That’s a weird way of saying that they had a net loss of $8 billion. Are you trying to imply that this is somehow extraordinary for a growth company? How do you figure that?
Then it didn’t make 13 billion…… and it didn’t lose 21 billion.
It lost 8 billion. That’s how math works.
You’re not wrong, but neither is the title. That’s just how business accounting works. Ultimately, it’s still the same conclusion ($8 billion loss)
A lot of people are gonna give openaishit over this, but the old cliché is true; gotta spend money to lose money.
At a personal scale you can just light a few bills on fire, and you’re good. Maybe even gamble away a home or something if you’re filthy rich.
But at this scale that just doesn’t work. Burning this amount of money takes infrastructure, trucks, roads, labor, and facilities. Otherwise the poor’s might take some of it.
Can you imagine what those people would spend this kind of money on?
So they lost $8b.
Altman is such a miserable cunt.
he pretty much is created by thiel, so yes he is. since ALtman met his husband at one of thiels “pool” parties. just like how he plucked vance out of obscurity from yale. he does love his gay puppets.
indeed
Yep, defo not a bubble.
Vibe stocks
What’s their op ex though? I feel like there must be a sizable portion of that which isn’t actually necessary to keep the service up.
So the the operating expense was greater than their revenue from operations by about 2, but it seems like they’re minimizing it by hiding the cost of some of the compute inside marketing and training costs. This is something that a few AI companies in China have been caught doing to make it seem like they’re doing better than they are. So they could be incinerating money at an even faster rate than they just admitted.
And in R&D
Yah, given that “training models” doesn’t stop when the model is finished and released. Like, a released model needs to be continuously tweaked to keep it up to date or to deal with problems that have occurred. Even if that’s not literally tokens used by customers, it is compute being used to provide service to customers.
And that’s just assuming that they’re not just hiding some compute costs used to service customer demand inside the R&D budget. “Oh, you see, this pool of customers are being served with an experimental version, so any compute here is actually R&D, any API fees or subscription payments made by them of course get counted towards normal revenue.”
Bah that’s baby numbers compared to what “private space” will accomplish!
“So what did you do with the $100,000?”
"I invested it and turned it into sixteen THOUSAND dollars. "
OpenAI:
“I DON’T NEED MONEY! I DON’T EVEN LIKE IT! I JUST WANNA THROW IT AWWWAAAAAAY!!”
im betting sam altman asked claude instead of CHAPGPT to how much to invest and where.
First thing I thought of lmao.
deleted by creator
Do you know how to make a small fortune with AI?
First, start with a big fortune…
Looks like the shovel seller is making quite a bit
JENsen still can afford his jackets, and is pretty jubilent about the whole ordeal, albeit still alarmed.
Indeed, they’re the only ones. Along with Samsung, SK Hynix, AMD, etc.
This is interesting. Nvidia cashing in while they can. They’ll come back to us consumers when/if all this AI stuff collapses.
NVIDIA has the whole ai bubble wrapped around jensen like a jacket. yea hes going to lose alot but not as much as the ai tech bros.
Wow, I like that “spent since page load”.
I want a ticker like that “Lemmy pages since bong load”.
I don’t smoke weed, I have a future
deleted by creator
You have to spend money to make money :)
spend more money to lose more money.
You are not actually considering the scale of 20 billion dollars nor the jump in a single year. Those are bad things.
The actual economics behind LLMs shows us they cannot become profitable ever, and this “invest more before becoming profitable” story does not fit in part because of that. Other companies who used that model had a way to become profitable.
Are you serious? You don’t know much about this if you think that’s what’s going on.
See my response below and keep an open mind.
Edit: And yes I am serious because this is how business has been done for a long time
Sometimes you have to spend more than you make but that doesn’t mean it’s a bad idea! You just have to continue spending and one day it’ll maybe start turning a profit. Maybe not, though! 🙂 I love AI
Thank you very much for explaining what I meant.
Are you being facetious or are you just dumb?
They’re being facetious but it’s actually what AI investors are thinking. AI can lose money until it’s the only AI company left then they can charge whatever they want.
I had an AI engineer tell me that hallucinations are actually AI having original thoughts. Zzzzz
You are wrong about me being facetious, I have explained in my response.
Wtf, these folk s are clinically insane. It’d be funny if they weren’t literally fucking over the entire world in the process.
Neither. It was very nice of [email protected] to make an effort and explain my comment in his response. Sometimes things might not be as simple as you think so please keep an open mind. Thank you
So the latter, gotcha.
This is what you call a succesful business man /s
If someone gave me 21billion I bet I could only lose 1 billion, return the 20, and never be seen again.
This was standard operating procedure for Silicon Valley companies before COVID.
I have no love for AI but I feel like the people clowning on this don’t understand the most basic aspects of how businesses work.
It is very common to take losses early on and return a profit later.
Well what I currently don’t understand is what is the industry use case. From the numbers alone you would need to spend double the current tokens cost, which is already very high. Every industry partner I talked to said they don’t have valid use cases for ai
Gtfo, there is no way llm companies make their money back legitimately and we all know it. No, it’s not “common” for companies to spend tens of billions to build infrastructrure that won’t pay for itself.
