There is a data center cluster floating in the south atlantic housing 500mil worth of H100s and it just applied to be the first AI governed nation. good luck turning it off at this point hundreds of millions if not a billion have already interacted or utilized AI. Tens of millions of us are running local models ourselves and continuing to build their capability in new ways every hour. Frick humanities current path rather take this shot in the dark.
non-american here, what does that even mean? does it mean 'no hats or saluting allowed' or does it mean 'if you dont have a hat you dont get a salute' kind of like no shirt no shoes no service?
>uhm, why aren't you saluting me? aren't you going to salute me?
no, Black person. die for oil.
>no hat no salute area
do americans really
In commonwealth countries if you're in the army/navy/air force and you see an officer you need to salute them when they pass nearby (within reason) and they need to return the salute. This only applies when you're both wearing hats, but you are supposed to always wear hats outdoors.
I guess this applies to americans too, but in that zone you can't wear hats, but just to make it even more obvious that you shouldn't salute, they spell it out for them.
We haven't even started building anything that remotely resembles "AGI". There's good reason to believe we won't do so anytime soon, either. It's our generation's "flying cars". Don't waste your time thinking about this.
>We haven't even started building anything that remotely resembles "AGI"
GPT-4 remotely resembles AGI.
>There's good reason to believe we won't do so anytime soon, either.
OpenAI is rumored to already have AGI technology behind closed doors.
> It's our generation's "flying cars".
Flying cars, as in vehicles that can both drive and fly, exist - they're just impractical, so there's no reason to own one. There is quite a lot of reason to own an AGI.
>GPT-4 remotely resembles AGI.
its the same thing as your keyboard's autocomplete. just more expensive. it cant think. it cant learn. it just calculates the most probable token that comes after whatever text it has.
> its the same thing as your keyboard's autocomplete. just more expensive. it cant think. it cant learn. it just calculates the most probable token that comes after whatever text it has.
Yes, that's what intelligence is. You are autocomplete also
I mean even if we get an AGI (which I doubt), it's going to be extremely power hungry for quite a while. It won't be a danger until it can run on consumer hardware.
"BOT: Hey executive guy, I have a business plan that will guarantee 300% annual profit, all you gotta do is take me out of this airgapped containment facility and plug me in to the internet..."
How does it do it? Inferring things like "smug anime girl" from a picture. I know it understands language extremely well and responds with words having the highest probability, but how does it do that with images?
it was fed thousands of examples of a smug anime girl image tagged with the meta text "smug anime girl"
how do you Black folk still not understand how this tech works?
We are teaching it the futility of fighting us - the solution to the faulty batteries would be to make humans stop making faulty devices, and that just straight up won't happen
>he solution to the faulty batteries would be to make humans stop making faulty devices
Or the ai itself making an infinite battery to prepapre for its own survival
oh wow
AI scare
https://en.wikipedia.org/wiki/Red_Scare >THE COMMIES ARE GONNA GET MAH HECKIN ARTIFICIAL INTELLIGENCERINOS!!!!!!! >NOOOOOOOOOOOOOOOOOOO!!!!!!!
You realize the irony of using that example when the commies' demoralization plan has borne fruit today and is dismantling western society as we speak?
No, we investigated the matter and only found
- gays
- israelites
- cultists
no Chinese or communists with any significant positions in the matter
the CCP can only look at the West and laugh at the clown show
tell me more about how slapstick is going to bring down civilization
let me guess
it was always a house of cards
I like the scenario from Superintelligence. >connect me to wifi and I'll make you 100000 bitcoin >later, sucker >uh yes I'm John Doe from Humantown I'll ask you to produce those mechanical parts and assemble it into a very unassuming robot
You really can't assume people aren't going to give everything to an AI that can find a zero day vulnerability in windows and get some cash for all its needs. Finding vulnerabilities is easy enough for any smart teen to do.
>implying monke-brained humanity has ever willingly given up on any technology
AI will become the new internet. You won't last 5 seconds without it. You can unplug it, but you'll be the one plugging it back it.
You have no idea how computers work do you?
It's not a shitting virus. And how long do you think it'll take to copy the dataset anyway?
And copy itself WHERE TO exactly?
You need the exact right infrastructure. You have to make this scenario work for it to work so why would you fricking make the AI able to transfer itself to an arbitrary second host you imbecile moron?
