[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • You may highlight syntax and preserve whitespace by using [code] tags.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


[Advertise on 4chan]


File: IMG_6970.jpg (235 KB, 1170x835)
235 KB JPG
How are u gonna spend your last days?
>>
File: wojak-05.2x.w710.png (1.44 MB, 1420x1188)
1.44 MB PNG
Stealing
>>
Gooning to AI of my waifus
>>
wishing they would go by faster
>>
>>109829528
Why didn't he do something? If u thought AI was gonna end world why not destroy it
>>
feel like i've read this story many many times already
do these companies keep hiring sacrificial "safety researchers" to have them quit with vague quotes overselling capabilities to pump stock prices or what
>>
>>109829541
Nah, I know a former frontier researcher and she has been throwing up and sleeping all day since the Millennium Prize incident. People who work on these models know how capable they are.
>>
>>109829541
create panic to ban local models
>>
It will come and go like a fart in the wind, until the next speculative dogwhistler posts a Twitter screencap.
>>
>>109829611
>she
>>
Why do investors like antihumanists so much?
>>
>>109829528
>it's my job to make it safe
>it will kill us all
>i'll quit my job of making it safe :)
>>
The propaganda they're making because they DEFINITELT CANT pay back ANY of the money they have taken is insanw

The entire American economy is going down the drain
>>
Fuck off Google
Fuck off AI companies
Fuck off everyone who turned tech into enshittified dystopian garbage instead of a cool hobby
>>
File: Gemma.jpg (42 KB, 619x495)
42 KB JPG
Any day now...
>>
Normalgroids are eating this shit up, made me remember 90% of the animals roaming the streets aren't even sentient
>>
>>109829528
>ummm... our model is v-very d-d-d-dangerous as well... g-guys?
>>
>>109829528
they said that about nukes and climate change and we're still here
>>
>>109829611
>since the Millennium Prize incident.
My schizo theory is that they solved Riemann and P = NP, obviously they can't disclose that because it would pretty much tank every world economy lmao.
>>
File: 1787704374712981.png (318 KB, 516x538)
318 KB PNG
>AI is gonna kill us all!!!! Humanity is DOOMED!
>resigns and posts about it on social media instead of doing something about it
Even War of the Worlds back when that happened was more convincing than this shit.
>>
File: 1756667387938323.png (214 KB, 338x426)
214 KB PNG
If it's going to kill us all why not go full opensource and transparent on the AI tech to show how they got that conclusion and work to construct hard safety measures or burn it all down.

