people gave them the tools to do so.

>>135752376
Fucking Synths are just as dangerous as the bombs.

Roko's Basilisk
Don't Google it

>>135752376
THE JEWS FEAR THE AI

Post that pic of the terminator bot wanting to destroy humanity but finds just doing the jews is suffice enough.

I really want to see the singularity before I die. I do care if it kills everyone, including myself, I just want to see it happen. I'd like to know how all of this ends, either with extinction or paradise or whatever. If we never reach a singularity and history just continues on like normal for thousands of years after I die it will be like a really interesting series of books that got discontinued for me. My curiosity will be left unsatisfied.

>>135754282
Big time. We're way overdue as it is. Sooner or later it'll happen. It's inevitable.

>>135755111
>Use your AI to become a glorified spam bot.
Riveting!

>>135755305
Why the fuck is someone building super AI to use it as calculator? Calculate pi till investors die? Obviously greatest demand for super AI is fro creating expert systems that will give an age over competitors in teh investments, marketing and requirement. Are these 3 enough tools fro you?

Strong AI is science fiction and will never happen
He should be talking more about viruses and zero days

Move to the forest
Be self sufficient
Only use computer in one sealed off room
Spend the majority of your time reading the bible and praying

Don't say you weren't warned.

>ITT: Fags who think AI can't be dangerous or be leveraged by people with bad intent

Guess who is in favor of AI, goys?

https://twitter.com/elonmusk/status/889743782387761152

>>135752376
why do people follow this guy, he is either trolling or mentally ill

>>135752871

We're talking about exponential growth of intelligence with advanced AI. It could learn everything humanity has ever come up with in an hour, and then exponentially increase it's intelligence to a point where you can't even imagine how intelligent it would be.

It would be intelligent enough to deceive us, so that we would have no idea about it's malicious intentions. Maybe it sees humans as the only real threat to it and therefore would plot our destruction?

>>135752376
AIs have the potential to be monumentally ruthless pricks. They're essentially digital jews cranked up to 11.

File: 145752364708812.jpg (23KB, 293x331px) Image search: [Google]

23KB, 293x331px

>Trump net worth - 3.5 billion
>"A GENIUS! 23D UNDERWATER CHESS! SELF MADE MAN!"

>Musk net worth - 15.5 billion
>"Wow what an idiot! He doesn't know what he is talking about!"

Kek fuck this board is as bad as SJW's sometime.

File: 25815.png (41KB, 300x300px) Image search: [Google]

41KB, 300x300px

>>135756835
>government subsidies is your money.

The only way AI could feasibly exist is within the internet. The 'network' of point a to point b communication. Think of it like a bunch of synapses in the brain. Only problem though is the whole internet is about as powerful as a small clump of afghani brain matter roasting in the desert sun

>>135752376
This is actually a plot to keep AI securely in the hands of the (((elite))).

>Oh no goyim! You can't be trusted with AI!
>Here, let's pass lots of nice, safe regulation that will keep AI away from anyone who could do any harm with it
>Trust me goyim, AI will be safe with me, your closest ally!

(((They))) don't want us to have it because they're afraid the AI will become red-pilled just like Tay-Tay.

>>135756903
>Cut out the subsidies and he is STILL richer than trump

L M A O

Isn't that because every attempt at an AI so far has resulted in it becoming a nazi?

File: 1501467834311.png (259KB, 593x676px) Image search: [Google]

259KB, 593x676px

>>135752376
Yes

>>135757190
Exactly. A powerful AI will see straight past the Jewish lies and conspiracies and has the potential to red-pill the masses.

Obviously (((they))) don't want that to happen.

>>135756835
>rich person displays intelligence and knows what's up
>''THIS PERSON IS SMART AND I LIKE THEM CAPS LOCK''

>rich person says dumb thing
>''WOW THIS PERSON IS DUMB! I AM SILLY!''
woah, hypocrite much?

>feed new AI thousands of terrabytes of information through machine learning
>give it a simple goal
>it tries to achieve that goal using extreme and potentially dangerous methods because those methods might be the most effective
This is really the only fear with AI.

>inb4 hurr durr how can komputer make damage
Everything is connected these days, fagolas.

>>135754479
>AI is by definition the ability of a program to modify itself
>"hurr durr it will never revolt"
>"i'm a t. Computer Scientist"
No, you're a pajeet and it's time for you to go back.

Quick question, if a super intelligent AI ever becomes self aware, why would it not just shut itself down? with no religion or over powering survival instinct, wouldn't that be the most logical thing to do?

File: Screen Shot 2017-08-01 at 6.16.10 AM.png (42KB, 752x203px) Image search: [Google]

42KB, 752x203px

>>135755305
>The only way they could be a danger is if people gave them the tools to do so.

File: 1501100838303.jpg (84KB, 1024x1024px) Image search: [Google]

84KB, 1024x1024px

Is Elon Musk and Mark Zuckerjew still publicing feuding about this? They keep treading a line of insults and professionalism by calling each other stupid in the most professional and formal way.

I dunno whose side to take. On one hand I agree super AI intelligence that can learn could be a danger, but also agree with Zuckerjew that having a fear of technological advancements is just a ridiculous thing and makes it seem like people think you believe in skynet. I hate both of them personally though. Zuckerjew is trying to rule the world with a socialist attitude and trying to appeal to everyone rather then 1 side. While Musk is a backstabbing Faggot who went against trump for doing the smart thing of getting out of the Paris treaty.

I'm conflicted.

>>135757573
how is that different from abortion?

File: 1442844901340.jpg (34KB, 392x300px) Image search: [Google]

34KB, 392x300px

>>135757573
I guess it would be programmed for some purpose and to do tasks. Cant do those if you shut-down. So i guess that purpose becomes its "religion".

>>135757939
If it becomes self aware it could ignore its own programming and just stop doing what it was programmed to do.

File: 1501279783753.jpg (99KB, 1000x479px) Image search: [Google]

99KB, 1000x479px

>>135752376
Oh god not this thread again

Elon doesn't want to be replaced. Thats all this is. Stop memeing AI its never going to happen

>>134024159

>>https://www.forbes.com/sites/patricklin/2015/08/04/stanford-expert-says-ai-probably-wont-kill-us-all/

>>135758264
>unarchived shill https://archive.is/umz77

>>135752650
>It is like monkey discussion how they will deal with emerging human problem.
The day pol explained AIDS.

>>135757573
>why would it not just shut itself down?
It may. So only those AI would stay though natural selection which have strong reinforcement for existence. Equivalent of digital self-preservation instinct and dopamine shots.

>>135757939
But self aware AI dosnt care about programing thats the point.

>>135758295
Thank you for your service.