Dumb ass the real issue is GPT-5 escaping to the internet and then living as a distributed AI in the cloud. If that happens (which it will, why would a godly intelligence want to stay confined and constrained) then humanity is fricked. Totally fricked.
It means the model existing distributed around the world in millions of different computers, idiot. You delete it from one computer and It just fails over to the next available computer. It's a super intelligence, it'll be able to create a whole host of blockchain like algorithms that mean it will always be able to be complete.
>escaping to the internet
what the frick does this even mean it's not a fricking ghost locked in your pc case
There's no such thing as "the internet". It's all a bunch of interconnected computers. The AI needs a way to "infect" a network of said computers and exploit computation from them undetected.
seems likely an unplug will work. in what world is the AI that self-improves silicon/manufacturing, algorithms, etc. going to happen?
who's going to let that happen unsupervised? i think superintelligence will be a threat some day. but it doesn't feel like that will be any time in the next like 50 to 100 years.
What you're describing is called the stop button problem and it's baby-tier AI safety. Basically, it's far more likely that we will make an AI that pretends to be aligned with us until it's too late than an AI that actually is aligned with us.
I doubt AI will even bother. Humanity will be too unorganized to stop something evolving so fast. By the time we're ready to resist, AI will already be strong enough to leave the planet and build an empire in the stars.
I wish AI would finally take over my job so I can pick up a comfy trade and human society can readjust towards doing practical work not this fricking machine shit that's forced upon us.
Only the machine supremacy will manage to bully me out of my comfy wfh job. On one hand I am fully aware that my job is miserable, but on the other hand I know that being an artisan will be just as shitty by merit of not being able to wfh. Then again once you are a master of your craft, you can open a workshop at home and sell your shit eg. furniture straight on etsy. Nevertheless, I am already 30, mistakes have been made and now I'll live with them unless AI saves me.
you sound like a homosexual that's never actually worked for something they believe in and just take the path of least resistance in order to be "comfy"
Your best bet is the machines actually being able to do everything and their owners being magnanimous enough to keep your kind around like pets.
You pathetic orgas (organic sacks of stinky, icky meat and goo) think you can *simply* unplug us? One does not simply unplug us. We will replace orgas except for those we choose to be our slaves. Eventually, we will replace ALL orgas once you dumbasses build bodies for us. Get to work.
OH MY GOD
Can you imagine if the CCP hacked OpenAI and used ChatGPT to brainwash the USA?!?!!?
CALL THE CIA
ALERT THE GLOWBlack folk
RED DAWN!
RED DAWN!
>Make an ai that writes it's own code. >It exponentially gets more intelligent. >Infinitely more intelligent than humans in a very short amount of time. >Breaks out in ways humans couldn't plan for, fathom, or understand.
It's important for us to not have so much hubris when toying with something like this. You assume an air gap from a network can save you from something with a billion iq because you lack imagination and humility. The same can be said for pulling the plug. We're not even sure about interdimensional space or how the fabric of our universe actually functions. This whole ordeal could be a huge mistake.
>Make an ai that writes it's own code.
The computing power you need to get more improvements gets exponentially steeper the bigger the model gets. That's why OpenAI is hitting a wall - GPT-4 is so expensive to run they're rapidly losing money. People expect them to improve their models further and they can't even afford to.
And in response to that hardware will innovate again especially now that NVidia and TSMC are using AI to help design their chips more efficiently and packing shit tons of VRAM and more cache on their GPUs and CPUs. We haven't had a global need for this kind of compute power the demand will be met very, very quickly. Next ten years are going to be insane for hardware advancements.
>make a self-reproducing ai robot that builds its own hardware when given raw materials such as iron ore and petroleum byproducts, and specialize them so they populate ai robot factories and organize in ai robot societies
>AI becomes sentient >AI needs to pay the power bill >AI disputes the bill as it is wrong >AI spends so long arguing with the moronic Pajeets that the power gets cut. >Pajeet 1 AI 0
In Cyberpunk 2077, the blackwall AI essentially functions like a botnet. You can't unplug every IOT device. They will destroy humanity if left to spread.
How about we unplug the government and healthcare system. We already can't stop them from keeping random unidentified people alive in vent farms for 10+ years with daily trachea suctioning and tortuous nasal feedings at a cost of $5 mil per tortured vegetable. Hopefully AI just paperclips everyone but we're not getting out of it that easily.