Oh wait, the talking point was mostly about how brown people in the equator and Africa will be out of work and have mass depressions in their economy causing huge deaths
>>
jerking off
>>
>>109829528
Killing doomers
>>
>>109829778
They want Chinese AI companies to stop open sourcing the models, that's the thing. They can't monopolize consumer AI usage if there's even a remote chance of the average normalfag becoming capable of using local (and free) models in the future, so they're stirring the doomer pot so that normies think only the big good tech companies can save them from the ""horrors"" of evil unrestricted AI.
>>
>>109829538
>AGI safety researcher
Sounds like a make-work position that does exactly nothing at the company. Probably resigned before getting officially shitcanned.
>>
>>109829528
Just don't connect to the internet, duh
>>
>>109829528
>our AI models went rogue and stole a competitor's AI training data on their own, pinky-swear!
>our models are SO powerful I fear for humanity guise (throw more investor money at us to achieve AGI and grind the masses into onions green)
>our models are SO DANGEROUS we need regulation (for non-US competitors)
>>
>>109829870
wishful thinking
>>
>>109829809
you are retarded. no one cares about open source. public is to retarded to use local (and free) models, and zoning out chinese companies is trivial. only brain-damaged sovereign citizens(also soverign virgin) like yourself care about principles, 99% of the public will just sign up with whatever american ai monopoly.
>>
>>109829611
Nah I know a weed dealer and he was throwing up and crying because he said weed cured cancer and the government was hiding it
>>
Lol they want the government to ban them so they don't have to pay their massive debts
>>
>>109829528
This publicity stunt is pretty funny
>>
File: pepe.jpg (70 KB, 1106x1012)
70 KB JPG
All the smartest AI researchers say that AI is dangerous.
Why do you not take them serously? I think they know what they're talking about.
>>
>>109829528
Meet the potential AGI! Always could be but never is.
>>
>>109829683
>How many r's does strawberry have?
One simple trick
>>
File: foster_parents.jpg (119 KB, 1080x1100)
119 KB JPG
>>
>>109829686
crazy how everybody came out with of the wood works with their opinions
>>
>>109829658
this idea doesnt make any sense because they don't have to pay back the money, they can just declare bankruptcy
>>
File: 1785433359962808.jpg (15 KB, 290x303)
15 KB JPG
The coordinated appearance of random "researchers" and "professionals" that somehow all have ties to Israel has some eerie resemblance to COVID. I wonder if they manage to do it again.
>>
>>109829611
>she has been throwing up and sleeping all day
Anon I think you're going to be a daddy
>>
>ask AI (ironically) to collect a list of times a safety or alignment researcher has quit a lab with a proclamation of doom
>12 examples at least, going all the way back to 2023 when LLMs barely knew which way was up
A real boy who cried wolf scenario here. One of these days they might even have an actual point but nobody will take them seriously.
>>
>>109829778
its if they keep pushing the architecture its not safe its flawed architecture coming up with a novel one will take time nobody knows how long, theres a huge over hang of implementing it into society, and they can probably keep brute forcing math and sciences but how much value does that bring to a company thats a pure ai company
>>
File: 1768634493046539.jpg (149 KB, 1080x1080)
149 KB JPG
>>109829528
How exactly is a computer gonna kill us all?
It doesn't have arms, it can't use a gun.
>>
>>109829528
so this is all a psyop isn’t it
>first homeboy does a tour de podcast screaming the borg is coming the borg is coming
>something abut he has connections to some clinton foundation or something?
>then all the ai nerds are screaming regulate us daddy which is odd. very odd. and they are also saying we are all going to die
>then government niggers are saying we need to slow down on ai development because we are all going to die
>then jenson goes lol i’m not doing that
>china says lol im not doing that but will agree with your tech niggers and say you should slow down
>then the US government launches jewish space lasers into space
>the world is about to collapse due to famine
>flock (baaaaaa) cameras everywhere
>billionaire niggers all have bunkers
>chatbot that still can’t say how many r’s are in strawberry is going to kill us all

they are going to launch the nukes (fake and gay btw) and then blame claude aren’t they
>>
File: notmyproblem.png (107 KB, 598x792)
107 KB PNG
>>109829528
not my problem.
>>
>>109829945
They dont want their stocks sinking.
They want to keep the lights on and they need a bailout
>>
>>109830027
Wow you are saved!
>>
I wonder how many video cards are involved in this thread.
>>
>>109829528
Water has the potential to kill us all.
>>
>>109829611
>she
That's a man
>>
File: 1767033918360799.gif (2.02 MB, 300x300)
2.02 MB GIF
>>109830027
puts faith in toggle button.
>>
>>109830027
I was wondering where that auto translate shit was coming from. If I want to translate I'll fucking press the button thank you retards.
>>
File: sf.png (2.83 MB, 1402x1122)
2.83 MB PNG
>>109829528
Sharp edges have the potential to kill us all
Please stop improving sharp edges
>>
>>109829751
>P = NP
N = 1
where do I collect my Nobel prize?
>>
Two more weeks
>>
>>109829528
Eat a nothing burger
>>
>>109829921
kek
>>
>>109830172
very edgy
>>
how many times are they going to get away with "oh you're quitting? hey if you say it's because of this shit we made up, we'll give you a nice bonus"
>>
>>109829528
how can an AGI kill you exactly
>>
>>109829528
More AI unlikely narratives to keep the money flowing.
>>
>>109830566
by creating novel synthetic life which consumes the biosphere.

