“Only” 10 years makes it sound shorter than it should be. 10 years is way, way longer than any bubble should last.
“an economic, asset, market bubble lasts for about 5.6 years or about 67.5 months. 98% confidence interval indicates a range of 3.1 years to 8.15 years. Thus, as per the data, there is a 98% probability that a bubble should have an age of 5.3 years ± 2.53 years.”
p.s. I love that their data goes back as far the Tulipmania in 1634-1638.
More on that in the excellent book “Extraordinary Popular Delusions and the Madness of Crowds.”
(Public Domain!)
“Only” 10 years makes it sound shorter than it should be. 10 years is way, way longer than any bubble should last.
A lot of folks whose first language isn’t English use “only” in the way OP did. It’s not to signify a shorter amount of time than people think it ought to be. It means in this case that the AI bubble would not pop before 10 years is up.
Yeah, I’d say there’s a difference between “in only 10 years” and “in 10 years only,” though this is coming from a non-native speaker as well 😅
It is annoying, though, that there is no (clear,) concise way to say “pas over 10 jaar” like I can in Dutch.
I don’t think that’s possible. So much is being spent on AI, that the amount of revenue it needs to generate is unreachable. Either lots of money starts coming in soon, or the ability to spend $100s of billions more evaporates. If the bubble were to continue for 10 years, it would have to have generated enough revenue to survive, while continuing to spend more than it could ever make.
Yes this seems to be the case
https://wlockett.medium.com/you-have-no-idea-how-screwed-openai-actually-is-8358dccfca1c
OpenAI in particular sounds very unprofitable and keeping it running another 10 years will take a lot of cash injections. I wonder how long before investors give up on it? Probably less than 10 years.
This is common knowledge, just look up “inflation ai '34”. The time frame is different but it’s close enough of a projection
Your question presumes this is a bubble that will pop (which I agree with), so valuation is overinflated by hype and…let’s say clever accounting.
For that to extend another 10 years would mean keeping the hype train running. “GPT 12 is the one that will have AGI!” I imagine there would also have to be several Good Excuses™ along the way. “Data center constructors are taking too long.” “The local community won’t give us their water supply.”
I don’t think there’s enough wealth in private equity to continue this level of spending for 10 years, so we’d see tax dollars going into these AI companies.
That somehow continuing for 10 years and then we realize these things don’t generate profit would be bad. Deflation, retirement accounts wiped out, mass unemployment, people miffed that all the tax they paid into AI is gone doing a mutiny.
The housing bubble went on for about ten years.
The tech bubble lasted for about five years.
If the bubble doesn’t burst in 10 years, it might be that it’s not a bubble.
I’ve been hearing about the housing bubble for my entire life. It hasn’t burst. I think maybe we’re using the wrong word there.
AI is weird one because there’s such a mind boggling amount of investment in something that hasn’t brought any financial returns yet. Either the Visionaries see something we don’t, or it’s going to collapse or contract after one or two more earnings calls.
Housing is a bubble, but we won’t see problems until the boomers start dying off in quantity and there aren’t enough people to buy/rent their properties.
All real estate will start being devalued and it won’t matter if private companies are snapping up the vacancies.
we’d just go back to paper
This is actually more plausible than many think.
In that case it would be because it provides genuine value