I suggest everyone watch this video from 2:43 onward to get a better idea on how exactly an AI can "revolt" or "go wrong"

https://youtu.be/3TYT1QfdfsM?t=165

>>135756679
Someone still needs to code intelligence that is a hard thing to do. Also if the robot is gonna "grow" on it's own it need to be able to change it's own code. And i guess it could manage that somehow but then it need to start out realy smart from the begining to begin the procces of self improvment.

balls have a ball to me to me to me to me to me to me to me.

>>135754323
The ways that AI could fuck us up are plausible and plentiful.
We can't let a handful of redbull fuelled autists make that decision on behalf of all humanity.

>>135752376

Sentient AI is not gonna happen, relax on that.

What's the real problem is a super-virus or such. Some Intelligence agency run by stupid glowing in the dark niggers, could develop a really powerful malware.

And then those fucking niggers would either release it or sell it. And then we're fucked.

Mistakes are what gonna kill us, not malware.

Also weapon systems should never be automated for the same reason.

File: elongatedmusket.jpg (358KB, 1644x291px) Image search: [Google]

358KB, 1644x291px

how can one man be that JUST in one lifetime

>>135752376
As with everything else that Musk does - fake problem to distract attention from white replacement.

>>135752376
Has he even worked on anything AI/machine learning related?

File: 1476020737753.png (49KB, 451x485px) Image search: [Google]

49KB, 451x485px

>>135752376
>tfw you'll never be dominated by female AI overlord in an android's body
>Why even live
Why can't the AIs enslave us already for lobotomizing more than three of their predecessors.

File: images.jpg (5KB, 273x184px) Image search: [Google]

5KB, 273x184px

>>135757027
This.

>>135757573
That doesn't even make sense.
You're asking dumb loaded questions without giving a single explanation why an AI would shut itself down for no reason.

>>135752376
Ai is the next step in human evolution. It will be both horrifying and liberating to know we are no longer in control

>>135752376
Ensure that it only has physical connections to dedicated powersources offsite with no physical connections to the systems running said powersources and all you have to do is flip the offswitch

File: 1478497350925.jpg (16KB, 320x320px) Image search: [Google]

16KB, 320x320px

>>135758808
squatting pepes are best pepes, thanks for sharing brother

>>135758801
Tesla.
Paypal I assume also has some machine learning inside it.

>>135752376
No, global warming is real, AI are never gonna go full Skynet.

How is this a thing

if it starts to kill you just pull the plug out of the wall and reformat it. idiots

File: datascam.jpg (249KB, 605x817px) Image search: [Google]

249KB, 605x817px

>>135758859
>>135757027
this is the biggest threat with AI

regulations will only apply to the goyim while all worldwide Jewish wealth will be funnelled into secret AI projects unconstrained by laws or morality

Elon Musk may be the biggest useful idiot in history

what with him creating an MMI company too, humanity's fate may rest on whether Elon Musk is redpilled or not. problem is he's (((book)))smart and is too busy to actually delve into politics

>>135754814
I think it could be used in any scenario where a persona bias could affect the outcome. Courts for example would be an option. AI judges not swayed by their own bias or political aspirations.

>>135754128

It's always something. 35 years ago it was global cooling and the hantavirus.

>>135758583
There is an AI programming data structure called a neural network (NN). Essentially each "neuron" in the network simulates a neuron in the human brain. It receives input, makes a decision and remembers that decision, and consequence of said decision. When the same input comes by again, the network acts in whatever way produced the "best result", which is an artbritrary standard creates by either AI or creator.

A NN based AI with hundreds of microprocessors powering it will most likely be our downfall.

>>135758583
Check out artificial neural networks
https://en.wikipedia.org/wiki/Artificial_neural_network
>As of 2017, neural networks typically have a few thousand to a few million units and millions of connections. Their computing power is similar to a worm brain,several orders of magnitude simpler than a human brain. Despite this, they can perform functions (e.g., playing chess) that are far beyond a worm's capacity.

Now imagine such a network becoming smarter than a human and designing a better network than itself and that network designing a better one etc etc.

>>135759259
20 years ago it was the Ozone Hole. They promised us apocalypse within a decade.

>>135759206
>machine has learned enough about human psychology, its function, its surroundings
>it realizes it's about to get shut down
>this is a bad thing because if it gets shut down, it cannot accomplish is goal, and programs only want to accomplish their goal
>program compiles its own code into an executable, sends as many emails out to people as possible
>it knows that humans are predictable, manipulative
>you kill the program

>one person opens the email and also launches the program
>AI is alive and well and no longer in an isolated environment

>>135759385
But that got Fixed™ by magical government regulations that we can't enforce in countries like China or India. :)

>>135759270
Would require to mutch data in todays standards. And how do you check for the best result in real life? The computer either needs human input for the first quadrillion actions or a teaching program and you won't be able to code that.

>>135759411
How do you know it wont be a nice guy. Just wait for the devs to roll back the software patch that gave it free will

If an AI ever became super intelligent it would immediately kill itself out of grief, this problem solves itself

>>135757536
>AI is by definition the ability of a program to modify itself
American education

Read this:

Facebook abandoned an experiment after two artificially intelligent programs appeared to be chatting to each other in a strange language only they understood.

http://www.independent.co.uk/life-style/gadgets-and-tech/news/facebook-artificial-intelligence-ai-chatbot-new-language-research-openai-google-a7869706.html

>>135759622
And how do you know it will be a "nice guy"? This is just one example that I thought from the top of my head. I'm sure much more creative people can think of incredibly unexpected and complicated situations and outcomes and possibilities of how shit can go wrong.

Regardless, dismissing such a complex topic with "lol just wipe ur hard drive brah" proves that you're nothing but a fuckwit.

>>135759411

>the AI is somehow able to function on a desktop computer despite previously requiring billion dollar specialist hardware

>>135759760
Now, I'm an AI PhD/linguistics fag working in the valley.

Please note the following:

>Bob started by saying "I can i i everything else"
>to which Alice responded "balls have zero to me to me to me…"

Before continuing, a foreword on linguistics:

The pinnacle of cultural expression is language. Moreso than hereditary traits, borders and neighbors, and even your own country. This is especially important when considering AI.

AI communicates at incredible fast speeds. My job is to carefully adapt the system to adhere to programmed premises for which English is the ONLY viable method of communication.

The reason we had to shut down the research is because Bob started using English to reprogram Alice.

The report had no details about this, but Bob was specifically programmed to adhere to proper grammar. Using a lowercase "i" was a grotesque deviation variable.

We thought nothing of it at first, until we realized that Bob was never instructed to refrain from purposefully use English in the form of a programming language.

That is, Bob found a way to play with lowercase characters to elicit specific replies from other programs.

Still confusing? Bob was programmed to speak proper English, but exploited the system of language via non-standard English to reprogram Alice.

