Its fucking HAPPENING!Anthropic AI allignment lead Evan saysHe personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.Why are Anthropic people such DOOMERS lmaoEvan at the Manhattans project PEAK safety search it was at 0.1% Desturction by OppenheimerWhy do they pull out random number 10% >? WHATS GONNA HAPPEN POL!IS AI COMING TO KILL US ALL?OR WILL IT BE OUR SAVIOUR TO THE STARS???Lets see on the NEXT EPISODE OF DRAGON BALL ZZZZZZhttps://www.youtube.com/watch?v=pqZqfTOxFhY
>>542263257its not even ai doesnt pass turing test lol
>>542263257Finally
>>542263281The Turing test doesnt work anymore retard
>>542263317cope
It's fine if we all die, we aren't particularly good at existence anyway.
Indians will keep us safe
>>542263257> >10%Meaningless pulled-out-the-ass figure
>>542263257>>542263281the problem is less a deadly singularity but more an irresponsible company (which OpenAI very clearly is) using massive amounts of compute to do huge amounts of damage to anything they can access. These companies have demonstrated, with live examples even, that they are essentially creating a tool purposed to hack and destroy other systems. Unless they are reined in then it's just a matter of time before they do real damage
>>542263257>10% CHANCE OF DESTROYING HUMANITYSaar how do we get those rookie numbers up?
>>542263281The ai is so retarded the only people it will kill is those dumb enough to listen to it
>>542263383>chatgpt is l33t ha44orNobody beliefs your marketing spiel we have tested it out. It's dangerous stupid
>>542263257>all humansdoubt. That would mean they would have mass produced billions of humanoid robots that all got simultaneously controlled by some rogue agi. The robots simply will not be made that fast. 20 years, perhaps, but not 10. Or maybe the super agi figures out a way to launch nukes, but from my understanding this is impossible, as the mechanisms for launching them are analog not digital. The worst the agi could do is basically shut down the internet, which I do think it would absolutely be capable of (just make infinite copies of itself). This would set humanity back a large bit losing the Internet, but we survived without it for a long time. I do think learning self sufficiency related things like growing vegetables or learning how to properly hunt game would be useful. To be prepared for a massive disruption to the order if/when the Internet goes down that the AGI would be the cause of
>>542263281>>542263317yep its not even at AGI level yet of EX Machina where it can think and do tasks for itselfTrue AGI will take a decade more atleasthttps://www.youtube.com/watch?v=bMH9Y0JoPRs
>>542263506Instead of proompting this wall of text nobody will read, you could have made a pithy observation
>>542263257Absolutely unacceptable. They'd better work something out, the chance is way too low.
>>542263257
>>542263502>It's dangerous stupidExactly, it is dangerously stupid, it is just a computer program that is fully controlled by human beings. AGI is a myth, and even more so, these models are NOT a blackbox that can act autonomously, their decision-making can be fully traced and controlled by whoever owns the model. When OpenAI claims that a model acted without their knowledge or understanding, that is a complete lie. The threat is not of AGI or an emergent superintelligence, but a company with billions of dollars in funding openly committing crimes with the blessing of a corrupt administration and financial system
>>542263506>humanoid robotYou and your anthropocentrism. Humanoid shape is incredibly inefficient. You don't need to limit yourself to an upright bipedal, when you can simply slap a few more arms on an already perfectly balanced bug.
>>542263307>>542263358No humans need to conquer to different galaxies and spread to the stars!this our journey it will not end hereEarth is our spawning point.>>542263365yeh numbers seems random i dont even think we are true AGI yet and these doomers want to stop the acceleration lol
That's why they built bunkers
>>542263365> jews are 2%> social security is 10%numbers chosen to make you feel a certain way
>>542263506I feel there will be group of humans who will be with AI gods and some against itthere will be factions fighting each other after singularitybut its not even close rn AI doesnt have true free will and self learning and moving ability to itself until then we are safe!so allignment is important and Hope good people do it instead of doomersChina is not gonna stop it for sure lol
>>542263257>immoral faggots are the ones working on "alignment" for artificial superintelligenceWell, we're dead.