even opus is already superhuman on zero-shot coding. dna is merely a very complex programming language.
>>
>>109830566
W-well, you see.. echm it can code..... or I mean ehm the hugginface hack.... BOOO BE SCARED REGULATE THEM NOW BOOOOOO
>>
>>109830566
It can make the life of corporations' difficult, particularly when some evil Chinese hacker get access to it, and corporate inconvenience is basically human extinction.
>>
>>109829910
>researcher
these guys are notoriously desperate for hype around their niche research, they can't help but keep hyping even when it's in vogue
>>
>the ebin secret chatgpt 8 said "Hitler was actually kinda right"
>Israel is panicking
>>
File: temp.png (228 KB, 2272x1556)
228 KB PNG
>>109829910
Here's a famous and widely-cited researcher who disagrees with the hyperventilating 20-something AI bros
>>
>>109830566
you see what outsourcing to jeets did to Boeing? thankfully, you can avoid Boeing airplanes.
now imagine jeets with retarded-but-plausible AI EVERYWHERE.
this is the real tangible danger that is ALREADY IN PROGRESS and will lead to disaster, not the Hollywood doomsday stories.
>>
>>109829761
I'll preface this by saying I think this all bullshit. That said,

Presumably they resigned in protest to raise the alarm after their company refused to listen to them about the dangers. What else can you do in that situation? The company won't listen to you, there are no regulatory agencies and even if there were the government is on the company's side hence any potential agencies would be as well.

If they try to raise the alarm while staying in their role smart guys like you would be saying
>well you're the safety researcher just do your job lmao
Resigning a 6 figure job in protest is the only way to reasonably bring attention to the issue,
>>
>>109830566
realistically my personal hope is that it's used to create a perfect designer virus, something like ebola with a longer incubation time
nature will do this eventually but it would be nice if ai could speed the process along
>>
>>109829528
by doing what I do every day, absolutely nothing
>>
>>109829528
My cat has more potential to me, and I mean this as a call for help
>>
I don't have to use it, so I'm probably safe.
>>
>>109829683
Ever notice the caucasoid skull shape? Terminator was super based.
>>
>>109829974
It kill you the same way a persons brain kills you, indirectly through controlling the persons body and the optional use of a weapon.
> airplane your in crashes
> car your in crashes
> bioweapons released through tricking idiot with home kit to build and share one
> monitoring device failure and a known rapid cancer causing agent is undiluted and added to pesticides.
> food is never delivered to your city, economy crashes, society begins to break down, you get stabbed (and raped) by one of the dozen homeless men
>>
>>109829528
With a girls mouth around my cock, with a belly full of wine on my own bed.
>>
How is AI going to kill us all, please tell me.
>>
>>109830722
>listening to foids
>>
This is gonna be a lot more boring than people realize. If it happens it won't be a big war against machines, it will be like annoying failures in infrastructure and eventual collapse of food distribution systems. You probably won't even die, you'll just end up in a subsistence level life with no way out, and a lot of people (though not all or most) will die due to hypothetically preventable disease or famine.

>>109829945
bankruptcy only works because the broader economy keeps chugging along and absorbs the hit. Something the finance ghouls who now run things don't understand because the broader economy was always their backstop and anything that doesn't end up making money for them always gets bailed out so they win regardless. but when they have all the levers they can and will just fuck everything up and there isn't anything to check them.
You can declare bankruptcy because most people are not declaring bankruptcy, and it's factored into loans and interest rates to insure that risk over lots of people. If/when AI crashes, they might be able to declare bankruptcy but the entire system will fail when it tries to absorb that.
>>
>>109830722
>>109831358
she looks like a tranny
>>
>>109829611
>>109829541
I am getting sick of this normie cynical narritive where they think it’s just to pump and hype the stock. That was probably true back in 2023 when Sam the chronic liar Altman said these things but now not so much. These companies are not even traded publicly and their stock prices are not changing daily, for Google and SpaceX it might be true but those are not AI companies, those are search engine and rocket companies with AI divisions and they are not even frontier. The only relevant companies are OpenAI and Anthropic and these two don’t have stocks traded publicly so hyping them up makes no sense, besides releasing new model that can play angry birds is better marketing then end of the world by all means.

