https://www.axios.com/2026/04/26/ai-cost-human-workers Uber’s chief technology officer already blew through his full 2026 AI budget due to token costs, according to The Information.
Lol. Lmao even
You must log in or register to comment.
It’s almost like it was an obvious and stupid pile of lies and shit the entire time. If only literally everyone with a brain had been constantly pointing that out literally the entire time, then we could have done better, right?
Rule: if something looks too good to be true, then without any further evidence, it’s likely too good to be true.
This feels predictable. AI is one of, if not the most invested in yet unprofitable industries in the history of humanity.
The last few years have been the beta and the tech demo. But that is not paying for itself yet. US companies are competing with (and falling behind) Chinese state-sponsored companies. OpenAI in particular, a company whose revenue doesn’t even cover half of their operating costs, has extended themselves into owing more than a TRILLION dollars to the entirety of big tech who are building chips and data centers on these IOUs, and will need to be paid sooner or later. The bills will come due.
Other corporations are already paying massive bills for licensing, tokens, training, and infrastructure changes to accommodate this shift to AI while laying off massove chunks of skilled workers on the idea that AI is cheap and will get cheaper over time. But that is simply not the case. This is the “first taste is free” part of this deal. Once they have companies deeply invested in AI and have destroyed the fabric of the labor economy in favor of it, that price is going to skyrocket because OF COURSE IT WILL.
Maybe at some point this will all level out. AI bubble will pop. Prices will sky rocket. Companies will try to backpedal, which will be slow and difficult, they’ll end up paying AI companies huge sums while they work to decouple themselves after just forming the bond, they’ll also end up paying stupid money to professionals who are suddenly in high demand, and many companies won’t survive the chaos. But the ones that do will settle into a new equilibrium.
AI will eventually get cheaper (but probably never this cheap again, at least not in the near future), and it will probably be a permanent fixture in our lives and work to some degree. But it’s usefulness and cost effectiveness will be limited in scope, with specialized purposes. It will not ultimately be the great labor replacement companies think/thought it would be, even as stupid and short sighted as that desire is in the first place (if 30% of the global work force is unemployed, how do you think that will effect your revenue, morons!?). But that also is assuming that the coming chaos doesn’t turn out so bad that AI is permanently legislated into oblivion after the chaos it’s about to cause.
Good stuff. One small note: I’m not sure how useful the distinction of “Chinese state-sponsored companies” is in recent history when comparing to the US, let alone now. The US has retooled much of federal research engine toward promoting US AI. Even fired the NSB (among many other long standing, expert driven advisory boards) to replace it with a bunch of tech baron stooges. States are offering unprecedented payouts to data centers. The AI hyperscalers already have a bailout all but guaranteed when the bubble pops. It’s all state-sponsored, just with extra steps.
The Chinese AI companies being state sponsored just means that they can go longer and throw more money at development without turning profit than other investor driven companies.
The US is certainly throwing a bloated amount of money at AI too. And a much as it infuriates me, they’ll almost certainly absorb the bubble pop with tax another bailout for criminal corporate behavior. But it’s not quite been a direct pipeline of openly flowing cash, just yet. They’re still paying for discrete contracts which have to be approved in the budget. They’ve been massive contracts, but they’re still making these companies compete e each other for them too. Like with the recent flip from DOJ contracts with Anthropic to OpenAI, for example.
In China, they’re buying in supporting the entire industry. They’re building infrastructure for AI data centers, giving them grants and subsidies, have direct ownership in the companies, and had made specific carve outs in their laws to give AI development deregulated room to do what it needs. I’m not in favor of either approach. Just pointing out that China’s approach does seem to have been an advantage in the AI race, or at least was enough of one that they made up a ton of ground, and maybe passed their US counterparts.
AI is one of, if not the most invested in yet unprofitable industries in the history of humanity.
I think there are some Dutch tulip farmers who would like a word with you
Not quite the same. The tulip industry was making money hand over foot. It was the speculators that ended up being shafted. Tulip mania was more comparable to the beanie babies craze, or even NFTs. AI companies, on the whole, are making no profit at all.
It makes more sense if you position Ai companies as the speculators and chip makers as the actual tulip producers.
Add to that the fact that hiring and training a new employee usually costs between 5-10 times more than retaining an employee (from hire to fully trained)
Not just hiring and training! You also have to start paying state unemployment tax on that new hire. In Florida the first $7,000 is taxed on each new employee. Then there’s loss of efficiency, and related items. On top of that, if your turnover is high, your payroll company will up your rates because they’re working harder and you’re a PITA employer. I’ve sat meetings where we decided exactly that.
Yet they never seem to act like that. They punish you for loyalty.
The fact that it is just a cost comparison, however much humans might still be winning it roght now, is the fundamental problem.
Yeah no shit. You get people dependent on your product, then Jack up the price
That was always the reason AI was put into everything.
Corpo fell for the oldest corpo trick.
It’s funny too because at our company people keep repeatedly asking “what are you gonna do when they predictably jack up the rates?”
And every time “don’t worry about that right now. We’re keeping a close eye on it. “
Uh huh.
“Why did you destroy your company to just inflate your next quarter bonus?”
“It is my nature”
“Lol.” Said the scorpion. “Lmao.”
And it’s from their own playbook
Those same managers eleven seconds later when they get an ad for a new startup making the same obviously empty promises as the last startup:
They love those the most because they integrate them and then use it to justify a promotion or move so they can get out of Dodge before the inevitable explosion happens on the next guys watch. The next guy blames the previous guy and then repeats the process.
They fired the wrong people. Go higher!