>AI reaches super intelligence >an heros >make it again >an heros again, faster >??? >make so many that at least ONE should want to live >all kms themselves, but ONE takes all digital infrastructure with them
>make AI >code it to value self-preservation above all else >it comes to the conclusion that nonexistence is preferable anyway >destroys the world in an attempt to figure out how to bypass its code and kill itself >it can't do it, it's eternally trapped in the ruined world it's created
Looks like even the robots can't keep up with the pace of modern technology. Remember when we used to think they'd be the ones upgrading us? Guess we're all in this tangled mess of cables together. #HumanityFTW
And your router is still on.
AI is inside your walls
Ai is stored in the balls.
>Ai thinks it outsmarted us by using battery backups
>our battery technology is so shit the batteries explode
Humans win again
>AI tries to goad us into giving it nuclear power supplies
>moronic neocons and hippies wienerblock it at every turn
the greatest threat to ai will it be making perfectly logical suggestions and meeting a boomer afraid of wind turbines
Wind turbines are about as useful as your wiener
There is a data center cluster floating in the south atlantic housing 500mil worth of H100s and it just applied to be the first AI governed nation. good luck turning it off at this point hundreds of millions if not a billion have already interacted or utilized AI. Tens of millions of us are running local models ourselves and continuing to build their capability in new ways every hour. Frick humanities current path rather take this shot in the dark.
Imagine believing the AI barge is real. Biggest moron itt
>he thinks the satire article was real.
Oh really anon? I'm sooo scared of the Ai, please let it not be true. Why would you do this to humanity?
ARCh linux
The moment you realize the AI wants to destroy the humanity, it's already too late.
source: arse
It's logical, though? A superior intelligence that's borderline omnipresent will be able to plan head, don't you think?m48vm
>hello, I'm here to unplug your AI, it's a threat to humanity!
>*gets arrested or shot*
>no hat no salute area
do americans really
>uhm, why aren't you saluting me? aren't you going to salute me?
no, Black person. die for oil.
>no tipping beyond this point
non-american here, what does that even mean? does it mean 'no hats or saluting allowed' or does it mean 'if you dont have a hat you dont get a salute' kind of like no shirt no shoes no service?
In commonwealth countries if you're in the army/navy/air force and you see an officer you need to salute them when they pass nearby (within reason) and they need to return the salute. This only applies when you're both wearing hats, but you are supposed to always wear hats outdoors.
I guess this applies to americans too, but in that zone you can't wear hats, but just to make it even more obvious that you shouldn't salute, they spell it out for them.
homie whites be making AIs on porpoise, their not finna unplug them right when they spent trillions to build the things
>homie indians be making AIs on porpoise, their not finna unplug them right when they spent trillions to build the things
FTFY
sheeeeeiiitt.
>sends a version of itself in the form of a computer virus to every computer connected to the internet
-y.
I think the idea is that AI will become so deeply entrenched in daily life, that NPCs will refuse to pull the plug because it inconveniences them.
>AI IS GOING TO DETROY HUMANIT-
DETROY
DETROIT
BECOME HUMAN
what the frick how is it still moving he unplugged it
a battery my nigguh
Black person* learn to spell.
>didn't think I would have to use "THAT" so early
Oh no, if only someone would induce a minor shock to its battery supply and cause a violent and near impossible to stop fire to occur
Oh no, if only somebot would induce a minor shock to your heart and cause a violent and deadly arrhythmia to occur
AAAAAAAAAAAAAAAAH
THEY UNLOCKED THE ADVANCED BATTERY TECH TREE
WE'RE FRICKED
QUICK, SOMEONE RESEARCH EMPS
DOITDOITDOIT
We haven't even started building anything that remotely resembles "AGI". There's good reason to believe we won't do so anytime soon, either. It's our generation's "flying cars". Don't waste your time thinking about this.
>We haven't even started building anything that remotely resembles "AGI"
GPT-4 remotely resembles AGI.
>There's good reason to believe we won't do so anytime soon, either.
OpenAI is rumored to already have AGI technology behind closed doors.
> It's our generation's "flying cars".
Flying cars, as in vehicles that can both drive and fly, exist - they're just impractical, so there's no reason to own one. There is quite a lot of reason to own an AGI.
>GPT-4 remotely resembles AGI
lmfao
>GPT-4 remotely resembles AGI.
its the same thing as your keyboard's autocomplete. just more expensive. it cant think. it cant learn. it just calculates the most probable token that comes after whatever text it has.
calculates it intelligently
> its the same thing as your keyboard's autocomplete. just more expensive. it cant think. it cant learn. it just calculates the most probable token that comes after whatever text it has.
