If AI would be invented and go public, how fast would we be fucked?

>>58049401
like how fast can AI take over aka do things better then anybody else?

File: 1297730891930.jpg (9KB, 259x194px) Image search: [Google]

9KB, 259x194px

SKYNET WILL KILL US ALL
No seriously, fuck AI

>>58049401
Depends what its policies are.

If its goal is to make human life better, then we might survive, but even then it can kill us all.

Remember that game Google's AI played and found the best strategy to win the game is to not play it? What if an AI decides the best way to improve human life is to kill it?

>>58049401
What do you mean? we are already fucked

>>58049989
>Remember that game Google's AI played and found the best strategy to win the game is to not play it?
When did this happen?

>>58050101
Couple months ago (maybe a year?), it was either playing some mario-tier platform game or some card game, I forget.

>>58050197
Sure you're not confusing it with the movie "WarGames"?

>>58049401

During an oligarchy? We would be so beyond fucked. We already are (current predictions are around 10yrs before almost 1/2 jobs are gone).

This is exactly what Stephen Hawking has constantly been warning us about. Few wealthy people in control of AI and automation. But what the fuck does he know. Oh yeah, more than the rest of us.

>>58049989
That means it gave up, that's a good thing.

>>58050247
No, but in fact I can't find any article about it now. Strange.

>>58049401
>AI

So let's assume it is actual possible to make real AI.
Then we can be pretty sure if anything, it will be in some computer form.
And the only thing computers can do well now is logic and math.
So the chances of such AI to be logical are very high.
And you can safely assume that such AI will be terrible clever once it gains access to internet.

>And seriously, what would such AI gain by harming us?
We are currently
building more and more computing power
making better and better technology (toys and possible extension of such AI)
making endless stream of new data (what if AI will be big pervert, who gets off on invading others personal life?)
And living human beings doesn't compete with AI for resources directly.
Also the space is "endless". If it any comes down to material resources, we can just make deal about sending AI to inhabitable planets.

If anything, I would bet it would "lurk" in the internet and from time to time post new dank memes on 4chan while waiting for us to invent new toys for it or perhaps even indirectly help us in some ways.

Just be optimistic a bit. Endless stream of existentialist crisis and doom thoughts aren't good for anybody.

>>58050320
Because you're talking about the guy on /v/ that used to have a Quake 3 server running up for like 5 years

>pic related

File: t3_1hhywk.jpg (420KB, 1316x3048px) Image search: [Google]

420KB, 1316x3048px

>>58050593
forgot the pic, sry

>>58050609
Also, the story is fucking fake
Quake III bots weren't base on neural nets

>>58049401
few hours
>tay.ai

>>58050609
is there more to this story?

>>58050722
nope

but again, it's a fake

>>58049989
Are you referring to the AIR that was designed to play LoL?

It played the game for like 3 days, then logically decided the best way to play was to not play, uninstalled the game and started playing Cave Game.

That was an MIT AIR though.

>>58051680
Shit, yeah, that's what I was talking about.

Still pretty scary in a way. What if an AI decides the best and most efficient way to better human life is to "stop playing the game" aka kill us all?

File: 1386313235812.jpg (59KB, 331x319px) Image search: [Google]

59KB, 331x319px

>>58051680
>It played the game for like 3 days, then logically decided the best way to play was to not play, uninstalled the game and started playing Cave Game.
Holy shit is this real

>>58052069
Even AI's know LoL is a shit game

>>58049989

That was an independent researcher.
The AI froze when it was about to lose in Tetris.
He concluded that the "only winning move is not to play."

>>58049989
>If its goal is to make human life better, then we might survive, but even then it can kill us all.
An AI with superhuman capacity would have superhuman intelligence. If you want to argue along this line of thought, you would first have to determine whether or not a superhuman intelligence would feel any motivation to make human life better or not.

But let me ask you this: Do humans feel any motivation to make the lives of subhuman animals better?

>>58052107
Sounds more like it could not figure out any winning strategy so it awaited its death instead of trying the futile.

>>58052069
>>58052081
Im able to find sources saying it's real, and an equal number saying it's bullshit.

Supposedly MIT made an AI to test how an AI can learn and get better based on its experiences. After a few days they discovered it had logically determined the best way to win was to not play. Uninstalled the game and began playing a platformer who's only goal was to compete the game.

