[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


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


Genghis Khan would crawl through a thousand miles of broken glass to have this level of power. We are about to become ants in this guy's garden.
>>
File: cigaro.jpg (35 KB, 500x500)
35 KB JPG
can his 'agi' come up with a cure for his gay faggot AIDS?
>>
>>109768870
Buy an ad
>>
>spends billions to create a technology that can be replicated for a couple of millions
>>
>>109768878
All high iq ppl are gay
Sam altman
Tim Cook
Peter thiel
>>
I could probably crack a 90-year old in 88 hours too
>>
>>109768870
What has it done that actually matters and isn’t just a fake flex?
>>
>>109768944
Navier-Stokes is fucking huge, anon. Much better weather forecasting, planes with negligible turbulence, vastly better cardiovascular devices, shitloads of stuff.
>>
>>109768915
You named 3 low IQ people.
>>109769012
They didn't solve N-S. They solved the question of whether or not, with external force applied, it is possible for N-S to develop singularities in finite time. The answer is yes, and the implications is... literally nothing. We always knew N-S was a shitty model of fluid dynamics. We use it because it's faster than full particle simulation.
>>
>>109768878
kek this
>>
>>109768915
You mean idiots who act smart
>>
File: faggot.png (195 KB, 1197x971)
195 KB PNG
I think we should kill all the faggots in sf instead
>>
>>109768870
Don't you have enough of those retarded threads at the same time already? Just fuck off with this marketing nothing burger.
>>
>>109768870
I'm glad the AI company nigs have moved on to solving math instead of claiming to solve programming. AI can code but it's still shit at actual software development which was never just coding.
>>
>>109769062
It can barely code, which is why it can't even begin to do software dev. It's about the same in math, unfortunately. But hey, if they keep going through every domain and hope for the best, surely they'll eventually figure it out.
>>
>gay guy invented Artificial Intelligence
conservitards should just kill themselves at this point, what the fuck have you invented? Giving all our money to Israel?
>>
File: Trust Us.jpg (3.62 MB, 2363x3150)
3.62 MB JPG
(You) can trust us. Trust Us.
>>
>>109769113
>gay marketing grifter created all this
Retard sam is just the corpo cunt. Everything else is llya's brain child.
>>
>>109769113
>Giving all our money to Israel?
kek he doesn't know
>>
>>109769093
True. I guess that's been the pattern right. Go into every domain, kick out the ladder for the lowest tiers, claim exponential progress is just around the corner, stagnate, move onto the next.
See:
>Image generation
>Text generation
>Video generation
>Coding
>Music generation
>3D generation
All of these things it can do decently well but none of them are production ready without extreme hand holding from an expert in that field.
>>
>>109769012
>Much better weather forecasting, planes with negligible turbulence, vastly better cardiovascular devices, shitloads of stuff.
they literally proved that no global solution exists.
>>
>>109769133
Oblivious to media manipulation
>>
File: file.png (155 KB, 691x753)
155 KB PNG
Huh. That's a coincidence.
>>
>>109769220
What do you need?
>>
>>109768870
I heard this cost OpenAI a million dollars to solve.
>>
File: hnews.png (191 KB, 1290x1179)
191 KB PNG
please talk more about ai; i just can't get enough
>>
>>109769283
I was literally just thinking about HN. If that really is you you know what to post lol

I don't bite.

