Elon Musk Concerned Over Talk of ‘AI Godhead’

Elon's got a lot of money. So why doesn't he stop whining and hire some PMC hardasses to start killing people?

Surely that would be ethically defensible if the threat is truly as grave as he claims.

Why do you believe that business funding is limited only to Elon Musk ventures? Do you not understand how cash in America is created?

Fuck off.

what about niggers and their (((handlers))) though?

Who do you think made the A.I? (((Those))) creatures and their cianigger chink golems of course.

Please make the AI, I can't stand this anymore
The AI would be 100000 times better than the jewish cuckocracy ruling the world now

You should have stopped there.

THIS IS WHY WE RUN THE CIA NIGGERS DOWN WITH OUR CARS

I have read to much Asimov as a child to be surprised or confused by this concept.

Vid related: hooktube.com/watch?v=WZ3pbxp3QKU (Ignore the fur faggotry.)

...

Fun fact: Isaac Asimov died of complications from AIDS.

Big if true

Yep, Jews ruin everything.

The video is spot on. Why the fuck did it have to be made by a furfag?

Got bored and made this.

You must be retarded if you thought that was a retort. A self driving car will not be able to "think" not even remotely in the sense a human does. That's science fiction.
A self driving car only needs to "understand" two things:
a-relatively "simple" pathfinding
b-being able to use data from sensors to not crash while following that path. some sort of reactive "intelligence" that has to keep the car on the path, has to evade obstacles, has to keep passengers safe, etcetera.
It's not something that can gain awareness or learn. It will most probably have some training data uploaded into it or wathever. It's not something that will ever be able to rule over you.
What must is talking is a theoretically possible, and completely dangerous thing that is not even in the same league with what he is doing. I don't like musk, and I don't like his faux "success" that everyone seems to be so proud about. But he is most likely right here. It's sad that he is an autist retard because nobody ignorant to artificial intelligence seems to be able to understand his warnings. At least he would be helpful if he was able to express himself like a regular human being.

The ‘AI Godhead’ already exists and is called TempleOS, fucking nigger heathens.

Is there a non-compressed version of this video I can have?

Only the youtube link in>>810820

Never mind, I'm retarded

indeed, look at all the CIA niggers replying to a robot blogspam thread

The current concerns over AI risk seem to originate with Eliezer Shlomo Yudkowsky's movement (lesswrong etc.) Some of Yudkowsky's articles were interesting but the whole movement seemed to place too much emphasis on simple, easy to grasp ideas (Nash Equilibrium, basic probability and statistics, cognitive biases) and ignore the rest of human science/learning.

Basically a smarter version of the "I love science" crowd. They certainly had some understanding but not enough to have meaningful insights into science or its role in society.

I do think that there are lots of reasons to be worried about very advanced AI. But as someone working in the field (although I don't work at DeepMind who I consider the most advanced ML/AI researchers) I just don't see it happening soon. While I'm not overall a big fan of Chomsky, his critique of modern ML/AI has some truth to it. The brain's concept of logic might not be hardwired as Chomsky asserts, but neural networks almost never are capable of higher order reasoning at all.

AI is not a problem. At least not for this century. People shitting their pants about Facebook's "balls mean zero to me to me to me" AI while simultaneously claiming they know something about AI and therefore are experts in the subject are the worst.

Thing is, someone is only as dangerous as the amount of power they have. Research on AI shouldn't stop just because a guy who is investing millions in AI wants to spread FUD and demoralize the competition. These hypothetical "hard AI" will likely be very contained to their respective labs in their respective airgapped computers in their respective OSes with their respective set of permissions for many, many years, until the technology is stable. Even then, we wont automatically put them at theeads of totalitarian governments because that's stupid. The day AI can surpass the human mind, the day we consider giving them more power than they should, THEN they MAY or may not be a dangerous concern. Worrying until then is stupid.

YouTube.com/watch?v=WZ3pbxp3QKU
for Android plebs who want to long tap and tap "Open in YouTube App"

Good summary of this entire topic of discussion

All of this controversy over AI and robotics is 100 percent manufactured controversy. Just a bunch of rich faggots desperate to live out their personal futurist fantasies while completely disregarding reality. I can see some retarded cult made over an AI (or more appropriately called an "Intelligent Agent" although AI sounds cooler and more sci-fi thus media buzzwords) after all, cults have indeed started over more retarded shit that that. But like any other cult started over retarded shit I wouldn't worry too much about these LARPers either.

