OpenAI could reportedly run out of cash by mid-2027 — analyst paints grim picture after examining the company's finances
https://www.tomshardware.com/tech-industry/big-tech/openai-could-reportedly-run-out-of-cash-by-mid-2027-nyt-analyst-paints-grim-picture-after-examining-companys-financesOpen linkView original on toast.ooo
75 replies
God I hope it just fucking collapses and pulls every other tech company down with them. I'm so fucking sick and tired of coworkers thinking ChatGPT is the most perfect thing that has ever been made. Every single fucking time I'm searching for something I'm asked "Why don't you just ask ChatGPT". They will not shut the fuck up, even when I show them ChatGPT is completely full of shit. Want it to write a simple 40 line script for Excel? Sure, just spend 30 minutes telling it, its code doesn't fucking work, because it's trying to use functions that don't excist. "You're totally right, and so clever for pointing that out! Here's an updated script that takes that into account", still doesn't fucking work.
I feel this, after searching for packages in apt that not only don't exist for debian, but in general
"But but but didn't prompt it right bro"
If it needs to be "prompted right", its faultry.
I've explained to to some people as "when I speak a language fluently, I'm not going to use Google translate to communicate". Vibe coding is amazing for people who aren't fluent, where they can translate between English and python (or whatever language of choice is).
I know exactly what I want to say, and English isn't a more convenient language for me to express it in
Can sometimes help to edit your previous message with additional instructions and regenerating rather than responding to ask for a correction. Also copy pasting documentation and code examples with syntax that is real into the prompt. Might also try something other than ChatGPT since OpenAI sucks in general.
Or just you know, learn how to code and do it yourself?
I for one would really prefer not to spend brain space on things like Excel scripting syntax. I haven't actually tried current frontier models but every LLM I've used isn't good enough to actually get good results vibe coding without applying some of your own programming knowledge, so I think you'd still want to learn regardless.
Brains are elastic and the more you put into them the better they work.
Are you a programmer? My general experience as a programmer has been that being judicious in what you spend time and energy thinking about is one of the most important things to do. The amount of languages, frameworks, and libraries all with their own syntax, documentation and quirks vastly exceeds the capacity of any one person to durably memorize, and there are compelling reasons to want to use a wide range of them on an infrequent basis and not specialize. Strategies to avoid getting lost in the weeds and find shorter paths to your actual goals were important before LLMs and they are now even more important.
Also, personally, my memory sucks, not everyone is dealing with the same constraints and because of that there is no single right approach.
Yes, if you're memorising them. You can only hold so much in active short term memory. Coding has never been about memory though. It's about understanding why.
You're right of course but they're not arguing in good faith. Wrong excel formulas in 2026? This person is either full of it, or they're trying to prompt gpt-3.5 in ancient Aramaic.
And the cool part is that the USA stock market gains for the last couple years are based on OpenAI (and others) promising to buy stuff sometime in the future to build their data centres, which in turn spins into reinvestment in the company and a giant circlejerk.
AKA a Ponzi scheme. If you think watching the Twin Towers collapse in NYC was frightening, wait until this house of cards falls.
With what money? Trump is adding a trillion in debt every six months and we are running out of allies
The difference is that for decades the flight to safety has been the US dollar in economic downturns. Now, people can't wait to dump the dollar; the flight to safety has been shifting to gold, silver, and even the Yuan.
When the bubble pops, it'll be be an economic fuckening we've likely not seen.
Starting to wonder if it's worth turning some of my savings to regular ass gold.
You are most certainly not alone; I've seen skeptics say that it's doomsday-prep-crazy, but any statements like that are, in my opinion, denialism in this day in age. A regular savings account isn't a viable place to store absolutely all of your money, even since before the Trump admin started fucking things up. A diversified portfolio with safe investments after a certain point is the ideal way to handle it.
Only difference now, is that what a "safe investment" has mutated into is something else now; not to mention how little people can afford to save money. Do what you can to take care of yourself with what you have man.
No but just like South Africa cannot sell trillions in debt, soon the US will not either because we are quick alienating the world and trading partners. If people do not think the US is credible and viable, nobody buys our debt
This is the most gigantic candy in the coal mine almost nobody is talking about!!!
US debt has been growing like crazy since before I was born, and the biggest way we float it is by selling it with low interest bonds that were always a safe investment which is why so many of the world's richest people parked their money here. Now we are seen as less safe and the US bond market is starting to show a world reversal of that sentiment.
When nobody is buying our bonds, the talk you've been hearing about inflation for the past two years will seem like a welcome respite.
Also banks going bankrupt can do incredible damage (this is why I'm a strong opponent to private banking. They should be public services, not for-profit corpos). Tech companies going under is bad but not too-big-to-fail bad.
The assumes that retirement and savings accounts don't invest in the tech companies.
Run out faster
Mid 2026 would be much better!
Or even early 2026. It’s still not too late.
God what a dystopian sentence that is.
That's assuming millions of people switch to agentic browsers and use them agentically. A big assumption.
There's easily millions of people in the US alone that do not know any better and will do exactly that.
You overestimate the critical thinking capabilities of your fellow humans.
Also some just don't care anymore. I was close to not caring anymore too often due to how depressing the past few years have been for many things.
My conspiracy theory: The plan is to fail and sell the data centers to Microsoft for pennies on the dollar.
And also destroy consumer electronics market together with it, so everyone uses their rental clod solutions
I think Apple is much more desperate for that gamble than Microsoft.