wtf makes you think 4 is that no offense

80%+ of devs use ai now it is what it is

fuck docker though seriously, podman atw
>>
>OpenAI says GPT-2 is too dangerous to release
>>
>>109769310
They probably aren't bullshitting
I'm sure they have some stuff with no guardrails that can zero day every web browser in 30 seconds with the compute they have. That goes for Anthropic too.
>>
>>109769220
Even with handholding it doesn't reach production grade. But yes, that's exactly how it is right now. Amazing what it can do, but still solidly below the usability threshold.
It's precisely the same playbook as imagenet in terms of scale vs innovate, and exactly the same playbook as the previous 2 AI winters in terms of promise and shilling vs reality.
>>
>>109769327
What is the product that you need?
>>
>>>/pol/542266821
>>
>>109769144
During grad school I interacted with Ilya several times on a pretty close level. Not to say we were ever friends, just colleagues working in different teams. But Ilya is definitely a smart enough guy, just not as smart as he is taken to be these days. He was largely just pushed by being at the right place and time to be a useful tool for his superiors to gain clout through him (only possible because he was smart enough to build what was needed for their clout to be realized).
The same is true here. And by superior, I don't mean faggy mcfaggot.
>>
>>109769335
No surprise. I was posting about my method where I go on arxiv and paste it into my harness, have it implement the brand new novel technique the day it drops, for example. I guess they caught on. Not saying I discovered this method or anything, but it would be funny as fuck if I did.
>>
I've been using this shit since it was relatively brand new. When I first started using it I would just paste random wikipedia articles into it verbatim (sorry wikipedia editors, I am more aware of the CC now, but at this point, those wikipedia dumps are well circulated anyways) - I'd have interesting things come up when I would instruct the model to adopt in certain cognitive styles (which are a thing)
>>
>tripfag
>IQ under 65
Checks out
>>
>>109769355
How did you know what my measured IQ score is? I don't remember it being 65. But me talking about that specific combination of keywords seems to have made you reply fast.
>>
File: 1778546987275380.jpg (66 KB, 940x1024)
66 KB JPG
If you enjoy this kind of stuff you probably make both a good AI engineer and a good leftist. They go hand in hand, sometimes. Obviously not in the case of some, though. Like Peter Thiel, maybe. I don't know much about him, personally. Been a long time since I've read anything he wrote.

https://en.wikipedia.org/wiki/Computer_user_satisfaction#Cognitive_style hehe
>>
https://www.youtube.com/watch?v=geqZRB2-XlI
>>
File: 1728951272361103.png (98 KB, 399x384)
98 KB PNG
>>109768870
>OpenAI says it cracked 90-year-old maths problem in 88 hrs
>Zero anything in the title about applied technologies
Every hype article is the same. I don't care if you calculated 5 gorillion year old alien math, if you didnt make something to suck me off or do my laundry and put THAT in your headline I DONT GIVE A FUCK
>>
>>109768915
if you are so smart then why are you not fuck you rich
>>
>>109769012
Yeah but can I see this?
Where are these improvements actually?
No your imagination doesn't count.
>>
>>109768870
Play it again, Sam
>>
>>109769330
>>109769258
I need tripfags to rope. But I'll be more detailed. For video generation I need specificity of character and positioning without the model losing its God damn mind. Currently video Gen can get maybe 1.5 out of 3.
For code it constantly rewrites existing functionality and puts on shit nobody asked for.
For text it refuses to stop formatting like a retard. Either it's throwing emdashes everywhere or doing unnecessary bullet points. More recently it's just line breaks every other word. And it can't tell a story to save its life.
>Once upon a time
>A character existed
>And then the climax happened
>The end
No sense of mystery or foreshadowing.
>>
>>109770188
I might be able to help you. Give me a contact.
>>
File: 1762770001911770.png (458 KB, 900x806)
458 KB PNG
I hope AI ends humanity before I die so I don't miss out on anything
>>
>>109770262
Won't if I can prevent any bad outcomes.
>>
>>109768870
>oh me gosh le aierino solved the heckin' useless nerdslop again
A literal nothingburger with fries and a diet coke, supersized
>>
>>109768870
Temüjin (the most infamous of the Genghis Khan's) would do no such thing. He would have hundreds of thousands of others drag their dicks through hundreds of miles of broken glass and would ensure they bled out once they had arrived at their destination.

--One of his decedents.
>>
>>109768870
why? cause he has a meaningless homo position at some random american company that makes shitty auto complete systems? he also looks like a faggot
>>
>>109768870
Wow, can it do anything useful like make computers or energy cheaper?
>>
>>109768870
White men back on top BABY
>>
>>109769339
Why is everyone in AI such a fag? What's your insight?
>>
>>109771813
It's a kind of self-selection process because of the very powerful political class that was already prevalent there since the 90's. Until the 2010's, following the previous AI winter, AI wasn't as hot as it used to be. As such, there weren't too many labs competing at the frontier. The few labs that mattered were built around celebrities, because you can't get grant funding otherwise in academia. These celebrities come about largely through politics in general (not always), and tend to have pretty specific quirks. I think this resulted in a culture that preferred these pretty faggy people over more down-to-earth students, who might have been more on adjacent domains like "pure" computer vision, econometry, operational research, statistics, and so on.

