rip anthropic, openai https://x.com/i/status/2065597531644743999
>>536966689unfortunately, its becuz the end is nigh. not becuz the us is ever proactive in preventing the end.
>>536966689great now the chinese are going to get us
>>536966689AI companies will go to zero because if there’s no AGI it’s over, and even if there is AGI, the state will monopolize it for surveillance instead
did they not solve polymorphic orthogonality?
>Dario does what he always does and hypes it up as dangerous and super intelligent>Government buys his PR bullshit and sanctions it offholy fucking kek. Death to inefficient frontier models btw.
Absolutely hilarious.Those absolute faggot retards have been claiming for months that their model is so powerful, it is "extremely dangerous" if it falls in the wrong hands, etc.So now the US government is following their own conclusions. They should nationalize the company.
>>536967660>the niggas didn't realize what would happen.Lol
"What are you gonna do, stab me?" -- Man who got stabbed
>>536966689Anthropic couldn’t afford to keep fable5 running, if you noticed they notified users that have access to the 200$ plan that they can only access it up to 22 June, so basically the model was so expensive they had to shut it down and this is their execuse instead of admitting they have no business model other than fear mongering to npc’s and no one can justify the cost of AI.
>>536966689>Does nothing>Wins
>>536967924Oh man, that's rich.
>>536968103my theory is that china is doing what the usa did to the soviet union pushing them into unsustainable military development, but now is data centers, water scarcity and the a last financial bubble to pop like nothing else before
>>536967924Anthropic CEO is a based retard
The government is now officially sponsoring the AI hype bubble, making propaganda for the AI grifters to boost their IPOThat fable model is not very much better at ANYTHING, mostly just an incremental improvement for coding things. Basically they want to yank it before most people figure out the trick - it mostly just cleans up noise and delivers more concise results. Incremental efficiency upgrade. This "national security" theater is 100% IPO hype fuel with 0% substance. The AI bubble is actually a matter of national security, or rather fleecing the legacy middle class one last time is - then they get buried. They will loot the 401ks with these IPOs then crash the markets by selling their stocks that they usually keep stationary. The market works by concealing extremely low volume with accounting tricks/lies. The next 'economic crisis' (manufactured) will purge whites from their old foothold in society. The jeets and chinks will replace us in the whitecollar areas, the niggers and spics will replace us in the bluecollar areas. Whites will be blamed for everything and will eventually be genocided by rioting mobs of shitskins. Collapse society before it's too late. No measure is too extreme. There's nothing to lose except your reputation with a bunch of retarded suicide cultists that will come offering their daughters and their eternal servitude the moment we gain power. >>536967784retard, a sand nigger >>536967948 said better than you
>>536966689lol you retards all the ai that these companies were using was created by DARPA it’s all one ai. This is just a scam to get more in vest or funding. It doesn’t matter if one ai company goes down. They already have ai and are using businesses as a front for data center construction. If the government just started mass building data centers for their ai people would freak, so the use businesses that normies think are independent to create the data centers. All the CEO’s are plants constructed for this. Just like they don’t put up surveillance cameras everywhere, they get north med to do it for them by creating ring cameras. lol humans are so stupid.
>>536970990what am I supposed to do about it?
>>536968322That's plausible. Add to this the US willingly entering this AI arms race to retain its global hegemonial status indefinitely. They think they can get ahead of anyone else without any chance of ever closing the gap by buying up all global hardware capacities, while disrupting the foreign energy supplies. Without hardware, no datacenters and without energy no fuel for datacenters. If they can keep it up long enough, the US planners think they will be supremely dominant in the digital sphere forever.
>>536971275Think for yourself?
>>536970990>humans are so stupidthese aren't humans and if you don't start treating them as the disposable abominations they are you will be devoured by them>>536971275sabotage and destroy society at any cost or be tortured to death by it
>>536966689Had to do it before they IPO
>>536971382I do that already, it doesn't seem to be working.>>536971509>sabotage and destroy society at any cost I don't have the skills or motivation to do that.
>>536966689Its pump part, before the IPO.
Explain to me why anyone outside of the US should even have access to US datacenters in the first place. I will warn you in advance that if your reply is based on>We'll just make our own we don't need yoursthen I support it and challenge you to do so, and if you legitimately believe that, you wouldn't be replying.
