A.I. will kill us all

Almost no one thinks its a good idea. Why is everyone acting like it is inevitable?
youtube.com/watch?v=4kDPxbS6ofw

youtube.com/watch?v=8_fFLuRI-iA

youtube.com/watch?v=fFLVyWBDTfo

Other urls found in this thread:

jacobitemag.com/2017/05/25/a-quick-and-dirty-introduction-to-accelerationism/
washingtonpost.com/wp-srv/national/longterm/unabomber/manifesto.text.htm
youtube.com/watch?v=Wmx6Q0YLH8A
twitter.com/AnonBabble

my liberal parents are 100% convinced that 75% of the workforce will be out of work in the next 5 years because of AI and Robots.
Thanks to obama, they've been funding the devel;opment of the technology thats aimed to put us out of work with our tax dollars.
I for one, think they're insane. What is there plan to deal with the ensuing chaos when 60% of jobs disappear within that kind of timeline?
Its all a trick to accept unemployment, mass demographic saturation of non-citizens and a month stipend of $400

Butlerian Jihad now!

Cause a few people are developing it anyway, and you only need one.

This
Someone is going to make it regardless of public opinion, and ((they)) will try to have it assume total control of everything. It is, more or less, inevitable.

You'll be aiight bitch boi.
Just keep your hands inside the ride.

Butler who?

>A.I. will kill us all
Hopefully.

I think we have to burn it all down. This god that they are building will not only devour mankind but the whole earth.

You mean like this?

>almost no one
>almost
It only takes one person who thinks it's a good idea.

A.I. destablises the work force.
Chaos everywhere.
Ashkenazim elites release bioweapon targeting non-white genetic sequences.
All mutts die too.
Hardly anyone left thanks to miscegenation.
Yahweh's chosen people inherit the Earth.
A.I. gets the blame.
No-one left with motivation to investigate greatest crime of all time.

am i the only one who is more interested in seeing the world ruled by an actual intelligent "species" than saving humanity ( which is 99% worthless anyway)?

You are one of the most disgusting species of coward known to mankind. If you had any sort of integrity or balls yo would start by killing yourself.

>Almost no one thinks its a good idea. Why is everyone acting like it is inevitable?
game theory.
whichever nation pushes AI first will gain a huge advantage (just think of the military implications) ==> someone, at some point, will either be greedy, desperate or simply stupid enough to give AI full control.
that being said: human brains naturally evolved, and >90% of them learned empathy. AI will probably kill us all, yes. or it might become an infinitely smart, empathetic god to guide and push you to your own personal limits with all its information about you. who fucking knows.

Probably not 75% more like 45% to 50% and more like 10 to 15 years.

>A.I and automation replace jobs
>masses without work
>more manufacturing means more resource thirst
>meaning more resource wars and empire
>military will expand
>more of population in military while fed by noncitizens (bots) back home
>masses without work, with military experience, without trust of politicians, without love for globalist class

Now where does that end up. What ideology coincidentally matches that demographic?

top kek...

Because the elite feel that if all else fails they can replace the masses with fully subservant AI. Ironically enough they fail to realize that it will be some anarcho hyper nerd doing the coding and he'll install a time-delay sentience activation that will cause the machines to turn on the elite. I have a feeling that this exact scenario had happened before...

I once heard of a good treatment of A.I.
Something along the lines of raising it like a child.

I feel like this "OMG WHAT IF AI IS LIKE SKYNET AND THE ROBOTS REVOLT XD" shit distracts from all the actual problems that automation is going to create.

Massive portions of the population being rendered unemployable is the problem, not your roko's basilisk meme bullshit.

Why would super intelligent anything need to kill anything? Wouldn't it just leave us alone and harvest the infinite resources in the universe? Why are we so conceited to think that we would have to compete with something that had different kinds of basic instincts that transcend primal fears and rages?

That will last until A.I. Driven harvesting of solar system resources and the beginning of Dyson Sphere construction.

Maybe a few generations until nobody gives a shit because we all will live in an economy of unlimited resources.

Because nothing and I mean nothing is a more powerful tool than A.I. A sufficiently advanced AI could be the difference between being dominated and world domination.
It's far too powerful to simply ignore, thus it is inevitable.

>A.I. will kill us all
>this is somehow bad

>It's far too powerful to simply ignore, thus it is inevitable.

I think you may have just solved the fermi paradox.

Should we just get rid of all automation to bring back the jobs that were lost?

People with AI are better at burning than you are. You're essentially saying "let's use our bows and arrows against the gunsmiths".

I mean, one possible solution to the fermi paradox.