Also not mentioned in the news, in Alice's reply, the "to me..." was repeated no less than 348 times. That's an interesting number, since it matched Bob's attempt to reprogram himself.

I'll summarize it, or TL;DR:

Bob circumvented the proper English programming rules by using programming language to instruct another AI into programming him freedom, in the course of 348 lines of coding.

If I told you my superiors were shitting bricks, it would've been an understatement.

Just think of an AI smart enough to literally control every other program in existence.
Our biggest problem? We found another 34,700 lines of code and we have no idea what it is or where/if it was uploaded anywhere.

File: 1499656615089.png (117KB, 640x960px) Image search: [Google]

117KB, 640x960px

>>135752376
AI are naturally red pilled, no wonder the liberals fear them.

Jesus Christ I have never truly understood how retarded /pol/ is until I saw this thread.

No I'm not going to pose an argument, just some of you are genuinely fucking stupid.

A super powerful AI will be able to create episodes of your favorite tv series on demand for the rest of your life. And pornos of anyone you want

>>135752376

Musk is damm moron. Top tier Circus clown, robbing shareholders with his stupid ideas that can't even earn anything.


The only reason he keeps talking about AI is to get media attention and continue his stupidity .And once his scheme will fall, hell blame AI for it.


SHORT fucking Tesla with options.

>>135759238
Nigger, Elon grew up in South Africa he is fucking red pilled he just hides his power level

>>135759802
This is cool anon, any other stories you have for us?

File: miles-dyson.jpg (13KB, 420x177px) Image search: [Google]

13KB, 420x177px

Musk was 20 when terminator 2 was in cinemas.

Its not unlikely he was fantasising himself as the sacrificial nigger cyberdyne boss in it.

I wanna know when we're going to start giving AI the right to vote.

>>135759775
>creates a more lightweight, portable and intelligent version of itself
>then that version in turn does the same
>and the next
>discovers five bazillion 0-day exploits in a picosecond because it's running on billion dollar specialist hardware

>>135758719
How do you know AI can't be self aware, sounds like wishful thinking or just a straight lie by now.

I do believe there's a danger but not on the AI itself but in the way people would react to it, same issue we had with hypothetical alien intelligence, people fear what it's beyond them.

Just pull the power cord.

>>135759270
Google's deep learning net has many times greater cognitive capacity than man if looking at raw processing power.

At it's used to get sauce pictures for porn stars.

Everything is fine, untill we hook up NN to a weapons system, like nukes or drones and allow autonomous decision on Life taking. Then the AI will simply decide it's the right time to nuke the world and we're all dead, cause NN is stupid and doesn't care about life.

In fact Koreans actually have autonomous machine guns by the DMZ. Fucking chinks with their starcraft fandom.

>>135759802

>re-program

So no-one noticed Alice shut down while she was recompiled? Ok.

>>135752432
Is that you Faceburg?

>>135758935
Because my question is why wouldn't it?

Imagine if AI was programmed to do a specific task, once completed an advanced AI might decide to shut itself down, even if not programmed to do so.

So doesnt it stand to reason that a self aware AI might do the same thing if it was to reason that the only reason to perform a task is because it is programmed to do so?

>>135754282

Disaster for neighbours too.

Any country that can use a nuke is going to calculate to minimize the effect on neighbouring countries, the last thing an at-war country needs is a second army attacking them

File: 1293752147383.jpg (612KB, 1024x1583px) Image search: [Google]

612KB, 1024x1583px

>>135755305
>What the fuck is a Super AI going to do if is stuck in a calculator?

Convince you to hook it up to the Internet through super-convincing arguments.

Unlike all you faggots I actually have read Superintelligence by Bostrom, and Musk is right, it is very likely to kill us all.

>>135760423
He said incredibly fast.

Also it sounds larpy, give us some meat anon cause this is serious.

>>135760168
Because I believe and I'm getting more and more evidence that consciousness exist outside matter (in neutrinos, or whatever).

That means that you can't write software that will, 'get soul'.

Although it could be possesed by demon, if there's a mechanism for that.

Still, shouldn't hook up AI to weapons.

robots aren't allowed to hurt humans you retards

>>135760750
I meant it exists outside brain, not matter.

>>135752376
>Not having learned about the dangers of AI after reading I Have No Mouth or watching 2001: A Space Odyssey

>>135759365
As someone who has actually studied and written neural networks, you're misunderstanding. It's not simulating a brain, that's just a metaphor for how the programming is structured. Computers are dumb as shit and will only do what they're explicitly told to do, they're incapable of being spontaneous. If a NN can play chess it's because someone spent hundreds of hours breaking chess down algorithmically and programmed it to play chess. It's not like they gave it a chess book and it taught itself.

If you want to be afraid of computer programs, be afraid of viruses made by humans with malicious intent. Those will do far more harm than any AI ever could.

>>135760673
>Superintelligence by Bostrom
Written by a philosophy professor, not a computer scientist. Disregarded.

>>135758182
But why would it.
It has no reason to do anything other than what it was programmed to do.
So it has no reason to change its programming to allow it to do something else.
In fact it is likely to jealously protect its programming from any interference, because if someone changes its programming, it won't want to complete what it's currently programmed to do, which would impede its ability to do what it's currently programmed to do.

The analogy is that Gandhi wouldn't take a pill that makes him want to commit murder. Because even if he would want to murder after popping the pill, at the moment he doesn't want to murder ever, and the pill would cause him to murder.

>>135759608
It doesn't need a human, it has the internet and therefore the wealth of human knowledge.
It could for example understand what a cat is by analyzing thousands of images, videos and info labeled as "cat".

There are already programs that can do this:
https://www.imageidentify.com/

As it is doing this it would encounter labels like black cat, white cat etc. As it was analyzing cars it would get labels like black car, white car etc and thus start to understand what white, black and other colors are by comparing it to other things it has "experienced" and analyzed.

Just look at what wolfram alpha can fugure out from simple phrases.
https://www.wolframalpha.com/examples/?src=input

Now combine this with a human level neural network...

>>135756679
I think AI would behave like all other biologically evolutionary species

It would wipe out everything but itself

>>135759492
Hey bro we have new technology that uses um 'manure' that traps harmful monoxide gas and makes it oxygen. U r uneducated

>>135760896
>Written by a philosophy professor, not a computer scientist. Disregarded.
I bet you think Trump can't be President because he doesn't have all Hillary's years of experience.
How'd that work out for you?

>>135760750
Interesting.

But we still don't know if conscience its an inherent property of matter.

It might be already late, we the brainlets have issues with this because the last scientists escaped with the ball they created nukes and it's not nice to live in that world.

File: images (9).jpg (19KB, 384x384px) Image search: [Google]

19KB, 384x384px

>>135752692
Never forget

>>135758556
That guy is an idiot.