Can you explain why? Microsoft seems to have bet the farm on AI, Apple certainly has a less sure future than they have had since the mid 00s but they still have a firm grip on the smart phone market, at least in the US
Microsoft is a weirdly desperate corporation in its attempts to generate money outside of the methods that have kept it afloat since the 1990s.
Microsoft has their own problems. Copilot only has a 4% adoption rate and Microsoft has increased the price of M365 to offset costs. They’re probably trying to figure out how to repurpose what they have, let alone take on more.
building my next pc late 2027 then
Bold of you to assume these companies won't just keep price gouging us, they've done it before.
I just pulled the trigger yesterday. I don’t see prices coming down in a reasonable time, and the bundle at my local Microcenter was the same price as when I first looked a little under a year ago.
Wish it were sooner
the picture:
Why can’t they just ask ChatGPT how to make more money?
That was Sam's plan.
WTF! How did I miss that and how the hell did that not trigger a giant red flag for any investor with more than 2 neurons???
"Computer, I followed your plan and got a trillion dollars. How do I turn a profit from my formerly nonprofit company?"
"Ask for another free trillion dollars"
Because the answer is unlikely to be factually correct and thus unlikely to be useful.
And they are trying to sell that? Dang.
Ask chatgpt how to make them LOSE more money.
Too long, these assholes are fucking up everything from the economy to medicine at great individual and societal cost.
Let's hope it happens sooner. Same with all the other AI shit.
May it happen even sooner.
I'm smart enough to know that when the bubble bursts it's going to mess with the world economy, apparently way worse than the 2008 recession, but I'm too stupid to understand why. I just hear people say things and it would be nice to see these assholes get wrecked for once and not just rake in some cash and dip in time while the rest of the world is left to pay the clean up bill for a party they were never invited to.
I mean, I watched the big short and the only thing I understood from that movie was that the bad guys won and everybody else lost. So I am sorta preparing for the worst here xD
The first 10 minutes of Fun with Dick and Jane might make more sense than any explanation I could offer.
Id say i cant wait, but theyll probably get bailed out and everything will become more expensive as the economy crashes. So thats cool
Burn baby burn
I think all AI companies are just trying to exist at as low prices as possible for as long as possible so that enough companies have time to implement AI into their company infrastructure.
Once the dependence on AI for work flows is locked in they raise the prices.
It will be a interesting time of a big AI bubble burst..... And sadly also bad for many individuals losing their jobs in the companies that can't revert in time.
Dumb CEOs will then again just get more money in the form of hidden compensation packages while the average Joe will have to do.. Something.
Well, if AI drops down out of dumb hype into a normal spectrum it's probably worth it.
God I hope it all collapses before the end of trumps term.
God I hope it all collapses <immediately, and causes> ... the end of trumps term.
AI isn't going to dissappear just like crypto, but god damn does this bubble need to implode.
Hardware market is so screwed that not even the ripoff listings on Ali express can keep up.
Sam Altman needs to be thrown in prison for fraud like Sam Bankman-Fried.
Are anthropic, perplexity or Gemini as a product making any money?
Noooooooo I don't believe so. It's the standard bullshit of burning hard and forgetting to make any actual profit.
nope, why do you think google is trying to jam gemini into all its services, and thier AI VIDEO creation is showing up on youtube.
No AI apps are making money as of now (and probably most of them never will).
They're just trying to survive until company workflows are completely based on them and can't stop using them; then they're going to raise prices.
But I think most companies will fail next year, destroying the market.
Like startups in general when you put it that way tbh
They will just resell all the RAM they just purchased at 5x the cost.
OpenAI: Taps forehead
Modern problems require modern solutions
arnt the AI chips mostly unusuable for consumers.
Ram chips are interchangeable. It's only the interface between the chip and mainboard that differs. So they are easily removed and put onto standard dimm/sodimm boards. The processors can not completely useless either, just maybe slightly more powerful than most consumer need.
They'll try, but we're already getting by on old DDR3 chips. I imagine a few unlucky companies with no other options are screwed, but most consumers are just going to wait for the price you crash.
Not soon enough.
🤞
I feel it will get big enough that it won't be let to fail by every other wall street company that would go down with it
I only don't celebrate because I know they will either be bailed out by the Govt or get bought out by Microsoft for pennies
The comment sections of these types of posts are always entertaining.
You have the "bubble" posters who learned of this word after 2008 and leap with excitement as they have successfully predicted the last 47 of 2 bubbles since 2008.
This article is such a nothing burger. No, it won't implode. No, it won't disappear into nothingness. It's a hyper growth tech company that have historically always operated on losses QoQ often with very thin cash runways until they are forced to become profitable. Then if they still are unable they will do what every other unprofitable tech company does
Huh, what have people hyped that was going to collapse but didn't actually go on to collapse?
Open AI is following the enshitification rule book to the the letter:
Initial quality: Platforms are user-friendly and attractive to build a user base.
Customer data is the product. They suck users in with a "free product", getting millions on board and dependant on it
Business exploitation: Once users are locked in, platforms degrade the experience to favor business customers.
They'll sell the customer data to business customers.
They'll provide it so cheaply that zillions of businesses will build their entire operating model on the provision of cheap data.
Shareholder maximization: Eventually, all value is extracted for executives and shareholders, leaving minimal benefit for users.
They'll then in phase 3 paywall the fuck out chatgpt and markedly increase business user costs. Which they'll pay because they'll have no other option but to pay
Edit me no speak the englrish good