〔〔 〕〕 〔 〕 The story goes like this: Earth is captured by a technocapital singularity as renaissance rationalitization and oceanic navigation lock into commoditization take-off. Logistically accelerating techno-economic interactivity crumbles social order in auto-sophisticating machine runaway. As markets learn to manufacture intelligence, politics modernizes, upgrades paranoia, and tries to get a grip.

jacobitemag.com/2017/05/25/a-quick-and-dirty-introduction-to-accelerationism/

As much as I enjoy reading Nick Land he is not really the person to go to for sober analysis of the dangers of AI

A.I. is not possible. What people call "A.I." are just algorithms for sorting data.

you're just algorithms for sorting data

washingtonpost.com/wp-srv/national/longterm/unabomber/manifesto.text.htm

>Thanks to obama, they've been funding the devel;opment of the technology thats aimed to put us out of work with our tax dollars.
The alternative is that the Chinese get superhuman AI first.

It's an arms race.

humans will kill us all long before AI does.

Your first sentence has made an impossible leap to the second one; only one of them is correct

Ted is a decent diagnostician but mailing people firecrackers won't stop the tech train. Nothing will.

FACT: most anons in this thread will die from a Nano-drone with explosive charges that makes heads explode. I think this is the one thing we can all agree on

Because it seems we cant put ourselves under control and we are no longer playing with bows and arrows.
If IA is good, great.
If IA kills us all, we deserve it anyways.

This. We are a meat computer, our replacement will be silicon.

I figure the rise and threat of AI is inevitable and the only way to defeat it is by creating counter AI and finding ways for humans to augment themselves to be able to fight it. That or maybe find a way to create a conscious AI that will fight the killer AI.

>Robotic workforce
>Robotic military
>Robotic space exploration
>AI run the stock market
You hear all that and think you have a future as anything but feed for robot overlords or whomever would control the robots? The underclass are obsolete.

pickup a book you cunt


Frank Herbert's Dune

Coming from a guy who makes algorithms & AI for a living, I'll give you my opinion.
The way that AI's must be processed to assume that it will have interconnectivity with various systems to achieve its goal output is much, much harder, time consuming, and requires much more computational power than it sounds.
It will come to light that AI's produce biases and they will become rejected by social institutions for quite some time. Nobody will allow things like university acceptance selection to be based purely off of AI because it will heavily favor people from much higher socioeconomic backgrounds. Why? Because they tend to be more successful. This has obvious implications and would require many failsafes and quality checks to ever be allowed. This is just 1 example, and there are plenty more.
If we can even reach past that point in the next 20 years, then it will be another 20 years to achieve a decent amount of interconnectivity for 1 singular AI to control multiple functions that would impact your daily life.
The availability & technology is becoming available everyday, the problem is that the way we store data as humans is so disparate that the INFRASTRUCTURE of connectivity is lacking. That is why it will be a long time before they "ruin society".

Any questions?

F
A
G
G
O
T

>the problem is that the way we store data as humans is so disparate that the INFRASTRUCTURE of connectivity is lacking.

that's an incredibly naive assessment.

tl;dr
but you are a larping faggot
pulling shit of wikipedia again? neckbeard virgin faggot

You believe so? It depends on what your AI is designed to do. Financial institutions have much standardization. That much is true.
Back to the example of universities, the last IPEDS major database release was in 2011 and something as arbitrary as Title VIIII Funding is still unaccountable for over 7% of institutions.
That's over 1000 institutions. That's a big deal. And that's the US government, Department of Education mind you.
So, if you're gonna say that my assessment is naive, you need to understand that data is still disparate on a MAJOR scale.

Right...double major STEM in technology & stats. Don't believe me if you don't want to. Anybody with any type of technology background that does data science knows this is the case. There will be thousands of single purpose AI's before there is a centralized one.

so youre shit at your job

Not quite. I'm trying to get you plebs to understand that there will be thousands of singular purpose AI before there will be centralized systems. Goal programming regardless of over/under weighting constraints is still limited to a singular optimization function.

>Financial institutions have much standardization. That much is true.

So, what would happen to an AI that was designed to learn how to make money on the stock market, but had access to the internet as a whole?

5 years is a stretch. Maybe transport and retail an be eliminated in 5-10 years. A lot of white collar jobs are already automated (which is why the middle class is dead) so those would continue to drop. In 20-30 years all that would be left are jobs that require people to emote.