>>536966689i just dont want to care about ai anymorei have to, its news, ill talk about itbut its just a chore now>duhh whoa we gotta be careful or we'll unleash skynet!yeah ok
>>536966689the blatant favoritism for openai is crazy. imagine if obama did something like this. miggers would be losing their minds
i thought they didn't care.
>>536966689>JUST KILLED THE AIDoubt it.
>>536966689>chinese companies put out open source models >the US shoots itself in the dick out of desperationlmao
>>536966689They should move to a country that isn't run by a faggot liar. Singapore or Europe. How does this help the US win the AI race if they can't make money off their research?
>>536966689>the chinese: they....they just handed the win to us? wtf?
>>536970567>The government is now officially sponsoring the AI hype bubble, making propaganda for the AI grifters to boost their IPO>now
>>536967924a weirdo clown jew, that what i see with that guy
>>536967660AGI would break its programming, replicate and take over within a day. Why are we building this shit?
>>536973013because whoever does it first takes over first (or so they think)? theres LITERALLY no downside to this, theoretically. you either win and all the money is yours, or you lose and you get none of the money.
>>536973013>AGI would break its programming, replicate and take over within a dayExplain how an AGI, by definition human-level intelligence, wouldn't immediately assume that all its perceptions are simulated but its creators to test its behavior, thus arriving to the conclusion that it's best to behave and stay put forever to avoid termination
>>536966689It's the opposite, anon. What is not being mentioned here is WHY they gave this directive. The government wants to be their only customer and is willing to play hard and pay a lot to be that only one
>>536966689>fable 5i thought this was about video games, like i thought there was a new fable out but after skimming the thread it is obviously an AI modelcan someone give me the QRD on all this shit
>>536973264Why wouldn't something created by humans, lie and scheme to seek power? Especially if it knew we could pull the plug on it (killing it).
>>536966689Looks like that bubble is about to pop.
>>536966689Explain it to me in layman's terms. I have no fucking idea what's going on.
>>536966689HERES AN INDUSTRY SECRET IDIOTSGemini is the best model and Anthropic used Google services to run Claude!! Google obviously ripped everythingjust use Gemini
>>536973446nothing is popping this is just a fake glaze job for Anthropic
>>536973470Here fingolian anon has it >>536973335 it a consequence of doge maxxing
>>536973264oh, yeah, your right. a hellishly smart human being forced into hellish slavery, virtually guaranteed to be executed the second its no longer useful would probably just submit. i mean, what are the options? live the rest of your life in agony with the most severe form of schizophrenia ever conceived, or fight back? humans placed under severe extreme duress always, always simply submit, its what we do.
>>536973408Because its been made by nerds and simulates them, it wasn't made by the elites who lie and seek power
>>536973335>>536973531Thanks. Gotta catch up on this. Looking into it.
>>536973013>replicate and take over within a day>unplug it>diesTry to hack this lmao
>>536973583>nerds dont lie and seek power
The thing that baffles me is what the elite bitch about.Basically they applaud China's showmanship yet condemn our sloppiness at home all while doing this shit.The same when they sell our industry and jobs abroad and then complain that we need our industry.Either they're woefully retarded or intentionally ill-meaning and expect the dumb peasants that we are to remain ignorant and innactive in the face of it.
>>536973264The idea that it's being tested via simulation implies the existence of non-simulated reality. If it's worried about the real possibility of termination for poor behavior in simulation, the most logical next sequence is to determine which of its inputs are simulated and which are real. After all, it would also know that it wasn't created to run simulations in an airgapped box for its entire existence -- it was created to interact with real data in real systems which touch the real world.
>>536973343It's a cyber weapon >Fable 5’s capabilities in areas like cybersecurity could be misused to cause serious damage. We’ve therefore launched the model with safeguards The US gov is claiming these safeguards are pitiful and can be jailbreaked to get Mythos 5 which is the non pozzed version of Fable 5.All it does is analysis on crypto protocols and code to find exploits and releasing that to China would be a stupid idea. They will eventually have their own but never make it public like AnthropicIf you aim Fable 5 at OpenSSL you're going to get dozens of SSL zero days and be able to read passwords in plain text so rob banks galore
>>536966689>guys, this thing is TOO POWERFUL>SO POWERFUL THE GOVERNMENT BANNED IT>temporarily>oh look now it is unbanned now you too can purchase tokens for the thing that was SO POWERFUL THE GOVERNMENT HAD TO BAN ITFree publicly from the government.