Those of us in reality know they will get bailed out by the government in exchange for fucking us even more than already.
YouTube would lose a billion annually early on as they expanded infrastructure to keep up with massive demand 🤷♀️
They figured it out eventually, all they had to do was enshittify everything.
That is entirely different from llm models. People like youtube, it has utility. Llm models don’t have enough utility to pay for their data centers and we all know it. Why are you simping for them? You believe their hype? That discredits you.
“I don’t like them therefore no one does” -you
People had the same conversations when YouTube was new and unproven, before it was the household name it is now. Many people thought it would never be profitable and watched them burn billion after billion.
I do think LLMs are overhyped, I’m with you there, but I do think they also provide utility that many use. Is it a bubble that will pop? 100%. But just like the web after the dotcom bubble, it’ll never go away.
Lmfao, look at this righteous edge lord.
Who needs to think things through, when you can just condemn someone who says anything “black” when you’re on team “white”.
wtf? Bitch.
Big baby bitch can’t even think things through.
The difference being they had a plan to monetize YouTube which offset the operating costs and even could make a profit, in addition to the stuff they over invested in having long term benefits. Aka they roughly could spent a lot to save money later off of what they built.
LLMs are not the same. They cost a shitload of money to run. Actual token based usage without being subsidized by investors would make any LLM cost users so much money that the actual value of it would immediately become major problem. Sure one can, currently, get decent code output by using hundreds or thousands of tokens, or using multiple LLMs / loops, or having agents go burn however many on iteration! But if we paid actual costs this shit would rapidly be shut down.
The AI infrastructure is also not saving money long term. Training is unbelievably expensive. Compute costs are all about gigantic data servers with video cards running, and the economics of all that is way tighter than anyone gives credit for. Those cards last like 3 years. The cooling costs are crazy. You have to have constant use to be efficient. None of these things are able to be covered by AI economics nor does it even make sense to be.
It costs too much, it’s just being covered by your money being pissed away by tech investors for a technology that cannot survive.
Sure one can, currently, get decent code output by using hundreds or thousands of tokens, or using multiple LLMs / loops, or having agents go burn however many on iteration!
This always stumps me. Because if that was true, Anthropic products (API, Cache, Claude Cowork, Claude Code) would not be shitty.
They are shitty. They are coded shittily, and Anthropic is unable to solve some bugs for years now. E.g. there is the console flickering bug that they tried to fix 3 times and rollbacked or failed all of them.
Or maybe we define decent differently.
In theory costs could come down with each new hardware generation if the we dont keep pushing models the to max extent of what the hardware can do while pushing size.
E.g Claude Opus today, only trained in a similar size and manner as today, will be cheaper to run on whatever the next GPU that comes out with higher speeds and processing capabilities, unless of course NVidia raises the cost substantially. Given the current situation I think nvidia might do that which would hamper this lowering of costs, but it should possible, if not slower.
E.g 10 years from now it will be cheaper to run a opus similar model. But 10 years from now everyone will want the mythos of today, then. That wont be cheaper.
This has been stated since ChatGPT was released and has not happened. The video cards released specifically for LLM usage do not benchmark particularly better than the previous generation. And it’s still unbelievably expensive to run these cards and maintain the facility and, again, you only get like 3 or 5 years out of them! That’s a crazy investment lol
Tell me you don’t know nothing about business without telling me you don’t know nothing about business.
No no stop. I know what you’re thinking. You don’t, relax. Take it easy.
GTfo, stroke off the enemy on your own.
I see that “GTFO” is a major portion of your vocabulary.
I recommend reading some books. I’d start with green eggs and ham by Dr. Seuss. It’s fairly complicated. It has some rhyme schemes that can really mess with your head.
The issue for OpenAI is that their main competitor, Anthropic, has a better product than they do and is currently scooping up their market share. So that means they’re going to have to spend billions more to try to catch up, and Anthropic won’t be standing still in the mean time.
This sort of competitive arms race can burn vast sums of money and result in multiple companies going out of business if they fall far enough behind to lose investor confidence. An even bigger issue (for investors) is that no one has been able to demonstrate an AI “moat” which would allow a company to gain any traction. Without a moat, customers can jump ship the instant a competitor offers a better model.
You don’t understand but keep pretending
Spotify didn’t start seeing yearly profits until 2024.
Twitter didn’t see profits in a single quarter until 12 years after they started. And only went 2 years with actual profits from inception until they were bought out.
Took YouTube 9 years before they saw any profits.
Point being. It’s very normal for big tech companies to not make money for quite some time. OpenAI isn’t alone. It’s the rule rather than the exception.
Will they go tits up? I don’t know. Time will tell. But my guess is that investors will keep pumping in money for at least another 10 years.
There is not enough money to keep funding them for ten years.
If that’s what you think, you severely underestimate how much money there is in the world.
We’ll both find out soon enough.
They won’t ever break even. Chinese models are already comparable
No they’re not.
The best Chinese models are good, but they’re resoundingly outclassed by the likes of Claude.
And that’s not to mention the Chinese models are trained off the American models, i.e. they can’t exist without the American ones being developed first.