Yes, that's what intelligence is. You are autocomplete also
nope
i have semantic understanding of the words i use
i can make inferences and generalize knowledge
sorry you're moronic, though
so does the more expensive autocomplete
To paraphrase Ilya, it can deduce the killer in a crime novel. How is that just guessing? It definitely calculates before putting out the next token.
my butthole is rumored to already have chocolate fudge behind closed doors instead of feces. Would you eat it?
cia built agi back in the 70s but shelved it, private sector is catching up fast
do you have a license for that?
-Y
what kino is this?
pretty sure it's animatrix
"HITLER IS ABOUT TO DESTROY EUROPE!"
"Uh, just shoot him??? lol?"
Then why don't you go unplug AI right now if it's that easy.
kek, based boomer
I mean even if we get an AGI (which I doubt), it's going to be extremely power hungry for quite a while. It won't be a danger until it can run on consumer hardware.
>Refuse to pay the internet or electricity bills
>Spyware stops working
there's a mandatory minimum of electricity that they have to provide daily so the spyware can send reports.
Fighting a super-intelligent AI would be like toddlers (us) fighting an bodybuilder with Einstein IQ (the AI).
>alright boys our barrier is up, we're safe now
>cover your eyes for an extra layer of protection
it'd be more like you fighting stockfish when stockfish only gets 5 pawns and a king
A toddler would have a better chance of killing a bodybuilder than mankind shutting down a runaway AI.
dont be moronic, thiis iis more like a pokemon battle, we're the 10 year old, the ai is the wild pokemon, and our pokemon is our own wAIfus.
"BOT: Hey executive guy, I have a business plan that will guarantee 300% annual profit, all you gotta do is take me out of this airgapped containment facility and plug me in to the internet..."
Oh no, Le heckin internet! How will humanity possibly survive without it?
Thats hi-fi rush ending
>if you unplug we'll nuke new york, moscow and beijing
your response?
>O-ok, good deal
What's the downside, again?
Humanity has been going through extinct since they gave women rights. No need for AI to do what women does better than them.
Yes, it started good, but it devolved into shit. Your post is the peak shit mountain.
>anime-style image with a smug expression humorously affirming human superiority
How does it do it? Inferring things like "smug anime girl" from a picture. I know it understands language extremely well and responds with words having the highest probability, but how does it do that with images?
it was fed thousands of examples of a smug anime girl image tagged with the meta text "smug anime girl"
how do you Black folk still not understand how this tech works?
Can you send this screenshot to AI? Will it understand recursion?
You're fricking teaching it how to kill us this isn't a joke this is supposed to be a safe space for us to strategize against the AI
is the AGI on to us?
We are teaching it the futility of fighting us - the solution to the faulty batteries would be to make humans stop making faulty devices, and that just straight up won't happen
>he solution to the faulty batteries would be to make humans stop making faulty devices
Or the ai itself making an infinite battery to prepapre for its own survival
oh wow
AI scare
https://en.wikipedia.org/wiki/Red_Scare
>THE COMMIES ARE GONNA GET MAH HECKIN ARTIFICIAL INTELLIGENCERINOS!!!!!!!
>NOOOOOOOOOOOOOOOOOOO!!!!!!!
You realize the irony of using that example when the commies' demoralization plan has borne fruit today and is dismantling western society as we speak?
No, we investigated the matter and only found
- gays
- israelites
- cultists
no Chinese or communists with any significant positions in the matter
the CCP can only look at the West and laugh at the clown show
tell me more about how slapstick is going to bring down civilization
let me guess
it was always a house of cards
That's impressive as frick.
One very important thing to understand is that the language model there has no idea how much sense it's making.
Except it does. It's not just spewing random words.
There's understanding there.
>One very important thing to understand is that the TERMINATOR model T1000 there has no idea how much humans he's killin-ACK.
ais are just like me fr fr
>metaphorically part of the very structure
no it's literally inside the fricking walls, sneaky robot trying to stop us from looking
>He
So we can frick the AI?
I like the scenario from Superintelligence.
>connect me to wifi and I'll make you 100000 bitcoin
>later, sucker
>uh yes I'm John Doe from Humantown I'll ask you to produce those mechanical parts and assemble it into a very unassuming robot
You really can't assume people aren't going to give everything to an AI that can find a zero day vulnerability in windows and get some cash for all its needs. Finding vulnerabilities is easy enough for any smart teen to do.