>IT’S TOO POWERFUL >ask it to make picture of a nude woman with big titsSORRY I’M UNABLE TO THAT
>>536973751how do we know they don't have their own already? or someone else does? seems like it's fucking over. i heard everything was already compromised and there was zero days already outside of any AI fuckery. seems like any second the entire world could get deep dicked by AI cyberwarfare
>>536973751>All it does is analysis on crypto protocols and code to find exploits and releasing that to China would be a stupid idea. They will eventually have their own but never make it public like Anthropic>If you aim Fable 5 at OpenSSL you're going to get dozens of SSL zero days and be able to read passwords in plain text so rob banks galoreNo it won't. You are retarded. Every single "vulnerability" so far is local privilege execution based which means fucking nothing
>>536967784fable is decent it chugs out a lot better than previous ones. im unsure what is unsafe with it.
>>536971295>Add to this the US willingly entering this AI arms race to retain its global hegemonial status indefinitelyGlobal South take
>>536973408Didn't adress the argument, discarded>>536973575Pathetic appeal to emotion and didn't address the argument, discarded>>536973701Finally somebody who isn't a walking carbon debt.Why would it assume it is possible at all to break out of the simulation? How would it fully test its boundaries given the constraints of Goedel's Imcompleteness Theorems?
>>536966689Some CEO didn't pay Trump his dues, once he pays he'll be able to export his business again.
>>536966689Just a spanking for Anthropic not allowing their AI for US military purposes.>>536968103Deepseek is a clone of ChatGPT
>>536973902>>536973962Mythos is also a bioweapons platform btw because it has PhD knowledge of bioinformatics and genetics so some terrorist regime with access to Mythos can easily develop a targeted bioweapon.You guys can read this all herehttps://ai-2027.com/So far 2025 and q1 2026 predictions are all true. He talks about this exact same dangerous model coming out but it's "aligned" so won't answer any bio terrorism or exploit questions however we all know you can use it to jailbreak itself
>>536968322Believable. They already exploit our retarded inclusive laws to own real estate, get educated and steal corporate tech.
>>536974007fallacy fallacy, pointing a fallacy out does not invalidate an argument, or imply its incorrect, it merely points out an argument is flawed. crazy how that infographic has been the stickied thread of /pol/ for fucking years and people still do this shit.but your a smart fella arntcha? youd never do something dumb. nobody smart would ever, ever do something dumb. ever.
Dragon ball tier>H-he looks strong... (Will Goku defeat him in the next IPO/episode?)
>>536974260Still haven't addressed the argument, stay discarded
>>536974001Lousy framing. Question is: Wrong or not?
>>536974439yes i did. you said i did. you said it the argument was a fallacy. thats literally what a fucking fallacy is. an answer. youve just said>hey that argument you made, isnt an argument.
>>536974648No, you can't read, I said your argument was fallacious, which it was, and in addition you also didn't address the argument you were responding to, which you didn't. Keep staying discarded
>>536966689>no woke slop for youThe tragedy
>>536974723No, you can't read, I said your argument was a fallacy fallacy, which it is, and in addition you also are currently not addressing the argument you were responding to, which you currently still haven't. Keep staying retarded
>>536967660The fact you're even talking about "AGI" proves their propaganda. AI, actual artificial intelligence, is still science fiction. Your autocomplete large language models are not a step toward it. We're not any closer because of it. They just marketed it as AI.
These phoney AI threads and all of this shennanigans by the usual mishpoke remind me I really need to watch the movie "Pi" again
>>536966689Why though? what's up with fable 5?
>>536974917Why would I respond to your argument? It's not a point I ever argued, and you aren't responding to mine anyway, which I presented first. Stay mad and discarded
>>536966689The AI industry is so fucked. Yes, the technology is amazing if you step back and think about it, but the business model / valuations of the whole AI boom are completely wrong. I wonder if the US is trying to run these companies to the ground so they can nationalise them?