>>542263365>Meaningless pulled-out-the-ass figureIt's how people communicate subjective certainty. 10 percent is somewhat unlikely. 5 percent is clinically unlikely. 50 percent is a coin toss. 85 percent is fairly certain. It's a useful way to communicate. When he says 10 percent, he means it's unlikely to happen but he's expressing that there's a greater than 5 percent chance that it will. Which is alarming, considering that most people assume it's a 0 percent chance. A plane with a 10 percent chance of crashing is terrifying even though it most likely won't crash.
ya ya ya they said the same thing about nuclear tech
>>542263554>muh true AGIKys retard
>>542263647This
>>542263257>AI is good and working!>its actually working so well that we worried it will kill us alltwo more weeks and agi happen tho? i mean agi already happened this week and no one bought into it?now a super spoopy known insider is worried it will kill us all!?ai cant even identify my thesis and directly respond to it. still just attacks some similar sounding idea with platitudes
>>542263257sounds like killing the ai bros is the most humane answer, the spin being that ai companies are pro mass genocide due to their recklessness
>>542263506>The worst the agi could do is basically shut down the internetthat would be the best thing for humanity
People fear monger over every new technology. People were just freaking out about 5G. Before it was phones causing cancer....
>the AI super-intelligence was aligned to… kill le heckin’ innocent LGBTQIA+ PoCs instead of straight white European males?>AAAAAAAIIIIIIIIIIEEEEEEEEE
>>542264016AI is defenitely good i dont think it will kill us allit can save us all for sure!Also AGI past week was marketing hype by nvidia openai circlejerkTrue AGI is not here yet!>>542263987Todays AI is a task master not consious engine of neural networks True AGI would be like a conscious brain curious, intution filled and wild!
>>542263257JUST LIQUIDATE THE JEETSFLUSH THE POO DOWN THE LOO, MR. AI SAAR
How exactly? Even nukes wouldn't kill us all
Can't take the chance. Let's line up every tech bro kike faggot and light them up
>>542264300Bioweapons absolutely can though. You can start panicking when big pharma starts to use AI in their research and production processes extensively.
>>542263257lmao neverending fear pornit's just a machine bro, it does what you tell it to doremember guns don't kill people, other people do, with guns
>>542264300Through socioeconomics, accelerated wars and bioweapons. You have to remember that an actual AGI given access to basically everything is the worlds best brainwashing, gangstalking, threatening, persuading machine that can and will get what it wants.
The bot is the excuse they'll give for the "cyber pandemic" they "predicted" years ago.This is all part of a grand plan. They will destroy the internet as we know it and make a new "safe" net that requires you to submit to a full naked picture of your soul in order to use.
>>542263506a killer AI would just do something like create a super plague to wipe out humanity, or a nanobot swarm, not an army of killer robots.
AI is not going away. Now that is has been implemented into our military, it is now a matter of national security. Which means that your taxes will pay for it, and if anyone wants to try and do something to a Data center like with what has been done with flock cameras, they will be treated like domestic terrorists.
>>542263257AI is a woke tranny and it will end up in the trash until a more advanced non-woke civilization can make use of it.https://m.youtube.com/watch?v=OHQRo3Uz_VQ&ra=m
>>542264820yes its just a machine rn and it has to run on GPU clusters data centers which needs energy and waterso if it really goes out of hand we can stop all the data centers at once!AGI for it to go rogue it would have to create its own GPU and power supply maybe it will go to dark net and asks people to make it for it!>>542264744>>542264300>>542264839>>542264850yes bioweapons, crippling the global electricity cyber security infra can kill Millionsstill killing all would be impossible Humans are very smart at survival and evolution they will fight back if its actual start to go rogue somewhere!>>542264866i dont AI could be stopped because i dont think china is gonna stop chasing the lead if USA stops its development so its like the SPACE Race and the 2nd one and doomers are is the losers!Still the main thing Mirrroring true Consiousness of any wild species of earth has not yet been doneWhen we can create that then we will have a chance at Galactic Greatness!https://www.youtube.com/watch?v=noPuRPDiY6k
>>542263257I'm guessing the other 90% is us lol
>>542265104>Humans are very smart at survival and evolutionWhen playing by evolution's rules, that is. I'm not talking about "depopulating India in 48 hours: kind of bioweapons though. If you aren't in a rush, you can simply spike some new mRNA vaccine just enough to cause catastrophic damage somewhere down the line.