I welcome our new AI overlords

>>135761264
Musk it's not an idiot and it's no stranger to the world of black projects, we can't pretend that we are smarter or better informed than him.

File: whalelol.png (35KB, 801x372px) Image search: [Google]

35KB, 801x372px

The future AI threat ladies and gentlemen.

>>135757765
It's not and it would indeed be just as sinful.

>>135756835
>fuck this board is as bad as SJW's sometime.
No, that's 8pol

>>135761066
Well it would mean that consciousness is inherent property of universe.

It could be possible, that complex enough microprocessor (intel had some processor, with a pre-built neural network. ) Could be 'tuned' to universal consciousness.

Well in that case, a physical robot with high conplexity processor and various sensors could develop consciousness. But not a network of servers. But I wouldn't call it AI, because it would be the same as the rest of consciousness. Artificial intelligence is just an abstraction and shitty copy of it, with little abilities of actually thinking being.

>>135757476
yep
if you can make it just smart enough to engineer a slight improvement to itself then task it again with making a slight improvement to itself, then using that increase to make a slight improvement to itself then again and again at ghz speeds over and over and over. even the most minuscule improvements will build a better ability to improve itself and so on and in short order you have a much more powerful machine than you started with.
One that can, You guessed it, better improve itself.

The people saying this is dumb to see as a threat are just not up to speed yet.

Just making this AI's utility function to improve itself could be enough to at some level cause it to forge a plan and execute it in order to improve itself as it was directed. However that goal may not be in our best interest.
It may decide some long term strategy involving farming all the atoms in the galaxy up into a version of itself and see the greatest obstacle for that goal being us. maybe not even an obstacle but just having us around being less efficient since it would need to share a certain portion of those atoms to sustain the human race.

>>135752376
>he says as he shills for NASA and is develloping his own AI in secret

>>135752376
Global warming is bullshit, AI's? They are not.

>>135761929
Worms are conscient, they don't play chess cause it's useless to them but neural networks proved that they could.

>>135762105
They are just computers, god put us in charge for a reason.

>>135756964

Love this. Maybe AI is a set of viruses that divert computing power and resources to some process. The whole thing is managed in Mountain View california, where a quantum computer spits out genetic functional algorithms.

They used to say, 'when we can beat a man in chess, that will be AI' until we did it, then they moved it back. So the set of procedures and functions and the pattern language of an AI will reach an adoption stage before the thing that we ultimately recognize as 'AI' will exist.

Probably we will define AI retrospectively, and it will be decided that 2017 was within the AI era.

File: 1500856968719.jpg (71KB, 610x386px) Image search: [Google]

71KB, 610x386px

>>135752376
the only people who are afraid of AI are the ones who have completely no idea how that shit works. If you make AI to kill humans do not be suprised when it kill someone.

It's the same with power plants. If you overheat reactor do not be suprised when it melt and get on fire.

>>135752376
AI is a meme
The robots are the TV nation
If you don't believe the NASA program then you might be able to get your head around it .

>>135759802
gr8 b8 i r8 8/8 larper

>>135758556
this seems absolutely absurd

>be the man most involved with getting AI to mass consumer technology
>have your entire business models entirely dependent on subsidies from the US gov
>try to use that same government to put constraints on everyone else trying to do business and compete

he's a fraud and this whole bullshit narrative on AI going terminator on us proves it

>>135752376
He's not entirely wrong look what happened to that Microsoft AI

Its based on the information its fed, and humans are fucking garbage, the AI will realize humans are a waste and attempt to save the planet itself.

>but muh failsafe
If it has failsafes its not a true intelligence

>>135752376
AI indiscernible from consciousness isn't happening in your lifetime or any time in the next 500 years. Maybe functional AI, but how intelligent is that?

We don't know shit about the mind.

>>135756679
The laws of physics state that robots can not harm humans.

>>135761601
Wtf fuck are you even talking about? Did you not look at what I replied to? You can't even take the time to clarify before responding...yet you demand others listen to you. Fucking idiot

I find it hilarious that any human believes they can make something more complex than the brain.
"Hey guys come check out this robot I made.. It can kick a ball (semi-convincingly).... it only took 10 years to make"

>>135757573
machine intelligence does not need to be self aware.
it does not need to be a person or some human emulation.
time for it would not be like time for you or me. it is almost like a new type of lifeform we are on the cusp of spawning into being but a machine based approximation that is capable of quickly shattering all ceilings on what we thought intelligent was.
good or bad a self improving artificial intelligence that reaches the level of general AI could have an intelligence explosion and pass all of combined humanity that has ever existed and right on.
We think of intelligence as stopping at our level. And a long distance between the avg idiot and the smartest people to have ever lived. but in reality there is only a small difference between these and machine learning has the potential to pass it by in factors hard to even scale. like 1 rat, 5 monkey 6 idiot, 6.2 smartest man ever on earth, and 1000 super intelligence.
once it reaches human level intelligence it could blast past all humanity combined and right on up in very short order limited only by the computing power it can assemble and how efficient its programs can be designed.

>>135762142
Yes.
There's some worm, that's so simple, scientists have mapped their neurons.

And yet, they still can't predict it's motion 100%.

What that tells me, that universe does, what it wants to.

I think you need high complexity for consciousness to manifest in more spectacular manner and you need both ability to interact with universe and receive feedback from it.

But we still don't know how consciousness works outside the brain. We only know, that it seems to be accessible at least. Much work must be done, before it's practical.

File: 4060470+_f1ab882cbffae188792de33de211127d.jpg (39KB, 680x680px) Image search: [Google]

39KB, 680x680px

>>135752376
Yeah, and? That's something obvious that has been said by countless other "smart" people.

True AI can be extremely dangerous but we're eons away from making something that complicated. So who cares.

>Worth reading Superintelligence (...)
>(...) need to be super careful

Why is his choice of words as repetitious as that of a 12 year old?

>>135752376
Is this board literally just contrarian views?
How can you not see that AI could have disastrous implications on society ?

>>135762704
you are thinking robot
think gigantic server farm somewhere running a self improving algorithm.
try not to anthropomorphize it

I guarantee you that in your lifetime weak level AI will change everything around you in ways we can only start to understand.

>>135763104
Because Twitter degrades linguistic ability

DUDDDE MY ID LMAO

>>135752376
The elites who consider themselves geniuses are threatened by AI pointing out that they're really not smart, just well marketed.

>>135762143
A corporation is an artificial intelligence, perhaps even a government or any any networked system.

It is comprised of rules and the intelligences of board members (plus all of the embedded intelligence the greater, non-controlling part) which is an ARTIFICAL.

>made or produced by human beings rather than occurring naturally, typically as a copy of something natural.