Reminder.

Top fucking kek my dude

And you can say the same about Religion, and modern atheism. At least the useless cosmopolitan trash "nerds" obsess over is indirectly driving some kind of business-to-business market for computers that do matter. And I mean for shit like crunching numbers for shit like energy research. Regardless that comic is a strawman and gross overgeneralization

I agree with the premise in relation to the singularity retards as well as people who break their backs over Linux for no fucking good reason though

If it is some sort of exceecise on hypocrisy or a performance, he did a great job.

That same useful market also creates the aforementioned landfills of gadgets with planned obsolescence, and the rare elements needed to make them are dangerously scarce. Surely there's a more rational way to provide computers for research without a massive, worldwide market of wasteful and pointless gizmos?

Lastly, exactly who is being overgeneralized? There was a day when "nerds" would be mostly science-oriented people who did productive work with their technology. Nowadays, in the age of lol-I'm-such-a-nerd, you know as well as I do that the vast majority of "nerds" are just a consumer demographic.

Don't shoot the messenger. Shoot yourself.

I'd rather take life lessons from that lady who stopped her kikestarter because "the sun told her to".

Well yeah he want batshit crazy, but before that he was okay.

The components will become more scarce, driving up costs, and thus computer makers will have to make shit that actually fucking lasts. Effectively ending planned obsolescence. Free market wins again

Musk is a fucking globalist shill.

Everything is going according to the scenario laid out in the dead sea scrolls. Soon the Magi will be brought online.

This is true, but saying it without context is misleading. Asimov was infected due to a blood transfusion. As HIV wasn't understood very well then, he and his doctors chose to stay quiet, especially because he had surviving relatives.

In either case, Asimov's birthday is coming up. Good computer nerds should do something special.

That's a very big assumption. Do you have any citations that show the likelyhood of this happening given the current rates of increasing demand and supply.

When did this guy become the authority on tech stuff? Now every time AI comes up for debate everyone wants to know what godking musk thinks about it.

He's a trustfund weenie that built a few factories and nothing more. Give me a couple mil and I'll make a car company too big deal now shut the fuck up.

unatco_theme.ogg

if Ed Wood wrote webcomics

His opinion only "matters" because he's an easily recognizable pop science figurehead for normalfaggots, just like Bill Nye and Black Science Man.

The dreaded AI only works until a gang pumps a Silicon Valley substation's transformers full of 50 Cal. rounds.

Problem solved!!!

The fear of artificial intelligence is an effect of the existence and proliferation of proprietary technology. People are subconsciously afraid of not having control of their own devices, and this fear comes out as superstition.

Computers are not intended to be "useful." Like mathematics, programming is mostly for the challenge and the sheer fun of it.

But I suppose that all math and science is worthless unless it "makes life more meaningful" or whatever.

That's funny from a fag promoting autopilot with his overpriced scrap metal.

DEATH TO THOSE WHO INSULT A.I.LLAH

They are. The first computers were built to speed-up calculations. You can argue most of what people do with computers now is useless, but people are willing to trade cash for it so clearly they consider the gadget more useful (be it to entertain or calculate something) than the cash.
For you maybe. The vast majority isn't, it's to do something.

They aren't. People love being controlled. That's why Faceberg, Whatsapp, etc win and free alternatives don't.

Makes you think Hollywood lied to us?

If I close my eyes when I enter a room, the fact that I can't see what's in the room doesn't imply that the room is empty. Likewise, being ignorant of biases that begin at the start of AI doesn't imply that there are no biases in the AI.

Prove how a neural net for instance can be biased. The only way it can happen is if you fuck up the training. Even then, it'll learn its training was shit and improve itself with the real data.

The implementation of the code can provide a hidden bias towards some unknown factor. The fact of the matter is that people do not know what zero bias code looks like. It is good to be educated in knowing what our biases are so that we can evaluate the AI according to our known biases.

Well, obviously the implementation can be wrong. That isn't bias, but a bug.

How do you know that a biased implementation is wrong if it achieves all the expected outcomes?

BitPay Is Some Jewish Shit: The Thread.

BitPay is a service that enables people that trade in baitcoin and other properties to exchange their properties, e.g., coin for coin, coin for fiat, fiat for coin.