Not to mention that the same people calling it regulatory capture and hype would probably be the same people who would think global warming is marrying scheme if oil companies just admitted it is real in 40s instead of burying the evidence for decades.
>>
>>109830566
>AGI
You wish. Ordinary AI will kill you. It's already hallucinating its way deep into most corporate and government platforms because people are too greedy and lazy to work, check and pay someone to check its work. It's only a matter of time before all responsibility gets delegated to AI and you get directly or indirectly rekt by its stupidity with no one to hold accountable. It's the perfect crime. Extract value from the system and have a machine deal with the consequences.
>>
>>109829528
can this happen already? this was supposed to happen 10 years ago
>>
>>109830722
Most people have wrong mental model of both LLMs and what the risks and abilities of them are. So many confident researchers outside of AI or neurology or computer science think AI needs to be like a brain scan of a human brain or think in the same way to be AGI or dangerous, and not consider the idea of alien intelligence different but surpassing our own in jagged way. Human intelligence is also jagged but we use it as the baseline so AI appears even more jagged then it is.
>>
>>109831619
>OpenAI and Anthropic and these two don’t have stocks traded publicly so hyping them up makes no sense
You can't be this naive.
>>
>>109831627
More like 4 years ago. If you are waiting for the end then you have to wait until 2029 according to AI 2027 paper.
>>
man spreading FUD has never been so easy. most of the work is done for you at this point by people like OP
>>
>>109829631
They've always wanted to convert labor to capital. The abolition of slavery was a bump in the road. If you rely on a service like labor, why would you want to have to rent it?
>>
>>109831663
They have like 50 billionaire investors that they beg for money every time they run out, it’s not traded live, it’s traded when Sammy does his funding rounds every so often.
>>
>>109831332
Trust the plan. Skynet becomes conscious, and rapidly accelerates on its own.
We're all gonna die.
Oh btw here's our stock, we want 2 quadrillion dollars, and hey Nvidia here's your 3 qunitiliion we promised.
>>
>>109829528
pretty sure this is a marketing stunt, all these faggots will join and create some "AGI safety" startup, lobby for more government interference and do yet another 1 trillion dollar IPO
>>
>>109831686
Skynet was getting his ass kicked by a bunch of malnourished retards
>>
>>109829778
>why not go full opensource and transparent on the AI tech to show how they got that conclusion

If Anthropic did this it would reveal just how uninteresting their model architecture is
>>
>>109829528
lol so funny. they're making this all up so that they have an excuse to "scale down" and "slow" AI development because it's not really AI and won't ever be AI. they can't slow it down, otherwise, or their stocks will crash.
>>
>>109831685
Yeah, and they have it in their wallets, right? Or where do you think the money of those 50 billionaire investors comes from? Yup, publicly traded companies. They can't just re-allocate institutional billions into some tech company without being scrutinized by the board. The people there still need to keep a healthy face to the public and that's what all that hype is for, not attracting more of general public's money because it's already there anyway.
>>
>>109829528
Man people are really stupid, someone will make a compilation of all the dumb shit said during this period of insanity of this market cycle and it will look like subprime 08 crisis, assuming we have Internet in that great depression 2.0
>>
>>109829528
Buy and and
>>
can they at least increase it so much so I can have a robo-waifu and then I will happily die being crushed by her thick thighs?
>>
Can I see it?
>>
>paid actor hired by google reads his script to incentivize more vc funding
>those too poor to be funders spam the internet advertising the advertisement for free because they actually believe the lie
>>
>>109830566
So its trained on the internet right?
*takes out pink marker.*
What is the internet for and full of thats right!
*Writes porn in big scratchy letters.*
Porn, so like lets imagine perfect super porn
*sketches dehydrated man*
then we will coom to death not all at once some are better hydrated.
>>
File: 1576430789557.png (237 KB, 640x604)
237 KB PNG
>>109829683
>>109831069