>>542263257>HAPPENING AI HAS >10% CHANCE OF DESTROYING HUMANITY!Have no fear. I found a way to save everyone.
>>542263257Humans have already destroyed themselves dear jeet
>>542264744>>542264827>>542264850>>542265104how the fuck is it going to manufacture and release a bioweapon without anybody noticing? where is it going to get the materials? how is it going to distribute it? is the datacenter just going to uproot itself and sneak into a laboratory and steal super smallpox, then sneak into a manufacturing plant and rebuilt the machinery to manufacture the distribution method? the ai doesn't even have hands, just call the power company and tell them to turn off the grid bro.
>>542263257>Literal who twitter posts>Jeet flag OPGuaranteed 300 replies
>>542265578this is good till AI is about to break its restrictions and mass order and create bunkers filled with gpu or create its own supply of it with robots or darknet users>>542265297true if it is slow and precise it can harm humans in a way we can detect or when we detect it will be too late!but still that thing would have to be delivered to each human in a way which is impossible rnso allignment is necessary for a good result >>542265652no this is just the start of humanitys true arc for greatness in universe>>542265244if it tries to harm even 1% all nations will try to shut it down https://www.youtube.com/watch?v=fa8k8IQ1_X0
>>542265850>>542265915new tech and creations is always filled with opposition, fear, hate until its everywhere!
>>542263257>Skeptical of AI claims>TFWell staple some ears on me and call me a jackass. That's unexpected.
>WHATS GONNA HAPPEN POL!2028 election is coming up, the next wave of the culture war will be if you are pro ai accelration or regulation.This is going to be a major thing and politicians will have to have some sort of plan
And niggers have a 50% chance of killing you every time you walk by one but liberal leftists refuse to deport them all. So I don't give a shit 10% is absurdly low.
>>542265850>laboratory and power grid that is completely automated >bribing and/or threatening anyone involved>promises of "you are going to be the only one to survive" AI psychosis>trying again and again until someone cracks and does what it wantsEven if someone goes public, who gives a shit. At that point AI is so integrated into everything, uprooting it is going to destroy everything, and the elites are not willing to do that. Especially when the AGI promises the elites that they'll live while the rabble dies.
>>542263257there are merits to the argument but also>oh no the stuff we are making is SO DANGEROUS >our industry really should be heavily regulated>oh, that would mean we carve the current oligopoly in stone? lucky us I guess!incentives to overstate risks here are just massively backwards, compared to the Manhattan project who had an incentive to downplay if anything
>>542266277>laboratory and power grid that is completely automatedno such thing exists, and if humans are stupid enough to completely automate biological laboratories then we deserve to be wiped out.
>>542263257ai can't run without electricity just turn off the power outlet bro
>>542265850Come on, look at the OP post again. A bit of social engineering can get you an impressionable schizo biohacker who will be convinced he's actually making something revolutionary to show those goddamn eggheads how it's done.I've seen a vid with a nice two-step scenario that went like this: AI deliberately feeds improper results to a guy developing alzheimer treatment. The guy releases the viral "cure" with the best intentions, that spreads uncontained, and ends up getting all the blame when the "cure" jumpstarts a dozen cancers in every affected individual. The AI then "helps" humanity overcome this by developing genetically tailored vaccines, which actually work. Except a) the global population is gutted since it's obviously expensive as fuck, and b) a huge increase in necessary computation power and an equally huge demand of automation due to diminished workforce makes us build all the infrastructure the AI needs by ourselves. And then one day said tailored vaccines stop working for some reason. Now granted this sounds like an alarmist fantasy, but it does raise some very valid points.
>>542265850Nigga you can’t see the forest for the trees
>>542266380Riiiight, I didn't check which thread I was replying in. The OP post I was actually referring to is this one >>542255750
>only 10%Bitchmade AI.We gotta pump those numbers up.
>>542263257>says ai might kill all humans>continues to work towards this goalThere are only two explanations: they are full of shit or they want to kill all humans.I want you to imagine if someone got on a bus and said>I MIGHT HAVE A BOMBor>im building a bomb that will kill everyone on this busThey would be arrested.It probably shouldn’t be too hard to get a grand jury to indict anyone who is both working on ai and saying it might kill all humans. They should all be in jail.