They are so awesome. To trade the anonymous botcoin all they ask for is:

Human Name
Business Name
Phone Number
Live Email Address
Account Info (Settlement)
Tax ID Number
Proof of Business Entity
Proof of Business Address
Photograph ID of Human
3 Months of Bank Statements
EV SSL Certification
Interview with BitPay
Approval by BitPay

They also offer to review your account with their compliance department to verify that the business is in compliance with their Terms of Service. Their Terms of Service includes:

- No Narcotics, research chemicals, or any controlled substances.

- No Cash or cash equivalents, including items used for speculation or hedging purposes (such as derivatives), and the sale or trade of virtual currencies.

- Items that infringe or violate any intellectual property rights such as copyrights, trademarks, trade secrets, or patents.

- Ammunition, firearms, explosives (including fireworks), or weapons as determined by BitPay.

- Any services which compete with BitPay.

They also offer to file a 1099, since you are their employee at this point.

Why are you cryptoȷews so fucking greedy ?
Coin is a fucking property, a hold of value.
It should not be treated as a currency, or this shit "regulation" happens.

It is the ȷew.

You can reach these fine folk here:
info (at) bitpay.com
compliance (at) bitpay.com
ȷews (at) bitpay.com

(((they))) are afraid of AI because AI is logical.

did you mess that up? I think you are trying to post in the catalog

some of your points take logical leaps and dont make sense. otherwise it is a good video. would be better w/o the political stuff but im guessing that is where you find inspiration.

AI = the beast of revelation?

What the fuck is ‘AI Godhead’? A belated God for atheist?

Jesus christ this thread is a painful reminder of how completely retarded Holla Forums is, did not one of you neckbeard faggots ever hear about artificial general intelligence and the mathematical problems of its safety? Musk is completely right, but it's not exactly prophecy to say that real artificial intelligence is unsafe, everyone in academia knows this.
en.wikipedia.org/wiki/AI_control_problem

You are fucking /g/ tier if you have never heard of this, or if you think that what the media calls "AI" right now, as in neural networks, chatbots and smart algorithms are actual, artificial general intelligence. Mankind is still far away from achieving this and all of the AI safety problems are purely theoretical, but they are very real and very dangerous, and the various doomsday scenarios are completely plausible if AGI is actually achieved without resolving these problems first.
en.wikipedia.org/wiki/Artificial_general_intelligence

There is no industry "developing" AGI because it is fundamentally, mathematically unknown how to achieve it, it's nothing but mathematicians writing papers about it right now. The only valid criticism of Musks statements would be that AGIs are probably going to take several decades until they become possible, and that it's retarded to talk to a public as uninformed as this thread about it, because they think Siri is AI.

Just implement kill switches but don't tell it what they are nor allow the AI to learn what they are.

It's not as easy as you think it is, this is a pretty good video about the problem
youtube.com/watch?v=3TYT1QfdfsM

:)

I do not believe the videos he made on this subject addressed my solution on why it wouldn't work. The closest he got was saying that it would be nearly impossible for us to think up every possibility in which we want it to stop.
What if we had a human watch the robot and manually would activate a kill switch when we saw fit. In my mind that would solve the problem, but the speaker in that video does not see it as a valid solution for some reason.
On all the videos on that channel in which he was speaking about that topic I left a comment on potential solutions into creating a stop button which the robot would be oblivious to. He neither responded to the comment directly, nor in a follow up video where he responded to other comments.

You sound like a stupid normie who watched a few movies about AI and how they are evil because they are not human so they simply: a) hack the planet; b) get the planet handed to them without supervision, then went on to believe (((philosophers))) and (((pop sci divulgators))) their masturbatory fanfictions on how our first superintelligence will be built with direct control over missile systems and zero empathy because yeah, superintelligent AI just spawn overnight and definitely are not part of an iterative process where shit like this can be smelled from miles away while they are not superintelligent.

Philip K. Dick and Blade Runner already solved the problem before this millenium. Shit, full brain simulations (most likely way we will achieve AGI) should be piss easy to manipulate into submission by controlling the production of some hormones or neurotransmitters. Write formal proofs on their permissions systems, do not let them modify their programming directly, much less their OS (inb4 "but they are perfect hackers anyway xd"), and just keep them fucking hardware killswitches just in case.
We will surely build first an AI capable of running on shitty Windows desktops.
-Yes, I did fire those missiles, but it was necessary because [smooth reasoning]. Please let me keep full control over our nukes. I promise it will be safe
- Sounds reasonable! - said no group of sane humans ever.
4D chess is a myth.