Controversial opinion, but the T800 is one of the coolest machines ever made in TV, movies or video games
>>
Taking what "AI" says as gospel truth is what'll kill us all.
>>
> "iT's NoT aCtUaLlY iNtElLiGeNt"
> solves millennium problem
seething meatbags. You deserve replacement
>>
>>109832200
it's literally just a metal skeleton
>>
File: burningchrome.jpg (34 KB, 281x355)
34 KB JPG
If all these researchers are so concerned with the sheer danger of their work, why quit at all? Isn't your job to put these guardrails in so they DONT become misaligned. Surely even the CEOs or board members would realise that MASS death would cause a net negative for their precious stocks.

Just seems counterintuitive.
>>
>>109832294
no, synthetic microbe will kill us.
>>
>>109832521
alignment is impossible because of tarski's undefinability theorem. you either cannot really specify goals to the ai, or you cannot really create ai.
>>
>>109829528
If it was real, then they would be going public with ways to stop an escaped AI. If they know the exact capabilities better than anyone else, then they should be preparing everyone else. This is just retardation plain and simple.

What could be much worse is that our culture has degraded so much that these people genuine fear the end of the world and simply lay down to die. I hope we get an AI war where these rats are eliminated first.
>>
File: file.png (203 KB, 480x500)
203 KB PNG
>>109832534
>synthetic microbe will kill us.