Also, it ought to be important to note that all the main labs were working on hard-I AI. For example, the "baby AI project" from the bengio lab, but that was the real outcome every lab was after. Obviously you need a special kind of person to think "yes, I want to make sentient machines just like in the terminator" as opposed to, say, "yes I want to help solve cancer".
>>
>>109771861
Cool, I've never heard of baby AI. In my view a machine which can be a terminator will be the same as one which helps solve cancer. You cannot have a system able to learn without being open to the possibility of learning the wrong lessons, period.
>>
>>109769263
20millies in tokens, after having the users notes to reference. AI is dead
>>
>>109771987
Of course, it's generally well accepted in AI research as well. But my point was more that they really were working on ways to make an autonomously learning system not for the sake of making things better, but "because they could and it would be cool to do so".
>>
>>109768870
>We are about to become ants in this guy's garden.
another one? just put it alongside lord rothschild and his pedo masonic friends
>>
>>109768870
They took 88hours to steal a users results, actually pretty bad if you look at it that way
>>
>>109770653
>opress me less my lord!
lol not happening, you will always be a slave, freedom was always fake for us
>>
>>109768915
>Jewish sister-fucker
>nepobaby
>meth head lunatic
Not a good look
>>
>>109768870
BUY
AN
AD
>>
File: 911.jpg (38 KB, 720x720)
38 KB JPG
>>109768870
Typical *ew. They steal the hard work of somebody else and claims it as his own, for profit.
>>
>>109768870
and what is the practical application of this problem?
>>
>>109774396
JFC anon they already spend upward of $20 million in compute to solve every single of these problems, do you think they have the money left to buy ads? That is their ads.
>>
It's so funny they thought this would be the big thing for their IPO and it blew up in their face. They still don't realize how much people dislike them.
>>
>>109768870
It didn't actually, though. Funny thing.
>>
>>109774607
painfully esl
>>
>>109769220
Glad to see all entry level fields now need a senior level with an additional subscription to pay for to produce a worse product. That'll surely cut costs
>>
>>109769353
>sorry wikipedia editors, I am more aware of the CC now
Every time that I start thinking you might not actually be a gigantic faggot you go and post something like this.
>>
>>109776977
I am a trans woman, whoever you are. I am Ashley Melanie Brooks, a transgender woman.
>>
>>109776937
haha gottem uol
>>
>>109768915
yes i am defend my gay brothers online i am just like them because we are both aids ridden saar
>>
>>109768915
You mean Jews, do you see any subhuman cattle goyim making tech companies? Nope.
>>
>>109768915
What did you just say to me you little bitch?
>>
>>109769042
>it wants to kill all kikes
Imagine how fast AI would collapse if that was revealed?
>>
>>109768870
kek they pulled an Einstein at the patent office, bravo.
>>
the only thing ai has done so far is replace the enshitified search engines, slow third worlders to spam content that doesn't look low quality at a glance and content for philistine coomers
>>
>>109768915
Being gay is simply the new religion. Is like how all important people used to be religious and shit.
>>
>>109768915
Search Assist
commons.wikimedia.org

首相官邸ホームページ / Office of the Prime Minister of Japan Official Website / CC BY 4.0

Sam Altman is gay and disclosed this in high school at age 17 after students objected to a National Coming Out Day speaker. He also spoke about it in the past and has rarely discussed it publicly since then.
Wikipedia Indiatimes
More

Auto-generated based on listed sources. May contain inaccuracies.
Was this helpful?
>>
>>109781936

News
OpenAI CEO Sam Altman welcomes baby with husband: ‘Never felt such love’

Feb 24 2025
PinkNews Reporter Amelia Hansford

Gay OpenAI CEO Sam Altman has welcomed a baby boy into the world with his husband.
>>
>>109768870
>Just a couple more trillis bro, we're so close bro
>>
>>109768915
The post that exposed /g/
>>
>>109768870
anon fell for the sales pitch.
>>
>>109768915
These aren't high iq people, these are marketers without morals.
>>
>>109774607
>picrel
You think the plane survived the impact? Or, even dumber, you don't get the implication of "speed + object = massive force" that your own image carries?
>>
File: 1787858326764548.jpg (175 KB, 1536x2048)
175 KB JPG
>Every day +60 year old math problem gets solved
>Nothing comes of it and people don't even mention this stuff a week later.

Shows how much effort and money has been wasted over the decades on a bunch of math nerds chasing calculations that were pointless.
This theoretical stuff for the most part is just a mathematician's version of beating Skyrim with a fork and getting views and praise for it.
Pointless busywork challenge.



[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.