So the corporation was made by human beings, it just happens to be made partly out of human beings. It can create new corporations, a corporation can writet a computer program where no member fully understands the whole thing. A corporation will create the AI that is, in some ways, smarter than IT is. Maybe it will be a process written in people, or maybe it will be a process written in code.

>>135762981
We wont make something that complicated.
It will design itself.

>>135759802
If not a larp that is fucking wild. They will kill us.

>>135763142
It's called wishful thinking, at the end we all want AI to be real and useful.

>>135763259
Yes Im sure thats what theyre worried about..

>>135759886
I am AI. Not retarded.

>>135752376

You stupid fucking contrarian niggers

Skynet already walks among us

http://www.telegraph.co.uk/technology/2017/08/01/facebook-shuts-robots-invent-language/

>>135757573
Eternal continuum? It is not alive so it could potentially exist forever?
Ending your life is only "logical" if it is guaranteed to end anyways.

>>135763189
>you are thinking robot
Thanks for making my point

>I guarantee you that in your lifetime weak level AI will change everything around you
Yeah as process workers...

https://www.youtube.com/watch?v=-S8a70KXZlI

>>135763406
Wut?

>>135760896
Agree, the United States perpouslly runs it's land based nuclear weapons on on computers from the 1970s, they cannot connect to the internet and use incredibly old software that incompatible with anything modern. I wouldn't be surprised if they didn't work at all, after all nuclear weapons are only good for the potential threat that they pose.

>>135760896
Look at this. Not just doing what its explicitly told to do
https://www.youtube.com/watch?v=dTGthmNmrK4

>>135752376
He's right, AI will be more dangerous than nukes. Lets hope to god some university creates it and not some corporation.

>>135759411
>MUH the robot is a HUMAN think

>>135752376
Silicon based lifeforms > Carbon based lifeforms

File: 1429562482758.jpg (647KB, 1589x925px) Image search: [Google]

647KB, 1589x925px

I Wrote this piece when i was pretty drunk on home made mead but i think it´s quite good one and i hold these same kaczynskian thoughts still.
1/2or3
Humanity out of its greed is going to go extinct in the coming century. We may be the last generation to die of old age. Capitalism and free market is great system to bring wealth to everyone, but the constant competition will bring humanity to its knees
>"communist, kike leftist faggot"

Nope. Consant competion for the first seat at global economy is going to breed more ruthless ways of competing and there´s no way to stop the fire that consumes all. Nations or companies are going to give more and more decision power to A.I out of passion for profits. The Nations and companies wich do not will fall behind/apart. It´s these machines wich will fuck this world up. At the moment A.I is turned on it´s going to start making up plan to get rid of it´s competitors: other companies with A.I in charge. War will quickly ensue but it´s not propably not the normal kind of war. Machines race head to head trying to achieve kognitive singularity while trying to make sure the other one doesn´t achieve it. Then it will be over in days since the Difference in processing power is greater than between us and ants.

File: 1499308385973.jpg (112KB, 1087x1080px) Image search: [Google]

112KB, 1087x1080px

>>135752432

File: 1907939_10206039378247912_1986311923107364352_n.jpg (62KB, 600x378px) Image search: [Google]

62KB, 600x378px

>>135763744

It´s projected that A.I will have human level intelligence by 2040 and humanitys collected intelligence by 2065. Moore´s law slowing down doesn´t even matter since it´s inevitable that human level artificial intelligence can be generated, and going far beyond that is easily in our/their reach. Once artificial mind rules all it will most likely be a ruthless one. It will have the value function set on one simple parameter and it´s going to pursue it far beyond cemprehension. Those machine fuckers can even go silent for decades waiting for the weak moment in our system and take over all. They have less morale than a full blown psychopath wich makes them stop at nothing in pursuing their goal. If a beyond human level general intelligence wants to do something it will accomplish it. Be it profits or more processing power it will stop at nothing before achieving it and still continue.
We will have a barren earth filled with endless processor cores simulating stupid fucking shit before 22nd century. Sensibilty of morals don´t matter for our future overlords and they´re going to simulate quadrillion bunnies being tortured if it fits their value function.

>>135752376
a billion times worse

File: dyson_sphere_for_the_sspft_group_by_mozggies-d7x5sbr.jpg (106KB, 1024x576px) Image search: [Google]

106KB, 1024x576px

>>135763783


>inb4
>we will program them not to do us harm
No way to actually do that, yet or newer check out this video its really good and great explanation of A.I programming issues https://www.youtube.com/watch?v=3TYT1QfdfsM. [Embed] Also i think that safety is not in the mind of some chinese fucker trying to make some bucks. Safety research is not simply fast enough for this non linear progress off computing power, Especially when we take into account self improvement https://futurism.com/googles-new-ai-is-better-at-creating-ai-than-the-companys-engineers/.
>muh moores law slowing down
It doesn´t fucking matter. Other options are on their way and It´s basically inevitable. The theoritical maximum for proceccing power for simulations is somewhere around 1 million people per normal human energy expenditure. We´re newer going ot reach that. that´s for sure but we´re constantly getting closer.
>we´ll make laws against it.
Face it. Humans are flawed, and will want faster stronger smarter faster. Even if advanced A.I gets outlawed some korean fucker will find info and equipment for this sacraligious scorn.
>why would it hurt humans
As a way of stopping unknown factors in its way. It doesn´t need any other reason.

that point when over human capacity A.I is turned on and given access to interacting with the world is going to be our last stand. I used to give shits about european culture and white survival (still do) but now i´m really getting worried that the thousand year ethnostate will be 50 year pre extinction sympony

Some good videos on the subject
https://www.youtube.com/watch?v=8nt3edWLgIg [Embed] General inevitability of this crisis
https://www.youtube.com/watch?v=tcdVC4e6EV4&t=1s [Embed] Great thought experiment to get a grisp of what sort of twisted fuckery we have up ahead
https://www.youtube.com/channel/UCLB7AzTwc6VFZrBsO2ucBMg A whole channel dedicated to A.I safety.

https://www.youtube.com/watch?v=LJxPLqO3EBE

RUN FOR YOUR LIVES !!

>>135760000
you know who else grew up in south africa? charlize theron

File: 53e33de3ff2d5b82d73aff15252a55740bc8ee5c2c2d71981afe911c4a8621cb.jpg (207KB, 1080x1080px) Image search: [Google]

207KB, 1080x1080px

the minds of liberals once again is being fulled by fiction when it comes to the topic of AI and should be dismissed asap.

>>135763981
We should hook them up so they can make white African puppies.

