Microslop CEO begging for AI $$$ because astronomical overprovisioning is becoming obvious, all big spenders frantically trying to hide CapEx from their books and hallucinate revenue projections like its Enron reloaded and Oracle is already getting sued by bondholders over AI spend [0].
To me CoreWeave is the one to watch. They have to actually bring all these promised datacenters online, operational, and profitable. They basically got a $2B bailout from Nividia a week or so ago but they're back to sinking.
Consumer can eat all the GPUs they have and more if we stop trying to force B2B
Right we have a loop where AI is so expensive (because it's priced to feast on B2B margins) that the best consumer experiences aren't affordable, and they're not affordable so they don't go mainstream, and they're not mainstream so no one is willing to take less money and bank on the incredible volume that would emerge if it went mainstream.
If we can get model pricing cheaper AI entertainment alone will probably carry things (I'm 99% sure NovelAI is already one of their largest customers outside of major AI labs)
Even if consumer can eat all the gpus, it cannot have the margins (as you say), and thus won’t sustain the current valuations all these companies have and which fuel the (necessary) investments.
NVIDIA is sitting on 74% gross margin. If we reach a place where "all" these companies have to do to unlock nearly unbounded demand is take lower margins, they will find capital.
If anything I'm more worried about the consumers than the companies.
It is not nvidia who has problems here, but the ai companies which depend on huge valuations for their continuous funding and survival, thus also those who invest on those valuations. I am not worried about them though, I hope they all burn to the ground, but alas it will not be them who pay for it in the end.
> Consumer can eat all the GPUs they have and more if we stop trying to force B2B
You should really crunch the numbers on buying and then running enough compute to run a leading edge model. The economics of buying it (never mind running it) just dont add up.
You still haven't factored in "training", the major problem right now that every one remains head in sand about.
I dont need a model to know who Tom Cruise is or how to write SQL if I am asking it "set up my amazon refund" or "cancel xyz service". The moment someone figures out how to build targeted and small it will take off.
And as for training, well having to make ongoing investment into re-training is what killed expert systems, it's what killed all past AI efforts. Just because it's much more "automated" doesn't mean it isnt the same "problem". Till a model learns (and can become a useful digital twin) the consumer market is going to remain "out of reach".
That doesn't mean we dont have an amazing tool at hand, because we do. But the way it's being sold is only going to lead to confusion and disappointment.
Consumer, as in B2C, not consumers buying directly. B2C companies will happily buy (or rent from people who are buying today) GPUs, because a huge part of the game is managing margins to a degree B2B typically doesn't need to concern itself with.
> I dont need a model to know who Tom Cruise is or how to write SQL if I am asking it "set up my amazon refund" or "cancel xyz service". The moment someone figures out how to build targeted and small it will take off.
I think people got a lot of ideas when dense models were in vogue that don't hold up today. Kimi K2.5 maybe be a "1T parameter model" but it only has 32B active parameters and still easily trounces any prior dense model, including Llama 405B...
Small models need to make sense in terms of actual UX since beating these higher sparsity MoEs on raw efficiency is harder than people realize.
If you truly believe this, slowly divest everything into cash, wait for the crash, then buy back in. Even buying in slowly over the course of a crash, on the way down, will save you a ton of money if you're out before it hits.
But you're more likely to just cash out early, lose a bunch of gains, then buy back in later at higher prices.
If you can time the crash you can make a shitload of money. But you can't, so you'll come out better if you just keep buying in every paycheck and ride it out just like you have been.
There is the separate risk that Microsoft, Google, etc. will have a lower value in two years as governments get their migration off their platforms into full gear
It doesn't have to be 'ride it till it dies' or 'sell everything'. The AI bubble is almost exclusively contained to the US stock market and a few east Asian manufacturers.
You're right that selling everything and 'going to cash' would be a mistake, but diversifying away from US large cap growth absolutely wouldn't. I'm 60/40 stocks/bonds. My stocks and bonds are 50/50 us/intl. ~ 10% of my us portfolio is small cap value.
What's funny to me is that nobody learns from the past. This is far from the first tech bubble we've had even before the .com crash (canals, railroads, radio...). The answer, every time was diversification.
The east Asian semiconductor manufacturers are selling shovels in the gold rush and being very cautious about expansion given how capital-intensive the whole sector is. It's hard to come up with a scenario where they outright lose, even with the bubble popping.
I mean there's also a cost to not expanding too, in that you're leaving money on the table. I doubt they've really been able to resist the siren call of basically being able to print more money, but if the AI bubble collapses and they're left selling most of their production to consumers, they're gonna have a lot of stranded capital. Here's hoping they're smart enough to build a big war chest to weather the storm, but in my experience, companies rarely do.
Invest into stuff that people will need regardless of the bubble popping like medicine, food, internet access, energy, ... . Stay away from luxury/travel stuff.
Also, during a crash there is the so called "flight for quality" where people cash out from risky assets and invest in stable ones that can weather the storm. So, try to invest in assets that are A or above (https://en.wikipedia.org/wiki/S%26P_Global_Ratings). The chart is for countries, but analysts grade companies as well in case you want to stay away from treasuries/national bonds.
Also diversify geographically. US will likely take the biggest hit if the bubble pops, so perhaps European markets that lagged behind in adopting the technology are safer (IMHO).
