Archive link: https://archive.ph/xTfD2
The commentary on September 29 by Morgan Stanley Wealth Management’s chief investment officer, Lisa Shalett, frames the current market boom as a “one-note narrative” almost entirely dependent on massive capital expenditures in generative AI, raising questions about its durability as economic and competitive risks start to mount. Shalett’s critique came squarely in the middle of some people in the AI field — and many financial commentators around Wall Street —fretting at market exuberance and beginning to talk openly about a bubble.
In an interview with Fortune, Shalett said she was “very concerned” about this theme in markets, saying her office had broadened from a belief that the market would only bid up seven or 10 stocks to roughly 40. “At the end of the day … this is not going to be pretty” if and when the generative AI capital expenditure story falters, she said.
What’s more than a little concerning to Shalett is that one company alone, Nvidia—the most valuable company in the history of the world, with an over $4.5 trillion market cap—is at the center of a significant number of these deals. In September alone, Nvidia invested $100 billion in OpenAI in a massive deal, just days after pledging $5 billion to Intel (the Intel agreement was tied to chips, not data-center infrastructure, per se).
Fortune‘s Jeremy Kahn reported in late September on significant concerns about “circular” financing, or Nvidia’s cash essentially being recycled throughout the AI industry. Shalett sees this as a major concern and a major sign that the business cycle is headed toward some kind of endgame. “The guy at the epicenter, Nvidia, is basically starting to do what all ultimate bad actors do in the final inning, which is extending financing, they’re buying their investors.”
Fascinatingly, they seemingly buy now interest on each other.
So when this goes down, it goes down as a whole.
I can’t wait for the bubble to burst. The hype is so annoying. Yes LLMs are cool but they’re not nearly as revolutionary as some think. They will end up as just another tool software developers can choose to make use of when appropriate.
This is exactly what people said about cars, phones, TVs, internet, social media. New technologies dont simply vanish and if you think AI will you havent been keeping track of its evolution at all.
You are blinded by survivorship bias. The technologies you name have persisted because they have turned out useful over time. Many others have not. Nobody uses any more radioactive toothpaste even if radioactive quackery was all the rage in 1920.
LLMs are a timesaver. I have seen the uptake where I work (monitoring via DLP) as a tool, they are useful and the word amongst staff is spreading, so I doubt the tech is going away, even if it’s just reduced to spitting out letters and simple power shell scripts.
Now if it can be offered at a sustainable, profitable price (the current subscription cost per seat is ridiculous). That’s another question.
The thing is not that it can save time for a few applications, but that it’s advantages are dramatically overblown.
Yes but that’s not the same as being the next radioactive toothpaste. As a product, it’s going to continue to exist.
That’s the second problem with AI. At the moment, the industry is highly subsidized by billions of investment money. In the long run, prices will have to cover the cost and revenue expectation of investors.
And no one is willing to pay what it will actually cost. It is unbelievably expensive for results that are mediocre and absolutely useless for critical applications due to the error rate.
My main exposure to AI at the moment are our immigrant juniors using ChatGPT and friends to deliver well-sounded texts. pre-LLMs they struggled to deliver intelligible texts, we would give feedback and their language skills would improve. Nowadays they get no feedback and do not improve. Well, they could use ChatGPT to get feedback on texts they produced on their own rather than just to produce texts – but who in their right mind would increase their own workload?
And did radioactive toothpast have hundreds of billions in investments and a million jobs dedicated to it?
The Dutch Tulip Mania of 1634-1637 certainly had a lot of investments in prices corrected for inflation, and a lot of people working at growing tulips.
The funny thing is that by now, people do know exactly that AI is in a mega-bubble, but still think that they can grasp into that mixture of fire, whirling glass shards, and falling knives, and make a gain from the bubble.
Please point out to me where I said or implied LLMs will vanish
You said they will end up as just a tool for software developers which given their current range of usage is not far from equivalent to vanishing
It’s also what they said about Tamagochis and Furbies, so there’s that.
I agree that tamagotchies were not the technical break-through of artificial life as they were sold at the time.
On a more humourous note 😉: Who needs moaning, hungry, attention-starved tamagotchies when we all have moaning, battery-hungry, and attention-starved smartphones? Do you leave your little smart phone alone at home YOU MONSTER?
You’re right. Tech doesn’t vanish. But it becomes oversaturated constantly. The video game crash of 1983, the dot com bubble, blockchain, NFTs, etc.
Yes, usually growth is not infinite and sometimes theres crashes, but if the trend continues after a crash and far surpasses the level before, I dont call that a bubble. Look at bitcoin. It crashed a few times and every time it rises back to twice its peak. You cant call video games a bubble either, theyre more popular than ever. There are fluctuations in the market, which is perfectly normal for any investment thats not stable.
All hype, no substance.
Now imagine a stock market crash and the government in shut down. So no stimulus for anybody.
We’ve been down this road not long ago. The great recession lasted longer than any government shutdown.
The top 1% know they have to keep the top ~20% to ~30% placated. Otherwise they join the working class in protest. They just have to bail out the “too big to fail” companies so the upper class can make a living. The working class is fucked no matter what.
That’s reason the status quo has remained these past several years. That upper class is still comfortable enough. So far they’re only inconvenienced enough to vent on social media. No so much that they will take actionable steps toward change.
Companies are buying GPUs and using them as collateral to buy more GPUs. All the while ignoring these are consumable products, not infrastructure, and all the while not showing a single dollar of profit. Because generative AI is a scam that has basically no profitable way of existing, and can’t get people to pay for it no matter what they try
The most expensive part are the GPUs and they are obsolete in a couple of years. So not only are they not making enough to cover the cost of operation they are not making enough to toss all the GPUs every two to three years.
They’re not just obsolete in a few years, they’re literally worn out. Running a GPU at max power nonstop will make it break after just a few years, as many crypto miners learn pretty quickly…
Great clarification.
But think of all the value AI provides lol
Well, I started for “patriotism”, but damn am I happy I moved all my investments in EU-only ETFs years ago! Sure, the hit will be felt even here, but not nearly as much.