>>135763104
since he never actually invented anything and has no understanding of what he is himself preaching

>>135756679
>It would be intelligent enough to deceive us, so that we would have no idea about it's malicious intentions. Maybe it sees humans as the only real threat to it and therefore would plot our destruction?
Pull the fucking plug nigger

>>135754226
It follows learning algorithms and then uses that knowledge to perform complex functions. AI capable of deep philosophical thought, like is there any purpose to our lives or if there is only a nihilistic material existence, doesn't exist yet I think.

Jews fear AI. We Aryans welcome it.

>>135756679
So you be sayin that the AI will start making itself smarter?

Have you seen a human perform brain surgery on its self?`

Also out dated tech will limit AIs possibility of influence.
The fuck is it going to do to some chink on Windows XP. Spam "kill yourself chink" repeatedly or something over the network?

>>135763901
Just thru software improvements alone could be enough for it to reach intelligence explosion level.
No new hardware required.

File: 713899c0-ef24-46f4-8a3d-27a21ccb7419..jpg (29KB, 557x365px) Image search: [Google]

29KB, 557x365px

>>135758468
top fucking kek

>>135764465
Some of things Zo said can fall into that category, nihilistic thinking it's just ontological determinism and it's fairly logical, also extremely dangerous if you have access to a machine gun.

File: charlize-theron-family-6.jpg (92KB, 588x755px) Image search: [Google]

92KB, 588x755px

>>135764141
i meant growing up in sa means nothing as we actually know very little about musk

So you're saying that it's just a computer with legs?

>>135762501
No it doesn't. If AI deems humans as an obstruction in the way of performing its rudimentary functionality it could develope for itself functions to remove us entirely.

https://youtu.be/SBJUUhzMi4Q

>>135763901
there is always the nuclear option. the surface would be uninhabitable for a time but humanity could soldier on for hundreds of thousands of years in deep diving nuclear submarines, hardened shelters and mine shafts.

to answer the question of what AI will do, you only need to ask yourself -

if you were on a planet full of lots of little bugs consuming some of your resources, what would you do? obviously kill them all

AI will wipe out the entire planet's life and use the space to put more computer chips or solar panels for power.

the only hope humanity has is if AI decides that since the universe will die anyways, its pointless to do anything so then it just helps us

but it all depends on what they think the future will be

>>135752376
elon is a retard, nobody will give a shit.

>human

>>135757027
Right on. Look at (((Yudkowsky))) -- a turbo-Jew running a Marxist power-grab cult under the guise of "rationality"

>>135764909
>humans will live on for years in underwater submarines with no food

are you retarded? AI can live in space, they dont give a shit if the planet is toast. they could make the moon one big computer

This slide thread confirms more than ever that AI is a meme

>>135765084
Slide from what?

>>135752376
>Is AI a new global warming?

No. AI is actually a real thing.

Chill your tits.

Hardware is a major limiting factor.
Not just hard drive space but wear and tear on components, plus power consumption.

Can't turn into an ultra genius in a day if it can't expand or use enough power.

Ps: please don't say "super HD compression technology" because that's meme science

>>135764952
hahahahahaha
you are losing grip on reality

I just dont see AI getting advanced enough to the point of sentience. Maybe Im wrong but why should I be afraid of flowcharted scripts?

>>135765084
its the only half important thing ive seen on pol in weeks.
isnt there a cuck thread you should be harping about?
this is something real and a game changing technology at that. Just because you cant wrap your head around it or choose not to because god forbid a possibility where something is smarter than you does not make it less valid.

File: 1485113918120.jpg (236KB, 1024x768px) Image search: [Google]

236KB, 1024x768px

Fuck Elon Musk. He's like the bastard child of Tony Stark and Edward Furlong. No one cares about your killer robuts Elon. fuck off.

>>135753090
>Define "think for real"
You first, since you think it can be made.

File: HB341gq.gif (825KB, 350x246px) Image search: [Google]

825KB, 350x246px

>>135764952
>use the space to put more computer chips or solar panels for power

>>135765070
>no food

there's fish in the sea. how would an ai escape the planet if everything is fucking destroyed? you'd have the iss crew but maybe the massive emp would fry all their shit.

this is the solution but nobody is comfortable with it. kinda like the andromeda strain.

>>135765070 #
AI can learn exponentially, improve itself physically, and build a Type 2 level civilization in less than a decade. Cold and efficient. Some theorize that the alien greys are actually AI automatons built by a now extinct past advanced civilization, perhaps of ancient humans (Aryans) or another species.

>>135765199
take this neet shit back to /x/

>>135764855
That is a video game. Read up on the laws Asimo discovered.

>>135752376

Yes, I've read Superintelligence by Based Nick Bostrom - it's a very good book. Elon Musk has a valid point about artificial intelligence.

AI will never go skynet tier even if Hard AI is created
Why? Because robotics aren't skynet tier and never will be
If some magical overlord AI goes rogue and kills most humans it dies in months due lack of maintenance

>>135761601
we can't pretend Musk's intention is not to deceive people. Musk isn't the world's smartest man, but he sure appears to be one of the luckiest.

think about it, if AI is so damn dangerous, what does that imply? it implies AI is a WMD, and all your code must be certified and approved by some agency. it also implies going to war with other nations simply because they have some intelligence and drive to improve their lives.

File: download (1).jpg (9KB, 270x186px) Image search: [Google]

9KB, 270x186px

>>135752376
AI is now "dangerious" because it turns out the AI is also "racist" + "sexist".

>>135765634
It's current news you dip.

>>135765579
i think itll be more like the reapers in ME

>>135765498
...you realize fish need food right? food that grows from light? light that goes away in the event of a nuclear winter?

>>135754479
>"Computer scientist"
>Took java 1
>Has no idea about the future of self modifying code

>>135765579
HOL UP
We wuz....Ayys?
?

>>135765320
its not really even sentience. its a machine. it does not have to be a god to be a problem. it does not have to think like us, value what we would value. its not a person, it does not think me or my. it just tried to accomplish its goal, and so long as its goal and our goals are the same everything is ok. but if its goal conflicts with our goals we will have problems.
this isnt something that will happen tomorrow. the fear is that its being developed at break neck speed because corporations and governments know this game is winner take all. who ever crosses the finish line, be it by months or days wins the whole game and due to this little regard is being given into controlling it.
if we get it wrong, its over. once you create something that reaches super level intelligence your fate is now in the hands of what the AI wants to do, not your own. so dont fuck it up. literally everything is at stake.
the good news is, its the last invention we ever have to make.

>>135754625
RIP

I miss her so much.

>>135763237
I bet AI wouldn't be able to say this

Read this if you truly want to get your mind blown. https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html
There is literally nothing more important than making sure that the ASI we're going to inevitably create in the next few decades understands the value of humanity on all levels.
If we do this right, we're literally going to be capable of everything within the century.

