They sought it with thimbles, they sought it with care;
They pursued it with forks and hope;
They threatened its life with a railway-share;
They charmed it with smiles and soap.
Fits well with the release of Hunting of the Snark 1876 by Lewis Carroll
There are certainly many parallels with our current situation. But railroads were a much bigger segment of the economy then. AI's growth is benefitting a narrower set of people, so I find unlikely that the inevitable collapse of AI investment alone, absent other factors, could wreck the global economy the same way.
Are you accounting for the entire electricity sector relying on AI payments that won't come, and the entire world relying on computer parts that are no longer produced except for AI companies?
Do you think token demand is going to drop after the collapse? I seriously doubt it.
IMO the collapse will come because the insane profits that OpenAI/Anthropic valuations depend on will collapse due to margin collapse from competition from Open Weight / Chinese models.
IOW, a reduction in price of tokens. Which will increase demands for tokens, and the electricity needed to generate them.
This might result in bankruptcy of some of the holding companies holding data center debt, but with that debt discharged in bankruptcy the data centers will be quite profitable and keep operating.
Training demand will also collapse after a crash, but training is now a minority of data center demand.
IOW, those electric bills are going to continue to get paid. I wouldn't want to hold OpenAI/Anthropic stock or data center debt though.
I think the risk is a little different. It's not that data centers are being built and then unused -- the fact that OpenAI, Anthropic and pretty much everyone has to put a lot of effort into getting enough compute tells you the demand is there for inference. And it's useful, increasingly so outside of coding.
The risk is in the financial world. The worlds central banks printed money because they thought inflation was tamed, but it wasn't, we just don't measure stratospheric stock and bond valuations as "inflation". I mean, euro-zone bond yields went negative -- but no, that's not inflation-like price mania at all! A big correction in financial markets towards some semblance of reality could amount to a psychological shock on consumers, a big negative wealth effect.
And a hit at the ballot box, hence no major political parties calling for raising rates (or sound money). They rather juice the markets for as much of their careers as possible.
> A boom in railroad construction, 33,000 miles (53,000 km) of track were laid across the country between 1868 and 1873.
> The railroad industry was the largest employer outside of agriculture and involved large amounts of money and risk. A large infusion of cash from speculators caused spectacular growth in the industry and in the construction of docks, factories, and ancillary facilities. Most capital was involved in projects offering no immediate or early returns.
> After 1850 two banking institutions issued 70% of global bonds, the Rothschild family and the Barings Bank.
> On 9 May 1873, the Vienna Stock Exchange crashed because it was unable to sustain the bubble of false expansion, insolvencies, and dishonest manipulations.
There are certainly many parallels with our current situation. But railroads were a much bigger segment of the economy then. AI's growth is benefitting a narrower set of people, so I find unlikely that the inevitable collapse of AI investment alone, absent other factors, could wreck the global economy the same way.
Are you accounting for the entire electricity sector relying on AI payments that won't come, and the entire world relying on computer parts that are no longer produced except for AI companies?
Do you think token demand is going to drop after the collapse? I seriously doubt it.
IMO the collapse will come because the insane profits that OpenAI/Anthropic valuations depend on will collapse due to margin collapse from competition from Open Weight / Chinese models.
IOW, a reduction in price of tokens. Which will increase demands for tokens, and the electricity needed to generate them.
This might result in bankruptcy of some of the holding companies holding data center debt, but with that debt discharged in bankruptcy the data centers will be quite profitable and keep operating.
Training demand will also collapse after a crash, but training is now a minority of data center demand.
IOW, those electric bills are going to continue to get paid. I wouldn't want to hold OpenAI/Anthropic stock or data center debt though.
I think the risk is a little different. It's not that data centers are being built and then unused -- the fact that OpenAI, Anthropic and pretty much everyone has to put a lot of effort into getting enough compute tells you the demand is there for inference. And it's useful, increasingly so outside of coding.
The risk is in the financial world. The worlds central banks printed money because they thought inflation was tamed, but it wasn't, we just don't measure stratospheric stock and bond valuations as "inflation". I mean, euro-zone bond yields went negative -- but no, that's not inflation-like price mania at all! A big correction in financial markets towards some semblance of reality could amount to a psychological shock on consumers, a big negative wealth effect.
And a hit at the ballot box, hence no major political parties calling for raising rates (or sound money). They rather juice the markets for as much of their careers as possible.
Better question is at what rate is there demand for what amount of inference?
Quoting the article:
> A boom in railroad construction, 33,000 miles (53,000 km) of track were laid across the country between 1868 and 1873.
> The railroad industry was the largest employer outside of agriculture and involved large amounts of money and risk. A large infusion of cash from speculators caused spectacular growth in the industry and in the construction of docks, factories, and ancillary facilities. Most capital was involved in projects offering no immediate or early returns.
Sounds like the AI bubble.
> Sounds like the AI bubble.
Every new technology causes a financial bubble, and railroads weren't even the first:
* https://en.wikipedia.org/wiki/Technological_Revolutions_and_...
> After 1850 two banking institutions issued 70% of global bonds, the Rothschild family and the Barings Bank.
> On 9 May 1873, the Vienna Stock Exchange crashed because it was unable to sustain the bubble of false expansion, insolvencies, and dishonest manipulations.
well lets not do those things again shall we
Can someone explain what the leading sentence means? Why is it written like this?
> ... lasted from 1873 to 1877, continuing until 1879 ...
Isn't that just "lasted from 1873 to 1879".
Or is it written that way intentionally?
> that lasted from 1873 to 1877, continuing until 1879 in the French Third Republic and in the United Kingdom of Great Britain and Ireland.
Reading the entire sentence makes it clearer.
I read it as 1873 to 1877 for North America and most of Europe, but contining until 1879 in France and the UK.
Why do people sell catastrophes?
Bitcoin, or the AI bubble, or Peak Oil, or some panic or collapse?
No doubt something's going to fail, but I have zero confidence in the ability to predict any of it.
My prediction: they're all wrong.