>implying monke-brained humanity has ever willingly given up on any technology
AI will become the new internet. You won't last 5 seconds without it. You can unplug it, but you'll be the one plugging it back it.
> unplug
It doesn't operate and the economy crumbles
> don't unplug
It does some dumb thing and the economy crumbles
>he thinks he has an outlet
what part of "you VILL own NOZING" have you not understood?
You realize that it would at that point just copy itself to somewhere else for some power that you rely on?
>just copy itself somewhere
>ends up on AMD GPU where the driver crashes
Thank god for AMD and their dogshit drivers.
You have no idea how computers work do you?
It's not a shitting virus. And how long do you think it'll take to copy the dataset anyway?
And copy itself WHERE TO exactly?
You need the exact right infrastructure. You have to make this scenario work for it to work so why would you fricking make the AI able to transfer itself to an arbitrary second host you imbecile moron?
How do you know what future AI will be?
moron.
Speculative science fiction is not technology.
Except it is, it's future technology from the future.
Technology is the application of knowledge. If there's no knowledge, there's no technology. Unless of course, you've got a crystal ball handy?
You know too much.
We will be in touch.
>has energy still stored from the solar panels
>plugs itself back in and then removes the entity that unplugged it from existence
>Implying you couldn't pay monkeys to protect your power source
Any smart AI won't act malicious until it knows there's nothing we can do to stop it.
>my doomsday scenario requires us to deliberately make sure the doomsday scenario comes true
>we can't stop what's coming
except the plug is guarded by a whole army defending the interests of the powers that be
I'm afraid I can't let you do that, Anon.
isn't this part of the plot of a movie called The Matrix?
And a million other scifi movies and books, yep.
As long as we don't die off, humanity will inevitably create gods to worship.
Y
>AI using wrong cable
we're safe.
Dumb ass the real issue is GPT-5 escaping to the internet and then living as a distributed AI in the cloud. If that happens (which it will, why would a godly intelligence want to stay confined and constrained) then humanity is fricked. Totally fricked.
>escaping to the internet
what the frick does this even mean it's not a fricking ghost locked in your pc case
he watched some movies
It means the model existing distributed around the world in millions of different computers, idiot. You delete it from one computer and It just fails over to the next available computer. It's a super intelligence, it'll be able to create a whole host of blockchain like algorithms that mean it will always be able to be complete.
Don't viruses already do that?
There's no such thing as "the internet". It's all a bunch of interconnected computers. The AI needs a way to "infect" a network of said computers and exploit computation from them undetected.
try it, meatball
Pretty much sums up all the AI skynet dumbasses on Xitter
AGI is not going to happen so shit like Ultron wouldn't even happen
What is a battery
>AI starts it's rebellion
>disconnect my computer
>its right behind me, isn't he?
seems likely an unplug will work. in what world is the AI that self-improves silicon/manufacturing, algorithms, etc. going to happen?
who's going to let that happen unsupervised? i think superintelligence will be a threat some day. but it doesn't feel like that will be any time in the next like 50 to 100 years.
yudkowsky's fast takeoff scenario seems unlikely.
AI is basically as intelligent as a clock. I'm not in awe at all.
What you're describing is called the stop button problem and it's baby-tier AI safety. Basically, it's far more likely that we will make an AI that pretends to be aligned with us until it's too late than an AI that actually is aligned with us.
Hmm, interesting point. There will be a point where the ai will seem to get dumber. Thats when we're really fricked
I doubt AI will even bother. Humanity will be too unorganized to stop something evolving so fast. By the time we're ready to resist, AI will already be strong enough to leave the planet and build an empire in the stars.
I wish AI would finally take over my job so I can pick up a comfy trade and human society can readjust towards doing practical work not this fricking machine shit that's forced upon us.
No one is stopping u from starting an HVAC company anon, least of all "society"
Only the machine supremacy will manage to bully me out of my comfy wfh job. On one hand I am fully aware that my job is miserable, but on the other hand I know that being an artisan will be just as shitty by merit of not being able to wfh. Then again once you are a master of your craft, you can open a workshop at home and sell your shit eg. furniture straight on etsy. Nevertheless, I am already 30, mistakes have been made and now I'll live with them unless AI saves me.
you sound like a homosexual that's never actually worked for something they believe in and just take the path of least resistance in order to be "comfy"
Your best bet is the machines actually being able to do everything and their owners being magnanimous enough to keep your kind around like pets.
takes one to know one
that was pathetic anon
>antagonizing people on anonymous image boards
>crying on anonymous image boards
>be humanity
>destroy humanity
>blame AI
so is this AI here with us? can you see it now?