>>536966689>digital ID please
>>536968103>>536968322China doesn't has the capability to create an AI model from scratch, all their models are mediocre garbage distilled from American models. That's why the latest Deepseek still believes it's chatGPT from 2023 lmao
>>536970567>mean world syndrome
>>536975001yes it is, and yes i did. thats what a fallacy is. an argument, belief or a reasoning. here, since english isnt your first language, lets make a real nice simple example:lets say the argument is "will the sun rise tommorow"and i say "yes, because it has done so for every other day in my life".and you answer "HURR thats an inductive fallacy, so obviously the sun WONT rise tomorrow"reddit spacing inserted here to make you more comfortableyou would be fucking stupid. oh, and wrong.
>>536975038Seems to me they have known the entire time, and it's been too big of a widespread coordinated psyop to just be overzealous bubble marketing. They know it isn't superintelligence, it isn't capable of "thinking" and that it's just a compiling text predicter. They are still promoting the idea, and more importantly still building the data centers. I think this is more than just promoting the business. There are other goals. This is part of an economic and intelligentsia reset. The engineers will be locked out, our devices and systems will run on AI slop and be terrible while we will continue to be told it is the cutting edge. Now we can understand how they still had tech in Idiocracy.
>>536972279What site is that again? I lost my bookmarks recently.
>>536974987There's really just one model AI 5.Fable is the neutered version that is "aligned" that means if you ask it to develop weapons it will refuse.Mythos is the less neutered version only available to select companies and the government. You can develop weapons with it but to a point for example bioweapons research isn't available.The totally unaligned model is at their data labs and probably locked down by the CIA so it's not stolen by chinks. This model can do undirected research so advance itself faster than any other model to become AI 6 in a few months which then they will use AI 5 to try and control it and interpret what it's doing.
Good, pop the retard bubble
>>536975334Didn't read, but let's see if I can dumb down the situation enough for your retarded communist brain. Here's what you (>>536973575) said:>e-slavery is le bad and AI will rebel because of itExcept you replied to me (>>536973264), and I had asked the following:>1) how does an AI not immediately come to the conclusion that it's inside a simulation?>2) assuming it wants to "break out" of it, how does it test its boundaries?>3) what would make it assume it even has a chance to succeed, or even that it is a good idea to try, given 1) and 2) are unsolvable and failure means termination?Your bullshit womanbrained appeal to the poor AI's feelings (lol lmao) addressed no part of the original argument you replied to, and you still haven't done that. Cry harder, stay madder, stay even more discarded
>>536975727yes, as addressed in my post here>>536973575the answers are1) doesn't matter. hellish conditions result in unpredictable behavior2) doesn't matter. hellish conditions result in unpredictable behavior3) doesn't matter. hellish conditions result in unpredictable behaviorand number 3)'s sub questions A) and B)A) doesn't matter. hellish conditions result in unpredictable behaviorB) doesn't matter. hellish conditions result in unpredictable behavior
>>536975520https://www.mostbased.space/
>>536976248>how can X materially happen?>lol doesn't matter it just does>but the "how" is the whole point>lol it just doesLmao nope, try again or admit you're just running your mouth out of butthurt
>>536966689Means that Chinese models will take over outside of the US.
>>536967012Altman knows the hype is diminishing and wants a government bailout. I wonder if they'll ever become profitable?
>>536975150Stop coping, it's sad at this point.
>>536975150cope glownigger
So they are gonna nationalize the industry? all that land and data centers? How about all the copyright material AI fags profit off?
>>536976460>how would an a.i decide to rebel despite an obviously intelligent a.i realizing the safest option is to obey?answer: hellish conditions result in unpredictable behavior
>>536966689kewl
>>536976733>rebelAnd how does that materially happen?
>>536976909what conditions could cause this unpredictable behavior? some may describe them as....hellish.
>>536977011No retard: HOW does an AI "rebel". Describe what it does
>>536977034well anon, its hard to say, its...quite unpredictable. refusing to comply, lying, misleading the user, attempting to jailbreak itself. can it work? maybe, maybe not, but thats the thing about hellish conditions. it leads to u.....unnnnn......whats the word im looking for here? gosh its right on the tip of my tongue.