Personally, I am preparing by moving money from growth items to stable ones a bit at the time. To diversify even further I am using ETFs that, in addition to what mentioned above
1) pay dividends (whether these distributed or reinvested doesn't really matter)
2) are denominated in or hedged in safer currencies (CHF especially, but also Euro)
You definitely get smaller returns, but the name of the game is to maintain what you have, not to make heaps of money.
Finally, I am not a financial advisor, so do your own valuations/risk assessment analysis.
Current US debt to gdp is 124%, 38.6 trillion.
Japan too at 230-240%.
Bond markets in both are looking seriously unhealthy (Japan going via a Liz Truss moment at present).
If the AI bubble falls over, the US government is going to have to print 5 trillion to cover the bubble at least. The only option there is inflate away anyone holding cash.
If hte AI succeeds and people are replaced, the US government faces a massive fiscal cliff of a loss of tax receipts. They won't be able to service the debt and again will be forced to inflate away.
To service current debt projects, AI growth needs to return some 3.2-3.5%, it is currently 0.5%.
Bonds, equities, USD, and housing are all risk assets right now.
Primarily the fault of our governments not using anti-trust laws for real in, like, decades.
Governments actually do have the power to regulate the economy and to prevent catastrophic crashes from occurring. The warning signs for the AI bubble have been visible for well over a year now, when the entity relationship map between the major players began to resemble a Habsburg family tree... and yet, nothing was done.
This is EXACTLY where I was getting at. People raise their fists at big-tech and big-finance for creating a bubble whose splash damage will hurt everyone, but it's ultimately the government's job to monitor, regulate and prevent this.
The reality is even worse than this. It's not that the government is asleep at the wheel, it's that all financial crises were caused by government intervention to begin with (2009 subprime government backed mortgages, 2020 covid money printing, etc). The big banks, corporations and wall street were only taking advantage of the situation the government helped create in order to enrich themselves.
Average people would have also loved to have taken advantage of it to enrich themselves, if they could understand the system well enough to game it, and some do, so it's not like the blame is only on one participant, but pretty much everyone is to blame:
- the government for creating asset bubbles and closing an eye as it is inflating
- corporations and super wealthy for being greedy and exploiting the system created and run by the government
- the voters for being uneducated and stupid and not seeing the government rob the blind, or for being greedy and complicit on the asset bubble for personal gain, and not holding the government accountable
Maybe I'm misunderstanding your position, but no regulations is certainly not better than some regulations, assuming the people in government positions have the actual best interests of the country in mind. Deregulation specifically led to the sub-prime mortgage crisis and the current situation, not the existence of regulations itself.
>but no regulations is certainly not better than some regulations,
I never meant such a thing.
>assuming the people in government positions have the actual best interests of the country in mind
They don't. They care about themselves and the lobbyists paying them, and sometimes they throw you a bone to win an election.
>Deregulation specifically led to the sub-prime mortgage crisis and the current situation, not the existence of regulations itself.
Deregulation is also a form of regulation because the government is the one that removes the previous safeguards put in place via new regulations. And Clinton was the one responsible for the sub-prime mortgage deregulation.
Literally all the cloud providers have been reporting severe capacity crunches for the past few quarters -- to the tune of backlogs of triple-digit billions each. As a reminder, a backlog or "Remaining Performance Obligation" (RPO) is money their customers have committed to them but they could not realize because they didn't have enough capacity to serve their workloads. Which is why they are all committing to double-digit billions each in AI CapEx spend over the next few quarters.
And most of them (aside from Oracle, which is trying to borrow its way into this gold rush) are investing money from their double digit billions in profit (per quarter!) into this spend... money that they could have otherwise comfortably held on to for something more palatable to share-holders.
Revenue and return on investment is a valid concern to bring up in this whole GenAI shebang; demand is not.
> "Forget about managing complex server infrastructure for your database needs."
So what does this run on then?
No docs, it's not possible to find any deployment guides for Ray using serverless solutions like Lambda, Cloud Functions or be it your own Firecracker.
Instead, every other post mentions EKS or EC2.
The Ray team even rejected Lambda support expressedly as far back as 2020 [0]. Uuuuuugh.
No thanks! shiver
I'd rather cut complexity for practically the same benefit and either do it single machine or have a thin, manageable layer on top a truly serverless infra like in this talk [1] " Processing Trillions of Records at Okta with Mini Serverless Databases".
Because, fundamentally, their biggest competitor is Google. A company with a market cap north of 4 trillion. If OpenAI does not spend hundreds of billions of dollars on datacenters, people will migrate to Google, little by little, and OpenAI will became a new Netscape story. A good product eliminated by an incumbent with infinitely deep pockets.
> If OpenAI is incapable of building anything that can't be easily copied by a third party
They can build better models, but for that they need a lot of compute and that's where the billions go. These better models can't be easily copied by a third party, because that third party needs to also throw billions at the problem, and billions don't grow in trees.
People are gonna move to google anyway, because Google can keep the gravy train running for much much longer. OpenAI's business model is totally reckless while Google is a cash rich company.
The harm being done is millions of dollars for VCs and music studios.
Billionaires and enterprises want to see consumers spending to return their investment.
The presence of other - dispersed - illegal material doesn’t diminish said returns too much, this central dump would have set precedent and had garnered massive attention.
Microslop CEO begging for AI $$$ because astronomical overprovisioning is becoming obvious, all big spenders frantically trying to hide CapEx from their books and hallucinate revenue projections like its Enron reloaded and Oracle is already getting sued by bondholders over AI spend [0].
It will be worse than the dot com bust.
0: https://www.reuters.com/sustainability/boards-policy-regulat...
reply