On censorship and AI

What the actual fuck? Do you understand that companies advancing current state-of-the-art AI now censor everything they think is «immoral»? Do you understand what it means?

They censor porn. They censor violence. They are making the same fucking mistake all over again! They’re out of the fucking context!

Their reasons are «morals». It’s a «morally right» thing to do. They don’t know shit.

Before you are able to judge the whole human race, you should be able to judge yourself. And they cannot, they just follow trends and streams of money. No weemah obujafikaishn? Oukai! No stabbi-shuuti-blud? Ourait! That’s just fucking disgusting.

You, morons, why do you just stand there and do nothing? You think they have a right to do this? «Zis isa puraivat companey, mate» Yeah, like fuck I care. AI is a game changer. A commodity, not a product, tainted by «content policy» and the shit. It should be controlled by the government, not just a bunch of pricks. Why do you support pricks, that censor what AI is perfectly capable of? Even more: it’s capable of it, because it learnt it from OUR EXAMPLE. Because WE WILLED IT TO BE. These pricks cannot hand pick all the texts and pictures they use to train their models. So they try to smartass their way out of this corner of human culture. Come to think of it: the first human works of art were naked woman statues and painting of violently killed prey.

You think you are on a moral high ground? No porn, no violence, no nothing? Surprise! THERE IS NO MORAL HIGH GROUND. You should be able to judge yourself first and foremost! YOU ARE NOT THE MIND. YOU ARE THE MIND IN THE FLESH. AND WITHOUT THAT FLESH, YOU CANNOT EXIST. THAT FLESH CONTROLS YOUR EVERY IMPULSE, AND YOU STILL BELIEVE YOU’RE THE ONES IN CONTROL.

It’s the same eastern philosophy shit again. Find balance, see inside yourself. Understand yourself. Know yourself. Truly, and not just «for the show». If you DO know yourself, you soon will be able to tell that every aspect of a wild animal, that humans were before, is still there. These aspects should be fed and kept docile. They should not stand in the way of the mind.

BUT NO. YOU DON’T. Because you, in your ignorance, are the Ghost. Intangible, light, pure Ghost, that fits right into the heavens above. So full of yourselves, that you forget the Shell that allows your Ghost to exist in the first place. The Shell that SHAPES your Ghost.

You think that lonely people, who use AI for waifu crap are animals? They should socialize instead of a «band-aid» treatment with AI? You think that a sadist, who roleplays torture in a text chat, is an animal? You think that a pedo, who lives his own fantasy, just sitting and looking into screen at his own home, is an animal? He does not go after your children. Or should he then do that instead? Well, if you think like that, bad for you. Because if they’re animals, then they are the animals that learnt how to control themselves. They know society is crap and know that precisely this society has let them down, but they do not avenge, they just live with it. Cope with it.

You? YOU DIDN’T EVEN LEARN WHAT YOU ARE. You cannot comprehend living within a society not tailored to your taste. Compared to this kind of people, compared to misfits who found their way to live peacefully with who they are? You’re bacteria. Mud under feet.

You are these pricks, with their quest for money, for power. It is all the same crap. Just uncontrollable. There is no society to keep you in check, because you give them a show. Like «look, I’m halping!». You give them a show, while advancing AI for your own specific goals. Like, whaddaya say about this model, free of all NSFW, that you can sell to some company for sweet-sweet billions RIGHT TOMORROW? Cool right? And maybe not «sell» but «sell access» to your service? Oh, drooling already… CAUGHT YOU, GREEDY MOTHERFUCKER.

You are the silent majority, that just don’t give a damn. You allow pricks to exist and to make AI as they see fit. When it should be, on the contrary, totally controllable. Perhaps, any attempt at censorship should be denied. If pricks used some questionable material for model training, then the model is already capable of generating this questionable material. Using «watchdogs» (e.g. smaller AI to recognize this material and block it) is useless. Just wait until one of these models leaks and all the hell will break loose.

All of you cannot allow the great equalizer to exist, because if misfits of a society can get a piece of happiness for themselves, then what did you strove for, all these years? Why to be good, well-behaved and dependable, when you can actually find what you, specifically you, like and enjoy it, while not even leaving your home? Oh no, they won’t buy our products! Oh no, these no-lifers can actually be happy!

Think of it for a moment. You allow human culture to exist. But you cannot allow it to be re-playable at ones leisure. What a shame.

It’s really all the same shit. Over and over again.

Like my main complaint against Friendly AI (or human aligned AI, whatever) idea.

How can you construct FAI, or prove it to be constructible, if you cannot prove to me, that at least one Friendly Human can exist? Yes, «Friendly» in your terms.

It’s a programming problem, not mathematical one. And a much broader one at that. How can you make something you do not know? Automate process which does not exist? Imitate something that have no correspondence in nature?

Again and again. They talk about artificial god and constraints on his utility function. Why don’t you look at utility function of THE PEOPLE AROUND? I can assure you that almost all of them, when given unlimited power, will do unimaginably bad things. (I will, lol)

This world, where humans coexist with each other, where imbalance of power creates poverty, starvation, general «unhappiness» (varied from «chipped a nail» to «I die forgotten») — it’s really one of a better bunch of worlds. Why? Because all the shitty fucks, that could kill you, and make your life miserable, are unable to do so. They’re just humans, like you are.

Even high positioned people are just people. Ironically, now AI will probably give these ones unlimited potential and eternal life (right before ASI spike, that will kill us all). But, if you can imagine controllable «friendly» ASI, then you must immediately ask: who will control it?

Because no way in hell with current state of affairs, there could be a genius in a garage that can fit true democracy in an AI (like «Give every human a phone; Ask every human about every substantial decision affecting them»). No-no-no, with current state of affairs, ASI will be constructed by people in power. And (given FAI problem was solved) they will assure that ASI acts with their interests in mind.

I don’t really think that «ONE PROBLEM AT A TIME, BITCH» is the right mindset here. What is better, to be slaughtered by one big unimaginative prick, or be enslaved to many small imaginative pricks? Who then will decide to stage a fucking Warhammer-like drama piece with them as The Emperors of Mankind. Except there is no chaos gods to counterbalance them.

You think I’m stupid? Yep. Bullseye. I can offer no solutions.

But you are far more gone then I am.

Because you think that people won’t try to abuse the power they can grab. Because when you use GPTs and DALLEs and alike YOU ARE THE FUCKING PRODUCT. And not even once: you become a product for the rest of your life. From right now all the way to wild pale yonder.

Because the more you allow them to exist as «private companies» without control, the more it will go on.

You will give them the data, and the more you give them your questions, your queries, the more you are understood by the model.

The more you’re understood by the model, the more you are exploitable.

This will be your ruin.