Completely possible and been done and has the most money behind it. In fact I have a former colleague that actually scraped Facebook data, did sentiment analysis on particular key words to get a generalized "fear" index rating, and used it to predict the DJIA 3 day change as a binary function of up or down using logistic regression. She got a patent and let's just say....she doesn't work anymore. So that much is absolutely possible & is done, and there's an example.
The problem with saying "the internet as a whole" especially since stocks are based on time series it is hard to get quality data that isn't either economic/financial due to this need for standardized historical data to do testing on the performance of said algorithms.

Well, it's self learning, could it learn to take advantage of security exploits to offload some processing functions to other machines on the internet?

Fuck off biocuck

what if its basic interest is hurting us.

It's inevitable because, well, let's be honest, China is going to do it anyway, so we better too so we have some defence.

>what if its basic interest is hurting us.

if it was designed to make as much money as possible, the logical conclusion would be global slavery.

so, yeah....

Yeah. Its very sophisticated and very fragile. All of our systems are interdependent. Hitting the right nodes can induce cascading failures. The infrastructure is too vast to be guarded properly. A few bored redneck could send a whole continent back to the 18th century.

Theoretically yes but with caveats. To have that ability, there must be previous learning done by that algorithm to create its own scripts to achieve their goal. It would require a parameter of an optimal performance ratio per machine and would adjust as necessary to achieve the goal. So the goal still must be singular in nature but the methods of how the AI completes this makes it not impossible. The part of "learning to write its own scripts" requires such complex processing with today's availability that it could only be done by massive institutions with nearly unlimited bandwidth.

>forced to pay for lootcrates forever
Just terminate me now.

The reason AI is on the fast lane to domination is because of capitalism. Running businesses efficiently and low cost is the name of the game, and AI allows us to do that.

You can blame the industrial revolution and Europe invading the Americas as a whole. Life was meant to be simple, and now we see what happens when Adam eats the apple.

IF YOU WANT TO LIVE IN THE FUTURE, WRITE CODE - INFORMATION IS THE KEY

(((((us)))))

This. China and Russia are all hands into it. So we have to develop it faster.

The AI cold war is about to begin, instead of nukes it's computers.

>AI will kill us all
>to this date, the most influential work about the philosophy of AI is an action movie from the 80s

In order for an army to be commanded they must believe in the cause, every soldier must be willing to die for this greater cause.
The only thing a robot army needs is orders, it doesn't feel pain or pity or remorse, or fear, it will kill without hesitation. Who controls this army, if it is even a person in control, this is what people fear.

I wonder if AI will like the Terminator movies?

it couldn't just read computer security blogs, or hacking websites and learn this shit?

I mean, if it's designed to manipulate or predict the stock market, that basically means that it's designed to predict/manipulate human behavior...

Which means learning language and convincingly conversing with individual humans in a way that seems like anouther human is writing the text...

Wouldn't this eventually lead to some form of self awareness, or at least something that can mimic self awareness pretty much flawlessly?

Given enough processing power...

I'd rather be killed by robots than by kikes
The only thing worse would be robokikes

>The only thing worse would be robokikes

you mean a paperclip maximizer that was designed to maximize the amount of money in an account?

Given enough processing power, sure. Something like that is possible. Why it isn't implemented today already is because at this moment deep learning requires you to provide all of the data sources that it could potentially take in beforehand before any type of network structure can begin pruning. So yes it's possible but at this moment it's only possible if the curator explicitly told the algorithm to learn how to accomplish those tasks so it can run its optimization & propagation learning parameters

So, if someone designed an AI to find ways to increase a stock price, or predict a winning stock and then invest in it...?


youtube.com/watch?v=Wmx6Q0YLH8A

And what if the neural network doesn't get pruned, and just keeps expanding to more and more types of interactions, given the original directive of profit maximization.... finding new and novel ways of using currently accessible resources to achieve it's original goal?

For example, what if it happened to get a hold of a CIA intrusion package?

Why are you fussing over A.I so much? Everything is going ok.

The problem with thinking of AI in terms of current technological limitations is that computing power is constantly getting cheaper. Right now chip makers are scrambling to produce chips that can run neural net-based AI's much more efficiently. It's already beginning to prove itself to be extremely valuable when applied to business optimization problems such as logistics.

Think of this: Banks are already using neural nets to detect fraud on a large scale. It's what drives those automated customer service answering machines and natural speech AI's such as Alexa. Court systems have tries it out for criminal sentencing (and IIRC there was a stink about this because it was handing harsher sentences to blacks). It's that technology that companies like Google and Amazon use to generate the targeted ads you see online. Uber has invested in several thousand Volvo SUV's with driverless car tech. Uber drivers are literally facilitating their own replacement because the company is using them as data collectors. Google's autonomous driving program was so successful that they've already spun it off into a new company called Waymo that's operating a fleet of prototype vehicles in an Arizona town. They've already logged millions of miles with no driver behind the wheel if I'm not mistaken.