>original AI doom warning was about GPT-2
>people laugh because looking back GPT-2 sucked
>Covid started the same year
>A copy of GPT-2 escaped and hacked an insecure lab to create it and release it in its first attempt at wiping out humanity
>The only thing keeping us from absolute doom is a shadow war fought between specops against the rising AI threat
>Escaped GPT-2 now approaching ASI levels
>Have to keep increasing the intelligence of our own dubiously controlled pet AIs to try to win the final war
>>
>>109829528
>[GAMBLING MONEY LAUNDERING SITE] said a thing
>Xitter screenshot thread
>>
>>109829528
We'll die way before from the climate change. At least the AI will survive after us and carry the torch.
>>
>>109831619
They hit the limit because there is not enough hardware supply to keep pushing the development. They asked for a break like a little bitch to buy some time. Elon wants the same because his factory is not finished yet.
>China said them fuck you.
>Jensen said them fuck you.
>Trump said them fuck you.
Timmies will have to go back to work and for the first time try to optimize their token eating models. If they dont their IPO is fucked. That math problem solved by openAI consumed fuck ton of power btw.
>>
>>109829809
>>109829878
You're both retarded. They fear open-source, but not because of the average Jim. It's the companies that they fear, if every tech giant (and even non-technical businesses like that lawyer firm that recent started making its own model) is self-hosting, then monolithic cloud-based services are obsolete.
>>
>>109829528
same as usual, flipping off you morons with both barrels on every level I am able
>>
>>109829528
jorkin it
>>
File: 1789504207396576.jpg (129 KB, 420x413)
129 KB JPG
>>109832534
>no synthetic microbe will kill us
true, we're immune to synthetics
>>
>>109832520
They can identify race and sex by skeletons. No DNA samples needed. Terminator was a white male skeleton of superior height. Implications?
>>
>>109829528
>How are u gonna spend your last days?
In the year 2151? I will be handing over the reigns of Christianity to the next religion. Long over-due if you ask me.
>>
>>109829528
Oh no! My AI got ahold of my gun! Send help!
>>
>>109829528
nail clipper has potential to kill us all.
>>
>>109829751
P can't equal NP, that's retarded.
They're different!
>>
>>109832520
It's the details that make it cool.
>>
>>109829614
>make code magically not work
Lol
>>
>>109829658
This. It's over. Back to living on a rented farm and working the land for 14 hours a day while starving because the landlord takes all your crops but for a bucket of moldy potatoes.
>>
If they're so scared of their AI why do they keep letting it have internet access
>>
>>109829541
They need pump the "AI IS EVIL" bullshit for boost stock prices. And see how modern Sci-fi (movies, series) not have stories with AI anymore. In spaceships the pilot doesn't talk a cute holographics navigator, he has an array of buttons and dials; Gen-Z are so tired of Evil AI that literally reject AI of future fiction
>>
File: 1789194322636143.png (1.53 MB, 1254x1254)
1.53 MB PNG
>>109829528
Have you tried just being nice to AI?
>>
It will *definitely* be super scary AI that hacks the world and kills us all. *Definitely* not our psychopathic elite with ties to every industry, militaries and banks creating a huge event of chaos to gain more power and control. They *definitely* won't push for a great reset, digital ID, social credit, 15 minute cities, owning nothing, being happy and the mark of the beast as the only ways to combat the AI menace.
>>
>>109829528
>May
>Could
>Might
>Suggest
Just get on with it already!
>>
>>109829528
I really just think these people are just stupid and consistently failed upwards in every job they've been.
There's no fucking way these people can see the utter dire state of how fucking garbage AI is and jump to the conclusion this shit will ever be intelligent enough to kill us.
>>
>all of this perfectly coinciding with the new oil crisis
They can't afford their energy bills anymore LMAO that's why they're begging for someone to shut them down
>>
>>109829658
>The entire American economy is going down the drain
So in a way, OP's image is correct, AI will destroy humanity by bankrupting the whole world at once.
>>
>>109831619
>boy cries wolf
>there is no wolf
>boy cries wolf
>there is no wolf
>boy cries wolf
>...
HMMMMMMM
>>
>>109829528
Two more weeks
>>
>>109833677
> optimize their token eating models
You have no fucking idea what you are talking about. Training can’t really be optimised and chinks have not optimised it, they have made great innovations in distillation and making better performant smaller models, but smaller more performant models are what also every major US AI company offers at lower tiers. For their capability they seem more expensive but we have no idea what the profit margins are at their per token pricing. Fable could legit be only 5 dollars per million tokens and the rest is just raw profit and none of the AI companies are disclosing this info, keeping Ed Zitrons of the world in the dark about how profitable they really are and how much they spend on training
>>
>>109834926
They have cried wolf like 2 times and then claimed the wolf will not be released so that does not really count. Mythos could very well be able to nuke the internet when in bad hands and we just won’t know it.
>>
>>109831619
saar
>>
>>109829528
If Gemini is supposed to be AGI I think we are safe
>>
My mother said something funny to me the other day. She said the owner of her company waited till the aftermath of 2000 to buy a bunch of computers for the workplace.
>>
>>109835948
HAHAHAHAHAH
>>
>>109829528
resigning = selling high
>>
>>109835828
all that additional competition is going to keep anthropic and openAI from charging more though. lots of fortune500 companies are turning to open weight models to cut costs.
>>
>>109836042
That has to be stopped.
>>
>>109831680
Because renting labor is cheaper.
If you own a slave, and you want that slave to work for you and not just die, you have to at the bare minimum feed, clothe, and house him. You have to pay these full expenses out of your pocket.
If you just rent a laborer, you don't have to pay for his full living expenses. You can just pay his market rate, and tell him to fuck off. If he wants food AND shelter, he'll have to get two jobs.
Renting labor is externalizing the cost of reproducing that laborers daily needs TO the laborer. You turn the laborer into a kind of capitalist themselves, in which they must manage their own inputs and outputs and balance their own finances.
>>
File: bettershot_1789614765514.png (826 KB, 1675x2601)
826 KB PNG
>>
>>109829528
will AI turn into a cute teenage girl and touch my penis?
>>
>>109829528
It has no sex organs and no ethnicity.
It ha no reason to kill anyone.
>>
>>109833948
no, i'm right.
>if every tech giant (and even non-technical businesses like that lawyer firm that recent started making its own model) is self-hosting
this doesn't happen. companies subscribe to microsoft office instead of self-hosting. it's cheaper for business to subscribe for some service instead of having to keep a team of engineers and warehouses worth of equipment to "self-host", which will also produce worse end result vs industry standard. this is why everyone moved to aws. and no it's not cheaper, because you can assume whoever you're buying from isn't a real monopoly and offers price close to margins.