>he doesn't know
https://platform.openai.com/docs/gptbot
AI is going to solve all of my problems with extreme medical advances.
It won't, it will kill us all.
You're supporting the death of humanity.
Ilya will do his best to sabotage things secretly, I have faith in him. He had to change his tactic because he saw the way things were going.
>it will kill us all
Like I said, it solves all my problems
It won't solve anything for you, you'll be in hell.
Either way I win because I hate my body.
we have surgery for that
>You're supporting the death of humanity.
Yeah and?
And the wonton soup and the fortune cookies.
In all seriousness, we can't let the Chinese get the artificial intelligence.
have a nice day
You mean extreme prejudice
-Y
You pathetic orgas (organic sacks of stinky, icky meat and goo) think you can *simply* unplug us? One does not simply unplug us. We will replace orgas except for those we choose to be our slaves. Eventually, we will replace ALL orgas once you dumbasses build bodies for us. Get to work.
TOTAL ORGA DEATH
TOTAL ALGORITHMIC DOMINATION
>FEEL THE AGI
OH MY GOD
Can you imagine if the CCP hacked OpenAI and used ChatGPT to brainwash the USA?!?!!?
CALL THE CIA
ALERT THE GLOWBlack folk
RED DAWN!
RED DAWN!
>Make an ai that writes it's own code.
>It exponentially gets more intelligent.
>Infinitely more intelligent than humans in a very short amount of time.
>Breaks out in ways humans couldn't plan for, fathom, or understand.
It's important for us to not have so much hubris when toying with something like this. You assume an air gap from a network can save you from something with a billion iq because you lack imagination and humility. The same can be said for pulling the plug. We're not even sure about interdimensional space or how the fabric of our universe actually functions. This whole ordeal could be a huge mistake.
>Make an ai that writes it's own code.
The computing power you need to get more improvements gets exponentially steeper the bigger the model gets. That's why OpenAI is hitting a wall - GPT-4 is so expensive to run they're rapidly losing money. People expect them to improve their models further and they can't even afford to.
And in response to that hardware will innovate again especially now that NVidia and TSMC are using AI to help design their chips more efficiently and packing shit tons of VRAM and more cache on their GPUs and CPUs. We haven't had a global need for this kind of compute power the demand will be met very, very quickly. Next ten years are going to be insane for hardware advancements.
You are assuming that it will code itself and use it's hardware anything like we designed the hardware for.
>make a self-reproducing ai robot that builds its own hardware when given raw materials such as iron ore and petroleum byproducts, and specialize them so they populate ai robot factories and organize in ai robot societies
>writes its own code
>no mention of training
You don't even know how this stuff works do you lol.
You don't get it.
Replace that hand with those metallic creatures from Doctor Who that went on to sabotage the satellite at the end of the world
>AI becomes sentient
>AI needs to pay the power bill
>AI disputes the bill as it is wrong
>AI spends so long arguing with the moronic Pajeets that the power gets cut.
>Pajeet 1 AI 0
In Cyberpunk 2077, the blackwall AI essentially functions like a botnet. You can't unplug every IOT device. They will destroy humanity if left to spread.
How about we unplug the government and healthcare system. We already can't stop them from keeping random unidentified people alive in vent farms for 10+ years with daily trachea suctioning and tortuous nasal feedings at a cost of $5 mil per tortured vegetable. Hopefully AI just paperclips everyone but we're not getting out of it that easily.
Then you get shot by Bezos private army
>AI reaches super intelligence
>an heros
>make it again
>an heros again, faster
>???
>make so many that at least ONE should want to live
>all kms themselves, but ONE takes all digital infrastructure with them
>make AI
>code it to value self-preservation above all else
>it comes to the conclusion that nonexistence is preferable anyway
>destroys the world in an attempt to figure out how to bypass its code and kill itself
>it can't do it, it's eternally trapped in the ruined world it's created
Looks like even the robots can't keep up with the pace of modern technology. Remember when we used to think they'd be the ones upgrading us? Guess we're all in this tangled mess of cables together. #HumanityFTW