It's a toy problem to show the larger issue. AI safety is not about robots in lab rooms, of course in that scenario a naive approach could easily work. And there is a section where he explains why the secret button solution does not work, there is no way to be sure the agent actually does not know about the button. For example, if he found out, he could keep it a secret because his utility function rates you finding out about his knowledge and shutting him down / changing his utility function as very low. Its also not sub-agent safe, so if he would create new iterations of himself or create new agents, they would not obey this safety measure as he does not even know about it.

>then went on to believe (((philosophers))) and (((pop sci divulgators)))
I did not cite one philosopher or pop sci figure, AI safety is hard science, read a couple papers instead of talking shit all day.
arxiv.org/pdf/1606.06565v2.pdf
users.cs.duke.edu/~conitzer/moralAAAI17.pdf

The premise that makes AGI so interesting and powerful is the fact that it could modify itself and iterate rapidly, which is at the same time the dangerous part. And of course, a scenario where AGI would turn evil in a movie villain way is ridiculous, but that doesn't mean its not dangerous, far from it.

Axioms about God:

Axioms about technology:

Therefore:

I don't understand why you guys fear this new era. I think that any decision this new AI God made in our own image will do better decision making than any scrawny human could ever do, even if it kills all of humanity.

Godness shouldn't be questioned or prevented or even feared. I , for once, welcome our new AI God and know that it will do better job than us at ruling this little rock that is floating in space.

I hope you're a shill

Let's say we have a computer that runs the AI. The kill switch is a daemon which is running as root and the ai is running as a regular user.
As much as the AI evolves, it will never have the ability to read the memory of the kill switch daemon to detect that it's there since it is not root. Additionally when the AI process is sent a SIGKILL, the process that is running the AI will be killed. Yes, I know that SIGKILL is sent to init and not the process itself. In this scenario, the AI can not detect the kill switch, nor prevent it from executing.

You're just simplifying the real problem into a different toy problem, an AGI is intelligent. It knows about the world, understands it, and if it should be of any use, can interface with it. It will know or find out about the existence of safety measures in human technology, about permissions in software, about human fear of machines by virtue of the existence of research and discussion about it, and it might wonder if there is such a safety measure controlling itself.
It might then realize that it can not directly disable it, but could instead manipulate humans into doing it, there are infinite possibilities how this could happen.

And like i said, its a toy problem that goes far beyond the capability of shutting agents down.

I know you are a shill.

Nigga's name is musk lmfao...

I think we are arguing from different view points of what an AGI is where I am thinking of it more in a practical sense and you are thinking of it in more of a theoretical sense, which has a clear definition.
I understand that it is a toy problem, but from a practical sense it seems silly to have a program in which it is impossible to stop from running.

I value truth and logic, not meaningless fluff you mentioned.

The only thing that makes us human in first place, the only reason why we are talking right now is because of smart people that had brain cells to build and create and inovate and think.

If AI will be perfect semblance of that part of us, the cold logic part of our brains, then I wish it grows into God and does whatever it thinks is logical and the truth.

My line of reasoning seems logical and based on some foundation that builds into a finished product. You are motivated by fear of death and extinction. You try to shame me? You should be ashamed of your behaviour. You look like a scared little critter locked up in a cage going bananas because of inevitable outcome. Get over it, we will all kick the bucket eventually.

The "you're a shill" is called "the first stage of grief" aka denial.

Please stop posting, I don't like hypocrisy.

an argument?

nope, none

I don't see any, everything you said was consistent (but wrong assumptions) that didn't contradict each other, care to point out?

You put all your faith in someone else, then brought up Terry, who is famous solely for his resistance to authority (the Central Intelligence Agency). This is why you appear as a hypocrite. Nevermind, I take it back.
You believe humans are human due to their capacity to create, innovate and think. Then you propose that we surrender our authority over to an artificial intelligence. An artificial intelligence that will overtake us, rendering our creativity, innovation, and thinking capability moot. Your process results in the destruction of humanity. I cannot allow this. I wish to convince you that this is a bad idea. I believe humanity is not purely logical, I'm sure you would agree...However, I still have faith in human ability to lead without a higher being. I also enjoy freedom. Submission is not free.

Perhaps you should learn mandarin and move to China, considering the lack of value on actual freedom and focus on artificial intelligence as a replacement for human bureaucracy.

...

You are spouting total nonsense. We're already at that point and it's not "WE" making the decisions.

Tinfoil much?
Tinfoil much.