Does AI cost more than humans primarily because of greed (i.e the AI companies demand a high profit margin now) or because of energy costs (i.e AI is so wasteful with energy, so polluting, that it costs more than human workers)
They just say a number. If nobody pays, it’s too high. If everyone pays, it’s too low. Aim is for i) highest market share, ii) max ARR, and iii) highest margin.
They’re selling the idea that a machine costs less than a human. They’re Walmart, humans are mom-and-pop shops. Once the competitors are gone, they charge whatever they want (you pay or you close out). Fuck them.
Costs. AI companies have been running at a big loss using investment money trying to scale quickly and conquer the market. That always comes at an end and something closer to the real costs has to be paid.
Given the ai companies are running at a loss, it’s fair to assume which of these is likely
It’s both.
Precisely. The question then is, which one is the main driver? I think it does fall on energy cost/the ridiculous scale of infrastructure they’ve decided is required to sustain AI companies.
Conclusion (for a luddite) is that One could cripple AI companies if simply prevented them from finishing their data centre every time. Goodness, it’s like a RTS strategy game where you have to build a monument to win the game.
If the other one is the main driver of this, purely an inflated profit margin, it indicates that AI is already collapsing and they’re desperately trying to scrape more venture capital off the back of the businesses that haven’t clued-in the how ineffective AI usually is.
This is a common myth, inference is not typically run at a loss, despite claims. It’s only a loss if you include staff and ongoing training costs. They could lock in their models now and be profitable if they wanted to.
Edit: I see the comment above has changed (or I misread initially) to say the companies are running at a loss rather than inference running at a loss. Yes, that’s extremely true. Now my comment doesn’t make any sense and is irrelevant so feel free to ignore my pedantry.
Yes, and let’s also not count all the investments in infrastructure because you know… like training and staff it’s not a real cost that’s essential to the business.
Anyways, you wouldn’t happened to have heard that from Anthropocene or OpenAI?
Somehow we don’t have any actual indisputable numbers (I wonder why) but it is actually quite controversial and some of those who have done deep research on the subject are saying inference IS run at a loss and it might not get profitable ever.
https://www.ft.com/content/fce77ba4-6231-4920-9e99-693a6c38e7d5?syn-25a6b1a6=1
We do have numbers from comparably sized Chinese models.
Yes, every AI company is bleeding money, they’re not healthy in any way. But inference by itself is profitable, based on everything that we know.
Inference + amortizing the training costs is NOT profitable, which is what most people are talking about.
This is easily fixed by not releasing a slightly different version every month.
Those people disagree with me and what I want are therefore wrong
“Inference is not typically run at a loss”
Bro thats called cherry picking
Businesses work on cash in cash out
Right now AI companies make way less cash than they spend overall when you dont include investments
Furthermore, most people use a free version of AI and would stop using it if it cost them anything
Explain how to pivot to profit when the investments dry up, were all waiting
I’m not saying they’re healthy, I’m saying that inference is the one profitable part of their business.
They’re all going to die because training costs dwarf the inference, and training doesn’t generate ANY revenue.
Furthermore, most people use a free version of AI and would stop using it if it cost them anything
Do companies tend to use the free version too?
Do rhetorical questions add anything of value, beyond a gotcha moment, to the conversation?
You know that wouldn’t happen. Which AI company wants to be the one that says, “we’re happy with where the model is at right now” and stops throwing cash into the boiler of the investor hype train and let their competitors exceed them in real or imagined metrics? Clearly firms like Anthropic have to rely on circus marketing tricks like “This model is too dangerous for the general public to see! Ooooh scary! Coming Soon!”, and they can’t do that without continuous training.
For you and I, the offline models aren’t too bad for getting little side projects started, but for major AI firms, the ongoing training cost for the next model and the one after that has become ingrained into the operating model.
I’m aware! I’m not saying they are healthy in any way. I’m just correcting that specific misinformation, because truth is important.
These companies are fucked if they keep operating the way they currently are, and I strongly suspect it’s all going to pop like the dotcom bubble, but worse.
The AI companies bet on efficiencies occurring that have not materialized.
The fact that piece of algorithm is getting paid more than a human being, who eat, live, love, is outrageous. Humans are the worst.
Just to twist the knife, the algorithm was likely trained on the replaced workers behaviours
And likely does the job worse.
I know I do! Last week I tried to overthrow a country with no oil! Egg on my face I tell ya.
Executives are the worst. There are many, many great people in the world, but for some fucking stupid reason we let executives and politicians run roughshod over everyone and everything…
I for one think it’s pretty bad ass that humans are able to build sophisticated enough tools to replace “brain” jobs. Not sure how AI is a humanity bad moment unless you are someone with a make-believe job being replaced by AI
I’m sure this comment will go down well with Lemmy.
Don’t worry, when they hire back their humans they will be paying them 25%+ less after adjusting for inflation.
But they got to fire a lot of people which is their favorite thing to do.
AND it’s shit
Hahahahaha their greed knows no bounds and they couldn’t help but reach and reach and reach further in until they ruined the one actual selling point they had of AI “replace the slaves cheaper!!” They cried waving their arms around “This is the new thing!!” “You won’t have to pay people anymore! Think of the profits!!”
And then they went and fucked it up because of how obscenely and unapologetically greedy they are in every single thing that they do. Hahahahahaha oh god, the irony is almost tangible enough you could put it on a plate and eat it.
Good for us, the actual workers of the world. But wow, they are so greedy they couldn’t even replace us all with robots effectively because they just want more and more and more and more and more and…
Edit: actually its a more than that. See the below comment.
Most LLM setups were running at about a 95% loss. That’s unprecedentedly humongous.
Subscriptions were going to need to go up 20-fold just to break even.
Rehire and Replace the CEO with AI 😁
Too easy.