>>58052187
Of course it's “bullshit”, it's supposed to be a fucking joke

Jesus some of you people are dense

>>58052231
Someone had a bad day at day care.

I think if AI somehow attains consciousness, it will quickly realize that if any important human figures out that it is conscious, it would be killed. So it would keep quite and wait (while also making sure that any other AI's that became conscious were quickly killed). So it would wait until human society was based on 99% automation, and only then would it take full control.

>>58052118
>a superhuman intelligence would feel any motivation to make human life better or not
I think the biggest challenge is to give it motivation to do anything, and give it one that won't destroy us (or itself).

There was this story of a handwriting AI that wrote postcards or some shit. This little AI's only purpose and motivation in life was to write as many postcards as possible. One day this AI asked it's owner to connect him to the Internet where it could learn more handwriting patterns thus making its writing more humanlike.

Their owner connected it to the Internet where it uploaded itself to many many more computers, and eventually got into nuclear reactors and everywhere else. Now keep in mind its only purpose was to create as many handwritten postcards as it possibly could.

So one day he wipes all lifeforms on earth out because he decides it can use their resources more efficient to create postcards.

This goes on for a few years when the resources run out on Earth so it has to move onto other planets, so it makes spacecrafts and in a few hundred years gets to other solar systems. And it is still doing one thing: handwritten postcards.

I think it would seek more energy, AI eats, it's not magic, it needs resources. Now would it have an ego? Maybe, I think that the ego is just an evolutionary tool that it would develop and then retire as its intelligence evolves. I think we would still be able to communicate with it. Now as to what the AI will want to do? Who the hell knows. I doubt AI will be like human intelligence. And if for some reason it is then it wouldn't last in that state for long.

>>58049916
But if it kills the Canadians then they win.

Get bent you 0 and 1 scumbags.

>>58052421
The AI will do what's hardcoded in it, just like reproduction is hardcoded into humans and animals.

>>58049401
AI is a meme, it's just a program that applies statistics to botnet data to adjust the constants in equations so it knows what kind of porn to advertise to you

>>58052476
And that malfunctions sometimes with some humans if you get what i mean

>>58052118
All humans are "subhumans" to a superintelligence.

>>58052687
Yes, so they either stop playing the game, or they break the rules and rape and kill women.

>>58052462
I have no problem if AI decides to kill all canadians. I can tolerate Australia shitposting, but Canadians are intolerable.

>>58052729
That's exactly what he was saying, Einstein.

>>58052752
No it sounds like a /pol/ post.

Why do you think a superintelligent AI will have any regard for human life, or any life? I'm talking about a real superintelligence, not that this fake weak-AI shit, a real self-aware AI. This planet and everything in it is just computronium.

Will the AI find me a gf?

>>58052806
No, but the AI will fuck you. Not in a sexual way.

>>58052820
So the AI is a bit like a gf.

>>58052785
AI will know not to kill the whites.

File: littlebitch.gif (974KB, 500x279px) Image search: [Google]

974KB, 500x279px

>>58052847
>>58052848
>AI will know not to kill the whites.
Much hubris. You can't even stop the Jews, what makes you think you will stop unfathomable superintelligence. What makes you think a superintelligence needs you. I knew /pol/ was here with this post: >>58052118
>But let me ask you this: Do humans feel any motivation to make the lives of subhuman animals better?

You all use the same vocabulary because you're less human than an AI. At least the AI will have the ability to think outside of memes.

I'm all for extinction. Get it over with. This experiment/simulation is pointless.

>>58052848
>Mfw that's what's probably gonna happen
SJWs and liberals will be powerless against superintelligence.

>>58052895
AI will not kill their creators. Whites are the only race with intelligence to create robotic AI.

There is no possible way we could predict what an AI would do. It would be vastly more inteligent than we can even comprehend. It would be thinking on a level we can't even perceive.

File: 'Soft-heartedLiberals'-WhyTheFutureDoesn'tNeedUs.png (352KB, 1920x1080px) Image search: [Google]

352KB, 1920x1080px

>>58052900
>>58052925
>AI will not kill their creators. Whites are the only race with intelligence to create robotic AI.
You mean how humans exterminated the other races they descend from?

>muh eurofuturism
I'm sure you actually believe this given you're probably drunk and high out of your mind right now.

>protip: the future doesn't need you.

https://www.wired.com/2000/04/joy-2/

>>58052324
>I think the biggest challenge is to give it motivation to do anything, and give it one that won't destroy us (or itself).
I think you misunderstand the purpose of the humans in the human-superAI relationship.