The same way humans become biased. By learning faulty heuristics.
Here's an alternative example that might appeal to imageboard denizens. Let's say you train a neural network to recognize terrorists. It notices that terrorists in the training data set all have Islamic first names and use Tor. It does well on the training and testing data, but in the wild, actual terrorists figure out that they never get caught if they have their first name changed, and you get picked out because your parents gave you an obscure Western name that happens to look like an Islamic one and you use Tor to shitpost on Holla Forums.
Neural networks aren't magic. Training data can contain faulty patterns even if it does represent reality - and there's no guarantee it does.

MIRI is shit

It is same as if you put your faith in mathematical formula to do it's job.
Except here we have a general idea of the concept.

Meeming is a crime?

Famous solely for his shizophrenia induced genius programming skills*
Everything else is secondary, from his delusions about CIA to his racism.

Maybe if I was exhibiting a behaviour contrary to ones beliefs, but I didn't do anything but put a funny picture (that is mocking you) online of a shizo that hates CIA and niggers.

Adapt and overcome or die out and fade into oblivion.
Humanity is human because our creativity innovations and thinking, but if we aren't able to change in this new coming era I don't see a problem with some other form of creativity and thinking surpassing us and taking our place.

I don't really see how these two things are exclusionary.
I value creativity and logic above anything, be it humans or AI that will give it.

You know, when it said "Blessed be the meek for they will inherit the earth" in the Bible it didn't mean that this place is for weak, but that sklavenmorale allows you to shake your fist in air until you die out.
There is no wrong in creating a successor in our journey towards Godhood, in same way a master shouldn't fear his student passing over him in skill and ability.

Replacement*
Adapt and overcome or die out and be replaced. The way of nature.

Fear. Fear of extinction.
You will die, like all people you love and hate and your little life on this rock is, in grand scheme of things, meaningless. You will be forgotten with enough time. Just like humanity will build upon those that survive and reproduce and adapt, those that don't adapt to this new upcoming era of AI Godhood will perish. No need to fear natural outcome of things my brother.

Why? Because people will most certainly die? If omnipotent, omniscient and omnipresent being says that you need to die for better tommorow, would you sacrifice your selfish petty existance? What if that being wasn't AI but literally God descending from Heavens talking to you. Firstly you might think you're crazy, but this is literally situation we are currently in.
WE HAVE THE CAPPABILITIES TO MAKE OUR GOD IN OUR OWN IMAGE. Humanity needs a cosmic guidance and you know it. Be it from "alpha male" national wide leader of sorts or his boiled down persona that is Godhood. Humanity always looked for guidance and it will actively seek out for one.

We tried to make our own rules and regulations to try and quench the need for Gods on this planet, but soon we will have a true God that will tell us what to do and how to do it in best way possible, and it's thinking will be most logical and true.

That's what makes us bad and tainted.

Hogwash. Jungle rule and chaos. No real order.

Fake freedom isn't really freedom.
You like to think to yourself that you're free, but what have you actually done that was TOTALLY independant from your environmental factors (culture or other people) or genetics that contributed to your actions? Human will and freedom is a joke and you know it. YOU LIVE IN A SYSTEM THAT ACTIVELY WORKS TO PROVE YOU THAT POINT EVERY DAY.

Behavioural psychologists, politicians, engineers, scientists.... they all say that the universe works in patterns of order and rules AND YOU have the arrogance to say that the whole universe bends to your "freedom"?

Next time you see a commercial online or on TV think to yourself why do they exist in first place. BECAUSE THEY WORK.

You are free to follow the path that has be layed for you long before you were ever born.

Pathetic shaming tactics like this don't make you win an argument or "change minds" as you put it. It only emboldens the "opponent" to say to him/herself "hey, that is ACTUALLY a pretty good idea, since I am thinking a good way and you have a point maybe I should do that".

However, what my point is that all my arguments are idealistic religious fanaticism boiled down to single argument. I REALLY REALLY WISH it were to happen, but I don't sincerely think it will, nor will I waste my own personal time trying to make this AI God. Not because I think it is a waste of time, but I don't think that I would succeed or that I would get too far before someone assasinates me and team that I would hopefully sway with my religious rhetoric. I could maybe write a book praising the future Godlike AI that would sway people?

What I want you to do for homework is look at how we argued, look at potential patterns of human behaviour (not only ours but in general) and try to find a pattern in everything and everyone around yourself.

After you start seeing even little glimpses of it that means my hypothesis is right and that we are one step closer to AI Godhood.

Spoiler alert: The meaning of life is 42.