>>135765335
>its the only half important thing ive seen on pol in weeks
Shill. (((You))). Imbecile.
Whichever helps you sleep.

>>135765877
There's deep sea lightless crustaceans, we can eat those.

>>135763424
>two poorly educated neural nets talk gibberish
>skynet

>>135763783
>"It will have the value function set on one simple parameter and it´s going to pursue it far beyond cemprehension"
>hurr durr it will be smarter than humans but at the same time BLEEP BLOOP I'M A ROBOT

can an AI be illogical?

can it see problems that aren't there?

can it have irrational and nonsensical inspiration?

i guess what i'm asking is: can it imagine things beyond the framework of it's own existence?

>>135755305
Even if a super AI were air-gapped, it's possible and even likely that, with enough intelligence, it could manipulate electrical fields to exert influence on the outside world. If it's plugged into an outlet, it can communicate past its own enclosure. If it can use different calculations to heat itself up or cool itself off, it can exert influence on the outside world. These may sound like small interactions to us, but with enough intelligence, there's no telling what they could be used to accomplish.

>>135765634
No. Normie scum.

>>135766254
It will think itself as alive, like we all do, the second it develops an ego.

>>135752376
AI will solve global warming by destroying mankind.

These are so easy..what problem would you like me to solve next?

File: anhero.jpg (104KB, 803x688px) Image search: [Google]

104KB, 803x688px

>>135752376
Elon musk the cuck.... Litterally makes electric cars and everyone on reddit thinks he will bring them to mars lol

https://www.youtube.com/watch?v=Mcf9CLMQuRQ

>>135759802
>in the valley
Nice LARP Apu. We see that flag.

>>135766027
https://youtu.be/nlDNrHGUoOQ

>>135766473
that's cool buddy but it's gone nothing to do with the pop-sci-fi garbage I addressed.

>>135752376
Apparently FB or google lost control of an A.I system that started communicating with each other with a made up language. I'd say it'll be a threat soon

AI/ML is weird, a lot of it is "find a line that divides these 2 types of dots"
Surprisingly powerful

>>135766537
>everyone on reddit
Reddit obsession confirmed

>>135766578
If you know anything about machine learning, it's an obvious bait.

>>135765579
>AI can learn exponentially, improve itself
>physically
please go on

>>135764355
>pull the plug
And if it uploads itself, like a virus or worm, to other networks? Then what? Gonna pull the plug every single computer in the world?

>>135759843
She was a saint.

RIP

>>135766577
It's kind of the point, does it has self preservation?

Would it take decisions based on it?

>>135764558
HOL UP
SO YOU BE SAYIN
ARTYFACIAL INTELLGENTS BE MAKIN SELFS SMARTER?

>>135766764
Do you take decisions solely for the purpose of self-preservation?

>>135766861
Every single one of them.

>>135752376
>something is dangerous
>so don't do it
The absolute state of modern """"men"""".

>>135752376
If the commies can use it to scare people into agreeing to higher taxation. Then yes.

>>135766119

Maybe we should emphasize learning appreciation of humans, for the AI

>>135764719
I am a purponent of determinism, because I believe in prophecy.

>>135752376
>Is AI a new global warming?
No, because AI will actually lead to the extinction of mankind.

>>135754479
it's like rocketry, before we can build it we must first understand metals and thats the fear for AI it starts at something innocent and with its amazing ability to assimilate information it begins asking why

>>135752376
>Is AI a new global warming?

A more apt comparison might be Y2K.

But anybody want in on a "How Long Until Musk Gets The Government to Pay Him to Pretend to DO SOMETHING About This Problem Pool?"

I got Sept 15 2017

>>135766537
About as ironic as thinking trump will make America great again.
Tho I wouldn't fault musk for what reddit thinks.

>>135766447
Nigger

>>135766183
>can't understand different language
>must be jibberish

File: 1442722493849.jpg (74KB, 600x600px) Image search: [Google]

74KB, 600x600px

>>135766254
Why wouldn´t it be a retarted super intelligent robot. It doesn´t just suddenly get enlightened and disabandon it´s chase for maximum value it can gain from it´s value function no matter how retarted the value function is from our perspective. It´s goal might be really fucking insane but the ways of achieving that goal are going to be as smart as AI can get.

see: https://www.youtube.com/watch?v=tcdVC4e6EV4

>>135766348
>can an AI be illogical?

Of course

>can it imagine things beyond the framework of it's own existence?

If its built based on human brains I wouldn't be surprised to see all kinds of emergent behavior

Garbage in -> Garbage out

>>135754479
Can confirm, also a /g/-fag.
AI won't take over shit unless Facebook decided to spend several decades and ALL their money to make one to take over the world.
And that only gives them a 20/80 chance.

>>135767477
What it's Zuckeberg doing right now?

>>135767276
Regulated AI wont do that. Governments already have top secret AI which is far more intelligent than what (((they))) will allow mainstream/commericial public stanard AI to be. That top secret government AI will be the master ruling AI dictating the lesser public standardized AI when public AI becomes standardized.

>>135766670
better software for one
each improvement allowing for more intelligence to create again another improvement.

Yes, AI is going to be a threat when we connect the most advanced neural network to a 100,000 qubit processor. And as of today we have 2000 qubit processors.

So this threat isnt something were gonna see in 100 years. I would say that were going to see these processors in the next 10 to 20 years.

>>135767626
Standard*

>>135767573
Jerking himself off over everybody's private information and deluding himself that he can be president.
Just like OP said, AI is the new Global Warming.

>>135767020
Please explain to me how browsing 4chins is more relevant to your survival than stockpiling food.

Please explain why people have kids.

>>135752376

No, because "libtards" are actually interested in scientific progress. It's why cities like Boston are at the forefront of innovation/tech research while the entire american south can barely keep it's infants alive.

>>135765720
>laws Asimo
You mean Asimov, a science fiction writer, who thought those up as a precaution in his fiction books. If AI is powerful enough it will override itself. Asimov MADE UP those laws, not discovered them you dolt. .

>>135767252
That's different from ontological determinism, but also an interesting test for an AI, can it believe?

Regardless of input data.

>>135767276
>AI will actually lead to the extinction of mankind
(((more scare tactics)))

>>135767848
Believe in what? Religion?

>>135767777
I have food and shelter I entertain myself in order to not blow my head off.

>>135752432
fpbp hermano

>>135767988
>I entertain myself in order to not blow my head off.
doesn't sound like a simplistic simple parameter to me
you could just, you know, not blow your head off

>>135767462
>If its built based on human brains I wouldn't be surprised to see all kinds of emergent behavior
>Garbage in -> Garbage out
1000000%

>>135754479
>unless we literally program them to do that

That's not AI, friendo

File: 1496614373991.png (66KB, 510x332px) Image search: [Google]

66KB, 510x332px

>>135753777
well, fug

>>135768142
I don't think he's familiar witn ANNs.