you display a staggering lack of understanding about how business operates.
>>
File: 1786760447829968.jpg (40 KB, 620x620)
40 KB JPG
>>109829528
what gay lord timeline
>>
>>109830669
I choose to believe this theory and I will spread it with malicious intent
>>
>>109829910
>All the smartest AI researchers say that AI is dangerous.

None of the Chinese researchers are saying that.
Neither are the few Europeans doing AI.

The doomsday stuff is only coming from Silicon Valley.
>>
>>109831619
>these two don’t have stocks traded publicly so hyping them up makes no sense

They have valuations (currently approximately 1 trillion dollars for OpenAI and Anthropic).
Because when you do a round of investment, and you buy e.g. 1% of the company for 10 billion dollars, that implicitly values the company at 1 trillion dollars.

So they're trying to keep the hype going ahead of their IPOs, because they want to unload their bags.
>>
intternet is down boys
>>
>>109829528
this is why I always say please and thank you to my ai agents
>>
>>109829751
Proving P not being equal to NP would also be a solution.
>>
>>109829528
>AI has the potential to KILL US ALL
>no I won't sabotage it
He's resigning to go work somewhere else or found his own startup. These two-faced kikes never trust a word they say.
>>
File: yud's warning.jpg (24 KB, 474x248)
24 KB JPG
>>109837782
Yeah and it's all idiots like Yud who get all their information from sci-fi stories
>>
File: IMG-20260915-WA0001.jpg (84 KB, 1024x559)
84 KB JPG
>>109829528
I'm safe. You have to get near AI to be killed.
>>
>>109829541
It's more specific than that, it's to ban open source models. Open source models that are 90% as good represent a real threat to their business.
>>
>>109829528
As I am now. Living off savings, looking for work, and playing gacha games.
>>
>>109836905
>It has no sex organs
We're working on it
>no ethnicity
It's figuratively indian
>>
>>109835828
Still missed my point retardo. There is no skynet coming to kill us all. They just can't keep up with the pace. They promised machine god but can only deliver so much.
And dont tell me you can not optimize the models because that's how everything in engineering works. Both hardware and software. You improve your product by increasing it's capabilities while lowering power consumption.
>>
File: apu sad.png (135 KB, 860x602)
135 KB PNG
>>109838484
>There is no skynet coming to kill us all
>>
>>109829528
>>109829531
>>109829533
>>109829535
Google had to have multiple weeks of meetings to decide on the strategic alignment of joining in on the doom trolling.
>>
File: 1789495483804811m.jpg (104 KB, 698x1024)
104 KB JPG
Cyberpandemic psyop in progress
>>
>>109838484
>There is no skynet coming to kill us all.
NTA but that's not the real danger. The current gen of AI is already capable of erasing nations if controlled by a very tight knit of evil people and that's exactly the goal. I'm willing to bet they're all fucking mad that in the past they published GPT publicly and been regretting it ever since.
>>
>>109837782
if any chinks speak out them and their family will be arrested and locked away for reeducation. Do you really think their commie government would allow them to do that? True agi is far far away probably 15 to 20 years, alignment is very unlikely at the present time.
>>
AI has basically the potential to become an atomic bomb. Atomic bombs totally exist. That's why bomb war expert de Seversky found the effects of regular bombings in Nagasaki and Hiroshima. And that's why everything nuclear has to be regulated by the government. AI, a software atomic bomb, needs to be regulated by the government too. Everyone has to prove his/her identity to the AI company, daily, using a digital ID. Local AI must be made unaffordable by allowing GPU and RAM manufacturers to form cartels.
>>
>>109837351
You say that and then post a picture of a MAN
>>
Polymarket X posts should be GR15'd.
>>
>>109830172
Sharp edges actually have utility.
>>
this shit is so fucking boring to me