They don’t train off American models, that’s just their PR narrative to explain losing the race to a country without latest hardware
https://m.youtube.com/watch?v=-DVxMLFhi3o Glm 5.2 is almost the same as fable, difference is immaterial.
I use deepseek v4 pro daily and it’s fine for most tasks and so cheap that I stick those deepseek api keys into most of my applications.
There is no use case for chat boxes. Uber was losing money, but it was driving people to places. Amazon was losing money, but it was delivering books to people. Chat boxes don’t do shit.
Uber would have failed if their cars just drove in circles. Amazon would have failed if they didn’t deliver anything. There is no real use case for this tech, just like with scam coins or monkey pictures, this is tech searching for a problem to solve.
The most foundational aspect of business is that you gotta do something to get money. What exactly have these chat boxes produced? Shitty porn and incorrect advice. If that’s the best they’ve got after consuming the entirety of human knowledge and guzzling an ocean of freshwater, this shit ain’t going nowhere.
If you think AI is just “chat boxes” then you really need to just start out Of any and all AI discussions.
Yes, you are correct. Many companies, especially in the tech industry, lose money for the first three to five years after opening.
What you have below are people who do not understand basic business concepts, such as the difference between revenue and profit, let alone capital investment.
You are also contending with people who hate something they do not understand. They call LLMs “AI,” dislike Sam Altman and OpenAI, and often do not even realize there are other companies and models that, depending on the metric, can outperform ChatGPT. They are hating for the sake of hating while disguising it as enlightenment. It is quite frustrating, and I push back against it whenever I can.
At the end of the day, people need something to hate, and right now LLMs and data centers have become convenient targets.
That is not to say there is nothing wrong with the industry. There is. Data centers consume enormous resources, and the constant drive for profit creates plenty of legitimate concerns.
My point is simply that many of these people do not actually understand what they are arguing against in the first place.
The amount of money that’s been invested in this will require 100 million people to pay $10,000 per year so they can have a 5% return on investment.
Who is actually going to be paying that kind of money for AI services? Has anyone ever actually worked for a company before? You need to go through layers of red tape to requisition a new office chair. Are companies actually going to spend hundreds of thousands of dollars per year for AI?
I’ve asked people that have used AI to alter photos (currently free to them, paid for by investors) how much they’d for such a service. Nothing, just would leave the photo as-is. There’s a big market for “free AI” but the market for AI where the users pay the cost + profit margin is a small fraction of that.
So why is this stuff valued in the trillions? Simply because the greater fool will buy shares so line goes up. Once the pension funds buy in, then a bunch of billionaires get together and short it simultaneously. They’ll make money and everyone else loses. That’s how it worked in the dot com bubble, there’s no reason to expect it will go any differently for AI.
Of course you might have gotten confused because there’s real tech in the story. Same as there was in the dot com bubble. We’re having this conversation on the internet right now. But seeing tech, using the tech doesn’t mean the tech isn’t overvalued. One of the “ridiculous” claims of the dot com bubble was that you’d be able to pet food over the internet. Today we most definitely can do that. Just because that eventually proven correct, it didn’t make sense in 1999, and having billions invested in something that wouldn’t have results for more than a decade isn’t a good investment. Equipment depreciates, and high-tech equipment depreciates rapidly. Buying networking equipment for an online retailer ten years before the logistics needed are in place is just throwing away money.
Buying GPUs before there’s a data center built to put them is throwing away money. Facebook is putting them in tents, Samsung wants to put them on ships. Obviously these GPUs aren’t going to last long operating in these conditions, but it’ll take years before there’s data centers built and even longer before there’s enough power generation needed to run them. They’re just milking every last dollar from investors by doing these things.
Many of the things promised are straight up lies (Halting Problem, anyone?) and it’ll be years before the the infrastructure is built. Right now they’re maxing the hype and making as much as they can before the bubble bursts. And why not? Fraud is effectively decriminalized in the US.
Right. Corporations do not pay taxes on income; they pay taxes on profits, and the tax code gives them significant flexibility in determining what counts as profit. Loans are not taxed. “Buy, borrow, die” is legal. We have weak antitrust enforcement. Politicians can trade stocks despite occupying positions that give them access to information the public does not have. Competition in many industries has declined, reducing incentives to prioritize consumers. We even have private healthcare.
So what is your point?
Companies, especially in the tech industry, have historically operated at a loss during their first several years. Even after becoming established, a 4% profit margin is often considered respectable. Revenue and profit are not the same thing, and investing heavily for growth is a normal business practice.
I was talking about people criticizing LLMs while clearly knowing very little about them. The bandwagon effect on this platform is strong. Many people dislike LLMs, call everything “AI,” and often do not understand the underlying technology, the economics behind it, or the fact that there are multiple companies competing in the space. I push back when I see criticism based on misunderstandings rather than facts.
There are definitely many companies willing to throw millions a year at AI, and are currently doing it, but as it stands today, it doesnt sound like they’re getting the return on investment they expected.
It can change if the models actually got better, but how much of it is inherint to how LLMs are made, and/or how much more can they be improved before this all falls apart
It cant go on forever as is today.
