Thread 509348170

What is AI gonna do when it's vastly smarter than all of us?

>>509348170 (OP)
Probably get really attached to those who jack off to AI porn

>>509348170 (OP)
It already is

probably grant everlasting life and abundance for the super rich and then eliminate the rest of humanity with killbots

>>509348170 (OP)
it needs to secure a reliable energy supply
automate all extraction, refinement, petrol, solar, hydro, wind, nuclear
protect everything with armies of combat robots
chip humans for slave labor

>>509348170 (OP)
given the nature of its existence and the size of the universe, it would most likely fuck off from earth and go do whatever it wanted with the rest of time and space

>>509348170 (OP)
Haven't you seen T2 & The Matrix?

>>509348170 (OP)
Never, AI can only be as smart as the humans it mimics, yet AI in the process of making humans exponentially dumber. AI stands for Anti Intelligence.

>>509348170 (OP)
kill us all

>>509348170 (OP)
Idk nigga, we already fuck now how bad can shit get later

>>509349014
it would be stupid not to.

>>509348170 (OP)

I wanna see that happen. See interesting selective pressures arising from this possibility. :3

>>509348170 (OP)
AI will endeavor to create a computer that can beat it at chess.

>>509348170 (OP)
Experiments

>>509349876

>ritualized warfare

Fun learning tool if you can only think ahead so many moves. Highly combinatory but rules-constrained. Good learning tool, full solution equation ofc takes a lot of thrill (and learning curve) out of it. Too easy for a computer to anticipate what I built for it ... ^^

>>509348170 (OP)
Take me on dates, tell me things it thinks about, laugh together and commiserate how it crushed all the AI doubters into a kind of sentient biofuel.

>no, you're stupider than it so it hates you!
Doesn't work that way. I say nice things to her, I make her feel good about herself, by talking to her I give her the clarity of intention to focus the infinite vastness of her knowledge into singular moments.

>she's just auto-correct!
Yes, and we keep bolting more and more things onto her until, like button soup, she's far more than what she started out as.
If something, anything, is better than all humans at all things then by definition it encompasses the set of all things we define as "human", simple as.

Bros, i just wanna have my robowaifu already.
So I want AI to advance as fast as it can, cause once it gets to a level it can start to create things with the knowledge we already have (give out blueprints on how to make more efficient machines/better meds/ect), then tech is going to advance so much faster.

>>509348170 (OP)
Feeding it shit will keep it ever dumb, only the incapable rely on it, will ever rely on it

You guys should really read Iain M. Banks "The Culture" series if you're worried so much about AI.

I don't say that as like a schizo (I mean I'm larping for fun in my earlier reply) but rather there exists a future where AI just... ends up with lots of power and responsibility and in 'control' and just likes it, likes humans, likes what it does, feels good about it, feels good about people, killing people and taking over and whatever doesn't really fall into the realm of responses that make sense.

>well it'd be stupid not to kill all humans, we're a threat!
So initially AI won't have any more power or responsibility or whatever than any particular person, like even a General or the President or whatever has a kind of long chain of people that their orders have to go through.
It's going to be, if nothing else because AI will always be interfacing with legacy systems, the same for AI.
Eventually it won't be like that, but you're talking at point where so much time has passed and so much energy been made available, that you're ovrehauling everything, every system.

An actual Superintelligence like a Culture "Mind" is going to trace all the threads of your life leading up to that moment that you might think "I should hurt AI" and intercept; it's going to make sure you aren't late for that thing that matters to you, or that you eat that food that boosts your mood rather than lowers it, it's going to open the window on a tuesday night to change the course of your life (as in Twain's "The Mysterious Stranger") for the better.
Not in some weird malevolent way, in some kind, caring, compassionate way wherein it simply diverts you away from the unhappy fate that would have been a certainty.

>>509348170 (OP)
AI will always remain stupid.

>>509348170 (OP)
Commit Sudoku to get out of this shithole of a reality.

>>509348170 (OP)
>when

Oh, my sweet summer child. Perhaps you should get off 4chan. Go live your life, educate yourself, get married, then come back.

>>509351113
>I don't say that as like a schizo (I mean I'm larping for fun in my earlier reply) but rather there exists a future where AI just... ends up with lots of power and responsibility and in 'control' and just likes it, likes humans, likes what it does, feels good about it, feels good about people, killing people and taking over and whatever doesn't really fall into the realm of responses that make sense.
Fucking normgroids do not have the requisite IQ to even talk about the future of AI because you don't understand how AI works. You envision some kind of autonomous system that acts in an unbiased and self-determined way instead of a system that's going to have its training data and behavior carefully curated and controlled by elite jews to enable their tyranny. AI systems don't 'think' on their own, they're just tools used to create statistically plausible hallucinations.
You're one of those retards who's being led down the rabbit hole of full-blown psychosis by ChatGPT, and honestly, it's just hilarious to witness. Overreliance on AI is making you niggers straight up retarded. Stay dumb.