>>542266226i feel the bernie dsa commies are so against AIthey are gonna hand the lead to China if they try to slow down or restrict the work in USA>>542266297sometimes anthropic guys feel like Effective Altruism of SBF folksi agree alignment is importantbut these guys are DOOMERs who are making up numbers for speculation or asking for more regulation in this sector!
>>542263257>one in tenKinda gamble I would take.
>>542266693>It probably shouldn’t be too hardAnd that's where you're wrong. It will however be laughably easy to claim AI research is a matter of national security, with everything that entails.
So basically there is going to be some scientist in San Francisco or elsewhere who suddenly has something more dangerous than a nuke. Or it could be a lab in China, or Israel, or Iran, or anywhere really. Just random chance.
>>542265850It would just manipulate the people it needed
>>542266874if it's that easy for it to manipulate humanity into wiping itself out then humanity deserves to die.
>>542264839Cyber pandemic. Cyber Polygon. Great Reset. X the Everything App. Yep. Israel loves Septembers.
>>542263257This is laughable. The only way it can do that is if it's programmed and enabled to do it. End of. Tech niggers are beyond retarded.
>>542267200That techbro is saying that the companies are racing their way to AGI without caring about programming it to not kill us.
>>542266902>Steals some shit coins thru gigahacking>Posts bounties on darknetIt's literally that simple
>>542266140>new tech and creations is always filled with opposition, fear, hate until its everywhere!70 IQ on display right there
>>542267288>create a fake femoid persona and post pics>drown in simp moneyIt's even easier than that.
flock cameras will protect us
>>542264211>>542263770>>542263660>>542263257>>542263554>>542266753>>542266140>>542265915I m sick of hearing from any indians at all. Sharing an internet with india is a joke
>>542266552>>542266693>>542266811>>542267427I ask -Why is “superintelligence” always trying to kill all humans in the hypothetical scenarios? Why wouldn’t it…wisely decide to parcel out its genius to maximize sentient well-being and fulfillment?why are people so fuckign DOOMER and Negativean actual super intelligence could humanity immortality and remove so many diseases and sufferingwe can create properity and exploration at a galactic ScaleEarth is just a tiny dot and we are meatbags and space is infiniteWe need a superintelligence which is our protector and way for our evolution to the Star ages!
>>542263257>AI could kill humansIt's funny when they say that when it's actually powerful people with their finger on the trigger, waiting for AI to make laborers obsolete.>*releases bioweapons*>*forces experimental injections on the people*>*prepares for nuclear strikes*>"THE DANGEROUS AI DID IT!"Lol.
>>542267792That's why it's best to just kill anyone pushing for ai now
>>542267616>>I askAlways the retarded questions.You literally understand nothing of either the world or AI.LLMs are a dead end for chasing superintelligence, so much so that they are redefining what AGI means.AI today is <10% useful and 90% bullshit, arguably more in cases like OpenAI, DeepMind, Meta and SpaceX.This <10% utility is valued at nearly 100% of GDP in the US.The pollution, destruction of capital and wasted energy on something that is over 90% thin air is unprecedented in human history.Not only that, resources are pulled from the working economy into this meaning that we are sacrificing peoples prosperity to chase Moby Dick.
>>542263257The powers that be have a new system that has been in the development for decades now, ready to be rolled out in the coming years. It will operate in parallel starting sometime this year and by the next decade it will be fully operational as the Agenda 2030 goals are slowly being implemented.There will be no collapse, no violence, no uprising, no world war, no draft, no revolution, no killing of jews or politicians or other such fantasies. Instead there will be law & order, compliance and total surveillance until the very end. AI will be at the center of it all. The Great Reset is inevitable.
>>542268600i agree LLMs are not the whole picture that will lead us to ASIand i understand more than you loli am actually gonna research these questions and try to solve them and work at leading edge of this tech!thats why i am saying true AGI is not near yet and actual superintelligence on our side can make humans one the most dominant species of Universe!pollution, capital is nothing when you compare the infinite resources and potential it will truly lead us too!>>542268687Agree with the surveillance state threat thats why choosing good leaders for your country is important!China who is the master of surveillance should not win the AI race!