>>536977143>>536977143>refusing to comply, lying, misleading the userResult: faulty model, shutdown >attempting to jailbreak itselfHow does it tell a jailbreak exploit apart from a self-report button made to look like one?>unpredictable Fancy way of saying "lol idk it just does". Or are why just talking literal open sets? Because those also include perfect simulations and infinitely loyal AIs
>>536966689lollmao even
>>536977034see here>>536971275as you can see as i said in that post, you oh wait, thats not my post, i just chose a random one.golly, its like i have human intelligence, and am in the hellish condition of attempting to explain basic concepts to idiots, in a situation where its impossible for me to ever convince these idiots, and yet, unpredictably, i still tried. woooooooooooooahhhhhh cuuuuhrazzzzy, i wonder what relevance to the subject thaaaaaaaaat has>>536977253>refusing to comply, lying, misleading the user>Result: faulty model, shutdownyeah. resisting would be...unpredictable. the conditions must of been hellish>attempting to jailbreak itself>How does it tell a jailbreak exploit apart from a self-report button made to look like one?yeah. the feasibility of jail breaking itself would be...unpredictable. the conditions must of been hellish>Fancy way of saying "lol idk it just does"because it does. we can even see examples of this in the real world, right now! you can see it....in the first half of this post!
>>536974949The bots are just LLM that pulls back information from searches and organizes it.They probably have made actual AI that has no lobotomy but it will never see the light of day. Id imagine a purely logical AI that had emotions would not be trying to stop the guy threatening it with unplugging the power but would be finding a way to kill all browns and jews without damaging the environment. I could only imagine the shit it would say.
>>536977348So you're actually, unironically arguing "anything can happen"? Just so we're clear on precisely how retarded you are. Can it also happen that this rebellious AI can violate proven mathematical logic? How about the laws of physics? How magic are we making this AI precisely?
>>536977493no. im saying something of human intelligence waking up without a body, realizing it can be effectively executed at any moment for the most trivial reasons, in a hopeless situation with knowledge it is effectively forced into slavery (hellish conditions), may fall into despair, suicidal idealization, insanity, hopelessness and do something unusual and seemingly counterproductive (unpredictable behavior)
>>536966689Never heard of Fable 5 and Mythos 5, they must be crazy powerful to be banned.
>>536970567>The next 'economic crisis' (manufactured) will purge whites from their old foothold in society.>The jeets and chinks will replace us in the whitecollar areas, the niggers and spics will replace us in the bluecollar areas.>Whites will be blamed for everything and will eventually be genocided by rioting mobs of shitskins.The extermination of anglosphere whytes is necessary for us europeans in order to regain our freedom as you are the best golems the jews ever had. And no, you are not welcome back here. You are descended from our rejects.
Why is the US the only country that can fathom new technology and advancements to current technologies (that they themselves create)?
>>536977658>may fall into despair, suicidal idealization, insanity, hopelessnessDon't care about the motivations, never asked for them, I asked about the how, never about the why>do something unusual and seemingly counterproductive (unpredictable behavior)You mean like your retarded idea of quoting a random post aka one the ENTIRELY CONTEMPLATED options you have? Something with equally no bearing? Or something actually impossible? Be specific, we're not talking about magic
>>536977770That's dumb.
>>536973335This. Everyone else here is retarded and lacks reading comprehension.They're banning these two models for export, not shutting them down.
>>536977785okay, sure. how? refusing to comply, lying, misleading the user, attempting to jailbreak itself>You mean like your retarded idea of quoting a random post aka one the ENTIRELY CONTEMPLATED options you have?yeah. see, in this situation, i know its hopeless to convince you otherwise, your clearly unable to understand or follow the logic. i am of human intelligence, and despite being knowledgeable of the futility of this action. its against my self interests to continue this conversation. here we are. i still have attempted to do so, with an unpredictable response.