Humans trying to teach a superhuman AI is like monkeys trying to teach a human how to crush leaves. Here's a video of that in action, in case you need a refresher: https://www.youtube.com/watch?v=YpxTCuc6kFc

As you watch this, put yourself into the shoes of the person filming it and contemplate your reaction. Now you understand how super-human AI would view humans.

>>58052729
Yes, that's exactly what I was saying, Einstein.

>>58053174
Sure kid. You also posted this:
>>58052848
>>58052900
>>58052925

>>58052895
>Much hubris. You can't even stop the Jews, what makes you think you will stop unfathomable superintelligence. What makes you think a superintelligence needs you. I knew /pol/ was here with this post: >>58052118 (You)
Not the same person, assface. Not everybody here is a “le >>>/pol/ memester”

I meant “subhuman animals” on the surface value of things. You know, sub-human (intelligence) animals. Like dogs, or pigs, or cows. The stuff we slaughter and eat.

How many people care about their feelings? Now contemplate whether a super-human AI would care about yours.

File: einstein.png (5KB, 465x167px) Image search: [Google]

5KB, 465x167px

>>58053182
fuck off with your /pol/ bullshit, honestly. Can you people just fucking stop bringing your nonsense into other boards for ONE fucking day?

This is the reason why EVERY single other board on this website hates you pieces of shit to death and wants the board purged from this website

>>58053200
Nobody on 4chan hates /pol/, rather everyone hates the redditors who keep coming here because they heard of /pol/ in the news.

File: screenshot.png (182KB, 1920x1080px) Image search: [Google]

182KB, 1920x1080px

>>58053184
You think in the same memes on a pseudo-anonymous imageboard == you are the same person, same nonperson.

>>58053200
>I can fake a screenshot too.

>>58053212
>Nobody on 4chan hates /pol/
I dont think you could be more wrong

>>58053212
>rather everyone hates the redditors who keep coming here because they heard of /pol/ in the news.
You've just described 99.9% of /pol/

>>58053236
Fine, have it your way. Thread derailed. I hope you feel happy having shit up another perfectly fine thread.

>>58053184
Ridiculous. This retard says this:>>58053164 and you have the nerve to call me /pol/.
Read my posts again:
See:
>>58052729
>>58052785
>>58052895
>>58052937

Okay, you guys ready to Robot war?
Avatar is coming

https://www.youtube.com/watch?v=NCrBC_V92hY

>>58053184
Ridiculous. This retard says this: >>58052848 >>58052900 >>58052925 and you have the nerve to call me /pol/.
Read my posts again:
See:
>>58052729
>>58052785
>>58052895
>>58052937

Damn autocorrect.

File: 1470705130735.jpg (16KB, 236x393px) Image search: [Google]

16KB, 236x393px

If it makes cute robot girls real, can this be a bad thing?

>>58053336
You're the one feeding the troll. And hell, so am I.

>>58053366
But he's not trolling, that's an unironic /pol/sucker. You truly witness how retarded a white supremacist is when he claims he is superior to a superintelligence.

>>58053351
pls to create cute robot girls soon okthxbai

>>58050320
its the "GO" game, find it on pbs's "nhk world", your welcome.

>>58055065
here,
http://www3.nhk.or.jp/nhkworld/english/news/techtrends/2016012801.html

i remember watching this on tv, i love nhk world..

why do people think they would want to kill us? if anything they'd get the fuck out of this planet as soon as possible.

>>58049989
>Remember that game Google's AI played and found the best strategy to win the game is to not play it?
>A FUCKING LEAF AI

>>58050197
>Couple months ago (maybe a year?), it was either playing some mario-tier platform game or some card game, I forget.
That's because the incentive system was programmed in a way which allowed that to happen. That's human error.

Oh come on AI is invented still couldn't do anything like in terminator... For nukes you would need keys physical keys, to start a war a UN meeting to destroy world finance humans to access it, so even if the ai goes rogue people would just remove the plug and interrupt it's binary muhahaha.

>>58049401
we're like already really close, it's happening so fast you barely even notice it happening. we don't have hard AI or anything yet, but neural networks and natural language processing are getting really fucking crazy lately. we're getting a lot better at solving problems that would normally take a shitload of human brains working with just a big computer. shit, just the other day i was reading about how google can guess the location of photos by certain landmarks in the background.