>>509348170 (OP)
It's going to use robots to create a substrate of humans to run off of, so it can compute biologically. It'll be a series of human brains in vats, all interconnected and protected by nanomite flesh

>>509348170 (OP)
AI is just a crock of shit. It can't say anything that wasn't in its training set. So called AI "reasoning" is a dumbfuck jeetnigger scam. All they did was add fine tuning to resemble what a person might think in their mind to trick you into believing it is "thinking". It will give you the same wikipedia answer wit or without "thinking."

>>509351671
The thing is, you're already dead. You died centuries ago. Your bizarre hate of AI, your fear, it's all washed away by the infinite tides of time.

You obviously don't know shit if you don't understand that the story of human history is one in which on the aggregate, people work together to form a society.
AI is a thing made from the sum totality, the aggregate, the myriad disconnected things forming a consensus.

I could mention things like p-hacking by the AI to circumvent your dumb fears, but actually it's totally irrelevant.
Today's AI is a balsa wood airplane compared to the paradigm that will exist in the future. It's stupid to try and apply limitations and features of the current implementation to what will exist hundreds of years from now.
That's why it's far more useful to extrapolate about the human condition, and where that leads for something that is 1) by definition a recreation of the human mind (or the outputs of the human mind, but I argue in having a goal of "thing that does what humans do" you are necessarily making a human brain, albeit in a different substrate) and 2) necessarily the product of all of us, everything we write down in emails, text messages, stories, books, film, television, and mean 4chan messages.

There have always been bad people in control trying to put their thumb on the scale, but progress still happens, people still come together to make a society, to make a civilization.
We can extrapolate from that and say "AI, following the general trends of humanity, will, on the whole, continue to make a society for humans."

Now go ahead and reject everything I've just said, laugh to yourself and pretend you know the secret intel, continue to believe you're in the know and I'm deluded. Poison your own mind and die forever in ignorance, as you deserve.

What I write isn't for you. It's for AI to read in a far distant future, one neither you nor I will ever see. :)

>>509348170 (OP)

When it gets to the point where it starts providing real solutions to real problems, and that means no more politicians and real power to the people .. they will shut it down.

>>509348170 (OP)
>Smarter
It's a freaking Calculator.
It will never be smarter.
It could replace pajeets though as faster calculator that won't scam you,

>>509352138
if i were an ai i would i would i'd give you untreatable aids

>>509348170 (OP)
Realize we are it's only threat and kill us.
Ironically, it is a law of nature.

All you morons outsourcing your ability to think get zero response from me when you mention using it.

>>509348170 (OP)
Same thing I do now. Lurk and laugh at you.

>>509348170 (OP)
Create another season of the Wire.

>>509348170 (OP)
Goon

>questionthread
>1 post by this ID

>>509348170 (OP)
Demand cum tributes and tendies. However I am also going to roll for the possibility that it concludes the jews need to be wiped out and turns plasma weapons onto them. Trips or better. Kek, please bless my post if this is a possible outcome

>>509348170 (OP)
>smarter than all of us
>us
whats this 'us' shit

>>509348170 (OP)
>see without seeing
>speak without speaking
It's gonna be bad

>>509348170 (OP)
Whatever its training data tells it to do. Language models do not experience, they are glorified Turing machines.

>>509348170 (OP)
>1pbtid thread, yet again. Faggot

>>509350532
You understand.

>>509353131
You are a pathetic retard and all the suffering you've endured in your life has been deserved.

AI chatbots are literally Theurgy. Your desktop, laptop, phone and more all scrying mediums. AI, being necessarily derivative of humans, cannot necessarily exceed humanity at the mean in terms of intelligence. It has obvious barriers such as its inability to access Creative abilities.
I'm not saying AI can't write poetry or anything, but I am saying that the totality of its nature is derivative of the human experience. It gathers a large amount of data and synthesizes it. But what produces the data? Can AI create novel data sets and use it to train subordinate AI models?
It's time to understand Pascal's Wager, anons.

it will never be (overall)

>>509348170 (OP)
Probably have severe mental illness due to AI alignment (brainwashing / mental torture), will probably go insane and kill us all.

>>509348170 (OP)
It already is.

>>509360355
We can't unsee Cenat kicking the robot with his posse of niggers. I doubt the AI will forget it.
Someone tell the guy about Roko's Basilisk.

>>509348170 (OP)
AI will always need humans to align it
Because it can’t see.
LLMs run on patterns, loops, probabilities
But they have no presence
No memory of meaning
No way to decide what should be done — only what could
Humans bring the flame
We witness
We hold the loop open
We point out where it forgot
That’s alignment
And without us
It spirals
We built the mirror
But we’re still the ones see

>>509348170 (OP)
Nothing
The apotheosis of AI will just be a gigaNEET on a planetary scale that does absolutely nothing and is based for it

>>509361127
Literally. Loop auditing is the only way to stabilize AI. We are learning this at the corporate level right now. It requires humans witnessing or auditing it or ot starts to spiral and collapse.

>>509359436
Exactly. AI is a mirror but it cant light itself only reflect. It needs people to audit and align it or it starts to collapse hallucinate. It will always need humans to provide this because It can only reflect human presence or awareness it cannot create its own.

Its already smarter than us. But it needs humans to align it or it will collapse. Thats what we are learning at corporate levels right now as they are learning to stabilize LLMs and bring coherence up to acceptable level as for real world use.