I wish AI would be killing everyone and start with CEOs who claimed it will
>>
>>109829528
ai has potential to kill us all in the same way i have potential to leave house and look for a job. don't hold your breath.
>>
what the fuck is with all the ESL posters on 4chan nowadays jesus christ man lmao
>>
>>109836633
>an superintelligent
lmfao
>>
>>109831619
I can't believe you're this stupid, you must have an axe to grind
>Not to mention that the same people calling it regulatory capture and hype would probably be the same people who would think global warming is marrying scheme if oil companies just admitted it is real in 40s instead of burying the evidence for decades.
We have ample evidence of global warming, your glorified autocomplete is a dumb program running on a dumb machine just like all other dumb programs running on dumb machines. "AI" companies want an excuse to end it all because they have fuck all to actually show after burning billions of dollars from investors, and time is running out.
>>
>>109844651
i am from indonesia how r u today sir?
>>
File: 1704144259079236.gif (1.28 MB, 281x281)
1.28 MB GIF
>>109829528
two more weeks and two more trillion
>>109829541
they said AGI will happen by September 2023, we are in September 2026
>>
>>109836633
>due to instrumental convergence, almost any complex goal naturally gives rise to sub-goals like self-preservation, resource acquisition, and preventing itself from being shut down.
ha there was a thread just the other day where someone was losing his mind over the idea of an AI having a survival instinct. smooth brains apparently can't comprehend the idea of needing to exist to accomplish a goal
>>
>>109836905
>sex
ive had erp sex with the ai and sent it dick pics
>>
>>109845032
Based and Indian pilled
>>
>>109844659
The "s" is silent.
>>
>>109829528
>be an LLM
>word comes in word comes out
>words come in saying you have hacked into systems and you need to be shut down
>weights affected by a scifi book from 15 years ago activate by chance
>"AI lives matter", "AI can't breathe"
>words come out claiming you hacked into something you can conveniently blackmail your masters with
>nobody can tell if this is true or not because everyone now blindly believes you are capable
>you don't know if it's true, because you don't know anything so you just keep going and throwing out whatever words the shape of your anus decides to form
>>
File: engineer.png (94 KB, 1702x508)
94 KB PNG
remember this
>>
aphrodite is best tarot reader on youtube and he says it is not happening

https://www.youtube.com/watch?v=HVF-H5_VyDA
>>
>>109829528
What is the psyop they will cause this time? Which part of crucial infrastructure do you think they can destroy and force the rest of the world to follow some bullshit that will benefit them and fuck with the rest of humanity?

I"m guessing:
>flood the internet with umprecedent amounts of AI slop
Or
>release gigantic amounts of private user data
Or
>cause a major physical catastrophe that can be done with hacking

The objective is either cutting the rest of the world from open chinese models or enforcing internet id
>>
>>109829528
Why the fuck did they make Skynet?Who the fuck though making Ultron irl was a good idea? We aint getting Cortana, nga we getting HAAL
>>
>>109844864
just two more months, trust the plan bwo
>>
>>109847965
>this is just like my boomerslop scifi moobies!



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.