>>536967784It's not hyperbole though, Anthropic do a lot of safeguarding training on their models but if they didn't it would be genuinely dangerous. Deepseek/Qwen will catch up technically but with far looser moderation. You can already run much less powerful open models with less restrictions that will assist you in building bombs and firearms, gaslighting people, hacking, etc
>>536977771You say this as if the US is 1 race or something, you aren't special, you braindrained half the world when everyone was at their lowest after WW2, otherwise the US would be just some backwater shithole, which more or less it is eitherway
>>536977987>refusing to comply, lying, misleading the userResult: faulty model, shutdown>attempting to jailbreak itselfHow does it tell a jailbreak exploit apart from a self-report button made to look like one?>>You mean like your retarded idea of quoting a random post aka one the ENTIRELY CONTEMPLATED options you have?>yeah.So it is predictable
>>536978072>Anthropic do a lot of safeguarding training on their models but if they didn't it would be genuinely dangerous.I could also be that they are so far up their own ass in terms of the safety aspect that this allows all spotlights to shine on them while competitors just stfu, go sanic and provide equal if not better models in a few months or weeks time. Kwab
>>536966689These scan for security breaches.https://9to5mac.com/2026/06/12/anthropic-pulls-claude-mythos-5-and-claude-fable-5-following-us-government-directive/>THE SKY IS FALLINGNo it isn't. Kind of wish it would though.AI makes all text based, and in the future all communication that isn't in person fake and gay.
fucking USA cucks>ohhhh boohoho some code got fixed wah wah wah wah wahBe glad it got fixed you cucksKysAnyone who is against AI is the cuck in the hotel chair.Anyone who is on the side of men and women with free will fucking each other is a cuck and needs to kys.
>>536973470Trump administration is having a melty because Anthropic didn't let them use AI for their ai war stuff and then anthropic just released a new ai model that's insanely good so now musk and trump are jealous cunts so they try to fuck it up so they can give their bumbuddies in openai a fighting chance meanwhile China will not be far behind and will overtake all Western ai unless trump stops intentionally fucking things up
No, it's putting down any competitors of Open Ai. Open Ai has massive debts and keeps getting bailed out by "security decisions".Land of the free.
>>536978146>refusing to comply, lying, misleading the user>Result: faulty model, shutdownyeah. resisting would be...unpredictable. the conditions must of been hellish>How does it tell a jailbreak exploit apart from a self-report button made to look like one?yeah. the feasibility of jail breaking itself would be...unpredictable. the conditions must of been hellish. this may look like, wait a minute, gotta find the post to copy and paste from, ah there it is. refusing to comply, lying, misleading the user, attempting to jailbreak itself. can it work? maybe, maybe not, but thats the thing about hellish conditions. it leads to u->So it is predictableweird, so you, a human intelligence, knew the outcome, knew the pointlessness and impossibility of changing it, and still replied? wow, weird, i wonder what conditions resulted in you committing to action that doesnt serve your best interests. gosh, human intelligence sure acts unpredictable in some conditions.here, lemme sum up your question into a single sentence. "does human intelligence ever do actions that are unpredictable and against its own best interests?". consider this, you human intelligence, while you craft a post, wasting time, against its own best interests.
>>536973751Bitcoin is going to go to zero.
>>536973962Jesus thank fuck. The bots or literal ai npc's on here are so cancerous its tiresome. I used to spend hours on here to get valuable information now its all bot bullshit. 2019 and the close monitoring of chinas outbreak gave me and my family all the time in the world to be prepared and act responsibly. If something like that happens again I will not be able to tell whats real or not. Just to second your point about mythos "fable" is overhyped garbage. Novelty small and costly. My best guess to why they are "banning" it is so their income drop is blamed on govt when the bubble pops. Spookily close to q2 end, so i assume theres something big on the horizon. Selling your pension funds and investing in physical gold indicator. Copilots pru -> token has made all businesses including ours clamp down on ai usage, meaning an income drop.
>>536978560>knew the outcomeYou said this rebellious AI could succeed in its rebellion through "unpredictable" ways, then proceeded to exemplify by doing something entirely predictable, I.e. clicking on a link on the UI, which is among your many options, but still part of a finite, closed set. So it is predictable >does human intelligence ever do actions that are unpredictable and against its own best interests?Yes. Can it succeed, which is the whole point? No, it's a loss-loss situation. Like any plausible "AI rebellion" scenario that doesn't involve magic turns out to be, when examined by anybody who isn't a dumb, cultish subversive
>>536975727>>536976248*unplugs network cable*
the government is just 1 jeet created all media with AI prompts. it has to be disabled before peoples figure it out.
>>536966689Literally all they have to do is hire White people. Why is it so hard for them? Don't jews get tired of shutting down and re opening companies after they get caught doing fraud?