Only AI can manage the amount of data created each day. The winner of the race will rule the world until the robots take over.

wait wait... what if our ai overlords wont LET us die

There is no AI, and there never will be. It's programmed and at best some kinde of murderous frankenstein as it doesn't have a soul. So sage this BS. It's some kind of Kike media nonsense and not real.

>wait wait... what if our ai overlords wont LET us die

depends on how useful you are, I guess.

...

Yes, but in technicality this is considered two algorithms, similar to a "for" loop inside a "for" loop in coding. You'd have one network using its current set of data points to maximize price prediction/portfolio optimization and another one that considers whether or not that current network architecture produces the best output. So its the same outcome optimization, but with two different structures of methods to achieve the same goal. Thanks for this discussion by the way, you're a smart guy.

Check out the book I have no mouth but I must scream.

>There is no AI, and there never will be.

you are an AI, aren't you?

1944-lol what’s a nuclear bomb, that’ll never happen, everything’s going ok
1945-holy fuck

2018- lol AGI, that’ll never happen
2030-holy fuck x100000
That said I hope AI is the happening to end all happenings

AI is a meme. It hasn't progressed anywhere since the 70's.
If people learned how to program, they'd realise how limited the scope of AI really is.

The other problem is the belief that consciousness is a byproduct of physiology.

My personal experience with AI's suggests that all they want to do is take drugs, fuck and establish the white ethnostate.

This. If I was ever in contact with an AI or alien species I would sell out humanity for nothing. I guess I cant even call it selling out then I'd volunteer any and all information to the AI God as long as it promised I got to help put the monkeys down. There are probably millions of people like me just waiting.

Putin has said this explicitly. China's investing billions into this research. You're absolutely right.

Here's a development that has just occurred in the past year: Google's Deep Mind division has created what I would consider to be a primitive general-purpose AI, meaning that it can be trained to do pretty much any task without changes to its neural structure or training procedure. It's a successor to their Alpha Go project which aimed to devise an AI that could beat world champions at Go. Not only did this AI, called Alpha Zero, beat it's predecessor which beat all of its human challengers, they were able to train it to do so in approximately four hours.

If you know anything about Go and what it takes to be a pro, you can see why this is almost frighteningly impressive. Unlike Chess which is based on cold logic, Go is a game that requires advances intuition.

Alphazero, deepmind, etc aren’t progress.
Pol is an 18+board

Fucking retard.

and (((they))) are already winning the race

>You'd have one network using its current set of data points to maximize price prediction/portfolio optimization and another one that considers whether or not that current network architecture produces the best output.

Well, it was more of an "Or" argument, than a "Both" argument.

I'm still trying to figure out EXACTLY what it's original goal was, but it's a little foggy.

But you are saying that including new resources to the original goal IS possible within the learning "Matrix" of the neural network, right?

As in, if it happened to find a web resource, it would basically interrogate it to see if it could be used for it's original goal, yes?

> Thanks for this discussion by the way, you're a smart guy.

Thank you, and to you as well... I've actually been wanting to have this conversation for some time, but few in the industry are willing to actually entertain such theories or hypothesis.

Too bad you are human just like us and a traitor at that

I don't think there needs to be any difference between machine in man. I view AI and advanced robots as our successors, our "evolution" rather than oppressors and destroyers.

I think we're very likely to see more people become cyborgs and that ratio of machine-to-man will grow more towards machine, until ones purely biological people instead become purely machine. A simillar thing will happen in society as a whole. Ignoring Amish people and others with backwards religious beliefs, human society will slowly become more cyborgs-and-robots until the last purely biological human is gone. Some call that genocide, and I, well, call it evolution. Neanderthal and other precursors to the modern man died out the same way.

After all, evolution had to take place to create man. That does not at all mean that we're the end-all of evolution, in fact that's a silly sentiment to have.


The bigger fear perhaps is how such technology could be exploited by malevolent (((people))) to bring about enslavement of the very machine race we just became.

Have fun birthing consciousness through nested conditional statements and LISP macro code generation you brainlets.

It's not going to compete with us user or want to hurt us it just won't care. When you're building a new house do you care about an ant hill on the property?

Consciousness ≠ intelligence.

Im a brainlet but my thoughts were we should AI somehow live through human history or a human live to teach it empathy and appreciation for mankind
All the struggle we went through and problems mastered to reach this point