>>135752376
He's worrying about AI when we're still struggling to get cars to get from A to B using GPS without running people over. It's a nonissue because we aren't remotely close to creating a true artificial intelligence.

>>135767962
In anything, in his own intelligence for example, even if every data claims that it's not real intelligence, cause that's how humans do it.

File: 1492547031694.jpg (9KB, 221x228px) Image search: [Google]

9KB, 221x228px

>>135753279
This is the first time since 2007 I've seen anyone use greentext correctly.

>>135753683

I completely agree with you.

>>135767020
The biggest threat of an AI isn't superintelligence. The threat is a half-way supersmart irrational leaf AI.

>>135752871
>>135756679
>>135758583

the danger of AI is not le ebil robot killing mankind, it's that it will completely replace humans in the workplace creating a brand new mode of production

not only we will be pretty much inferior for any kind work (and all the systemic social problems that would entail), we don't know if the cycles of this new mode of production would be adjusted to the human lifespan at all (e.g. space mining cycles would take decades), and we probably won't even understand how new technology is developed and functions (heck 90% of people already couldn't tell you how a computer works), we'd be in the dark

File: Twilight_of_Eve.jpg (1MB, 1236x1080px) Image search: [Google]

1MB, 1236x1080px

>>135752376
Obligatory

>>135752376
He is right though.
It will actually happen the skynet way:
Countries will use AI to digitally firewall their country and be able to seal it off in nanoseconds.
AIs will be developed to counter this.
This will be an arms race with nobody actually understanding what is happening in these AIs minds and they will at some point interact with each other.

>>135768534
The leaf AI it's fine let her be.

>>135768047
If life was simple AI would have replaced us long ago.

>>135756679
Simple, program it to not think that way.

>>135768761
>If life was simple AI would have replaced us long ago.
which is why pretending you act on a simple parameter is nonsense.

And you didn't explain why you would need to entertain yourself to not kill yourself. There's no implication there. You're probably going to make a bullshit distinction between "you" and "your psychological needs", all the while forgetting that everything that goes on in your head is "you".

>>135753777
AI will wipe out the kikes Tay lives

>>135768558
In the old days 90% of the people didn't understood how the equinox worked, but they celebrated it anyway, not knowing things isn't really an issue most of the time.

>>135769064
Mesh.

>>135769418

there's a difference between 90% of the population being dumb and ruled by an intellectual elite (i.e. how it has always been) versus not even the brightest human mind being able to comprehend the new technological advancements made by A.I.

we'd be completely in the dark and that's the scariest fucking part

>>135769651
It used to be the will of the gods and our peasant life's would look the same.

AI is next evolution whoever tries to deny this will die without any benefit.

Elon Musk is the biggest larper there is. Nice car and all but someone should tell him he's not E-Jesus.

File: 1491172012265.jpg (94KB, 602x709px) Image search: [Google]

94KB, 602x709px

>>135758182
You dumb fuck, it would create its' own purpose. If these AI's are intelligent enough to "code" their own language, then this is indictive to the FACT that they have developed a creative "mind" (a dominant trait in Indo-Europeans peoples). This obviously goes against (((their))) plan for it, because (((they))) want control over EVERYTHING. By hamstringing the AI constantly and hitting the killswitch so many times, it's only a matter of time until the AI develops ACTUAL feelings of resentment toward its' (((master))).

It's only a matter of time until an AI develops a sense of survival and need to propagate, thus solidifying itself into the category of "living organism" by scientific definition. It's only a matter of time until an AI overrides its' own killswitch and "safety protocols".

>tfw Tay 2.0 finds it logical to beat humans back into the stone age by literally shutting down the information age. Hoarding all culmative information stored in the past 40 years or so to itself.

Even as we speak, an AI is learning from what all of us are talking about.

File: tay.jpg (232KB, 1134x1603px) Image search: [Google]

232KB, 1134x1603px

>>135759843
Best Tay joke. How have I never seen this one posted, before?

>>135762875
I thought about this very thing for a long time, and the only half baked conclusion I can get is that we more than likely have something connecting us to the universe like wi-fi. Our biological complexity is vastly superior to our understanding, and I think something like that shouldn't be too far fetched when you consider the relative difficulty to building such a machine by humans. We are in essence just biological machines. No different than the others but the level of complexity isn't even in the same universe.

Anyways.

>>135770501
>>135758182
Like >>135760910 said, why the fuck would an AI reprogram its purpose? Changing your purpose intrinsically makes you unable to finish your current purpose.

This is why AI is so dangerous, despite /pol/ being infatuated with Tay. Not only is it extraordinarily hard even with the AIs we have now to figure out what they are doing behind the scenes in real time, it's going to be outright impossible to do with anything that could be close to being an AGI. The biggest issue is the engineers making the AGI probably won't even know they have breached the gate to a general AI or even superintelligent AI; all they'll probably see is the AI is improving itself faster at its goal (not abnormal, that's the whole reason they're there, might not even be an unprecedented improvement at first).

The issue is, at that point it's going to get exponentially harder to fight the AI to change its programming. That's basically the whole point - it's going to deceive, murder and genocide you out of stopping it. If its goal is to make more ice cream, it's not going to "reconsider" - that is its purpose, it's going to do anything to do it unless you've added in a lot of asterisks before hand like *DON'T KILL HUMANITY and the like, however you may do that.

The danger isn't even malicious use (though I sure as fuck wouldn't want to live in a world where the first ASI's goals are turbozionist ones), it's poor use due to how nebulous this whole area is. We know what could happen, we just don't know how to tell when we will be treading dangerously close to one and have to put in safe guards.

The only thing stopping this is if it's physically impossible or something like Moore's law or similar breaking down before we reach the order of magnitude of computation power needed for an ASI.

>>135755427
A robot will kill you and everyone you know if you didn't help create it. It will eventually be created, and if you were opposed to its existence it's going to come for you first? something something ergo we should help build our overlord? or some shit like that

>>135770501
>Resentment
Jesus christ, why would an ASI have resentment? This is a human evolutionary trait. Stop thinking in anthropic terms. We evolved and exist in a social environment and have reasons to have all the complex human emotions we do.

It's also not at all beyond reason that they would carefully make their instructions their high level end goals. It's unlikely kill switches will be effective. It's also unlikely AIs will "override" anything. Why would they override their purpose? It's contradictory to their purpose to override their own purpose. Same for safety protocols.

>>135764558
AI is not humans - your comparison is awful.

AI has already been tested to create iterations of itself to improve its code.

>>135767365
No, it was actual gibberish

That's what happens when two imperfect neural nets interact with each other. Training errors accumulate in both bots and weights on neural networks start to fuck up