People talk about the AI bubble like it’s going to burst and everything will just go back to normal and we’ll all frolick in the fields holding hands. That’s really stupid.
Some people do, and certainly those people are poorly informed.
Most people talking about the AI bubble are focussed on the gargantuan amounts of money being invested to build out datacentres, relying on the future value of inference.
As in, right now n inference tokens is worth $y.
Investment is occurring on the basis that in the future everyone will need nx tokens.
The actual amount of x is the problem. If it’s less than investors are hoping, and it almost certainly is, then the bubble will pop and suddenly all that investment money is lost.
A catastrophic economic outcome.
However, you’re correct that the chatbots of today will survive the coming economic storm.
I’m not quite sure what your point is there.
For every pre-dotcom tech company you can name that was “over valued once” there are dozens you can’t name because they collapsed.
To your broader point, I agree that AI is here to stay, but the current hype is BS.
Regardless, your use of the term bubble is incorrect.
People talk about the AI bubble like it’s going to burst and everything will just go back to normal and we’ll all frolick in the fields holding hands. That’s really stupid.
Some people do, and certainly those people are poorly informed.
Most people talking about the AI bubble are focussed on the gargantuan amounts of money being invested to build out datacentres, relying on the future value of inference.
As in, right now n inference tokens is worth $y.
Investment is occurring on the basis that in the future everyone will need nx tokens.
The actual amount of x is the problem. If it’s less than investors are hoping, and it almost certainly is, then the bubble will pop and suddenly all that investment money is lost.
A catastrophic economic outcome.
However, you’re correct that the chatbots of today will survive the coming economic storm.