that is not much of a timeline and not much of a prediction. first you need to define what “crash” means in real terms. over the course of 4 years there will be market correction (especially after a bull run like we are on) so just saying “ai bubble, crash bla bla” is too lazy (although there are probably 10k+ such “predictions” on HN in the last say a year)
If you're looking for a short term prediction, I expect with, let's say 80% confidence, that the OpenAI IPO later this year will be quietly cancelled or face a WeWork moment and be loudly cancelled. Too much of what we think we know about the economics of modern AI is built on trust in people who aren't trustworthy, and the presumption that VCs would check the financials carefully when we know they strap on blinders once they see a revenue graph they like.
My feeling is that some event will happen close-to-IPO that spooks investors, that results in OAI not IPO'ing. Remember if there are under-writers involved they will not want to go forward.
Then they will face financial distress, and questions over how they get the funding to continue as a going-concern. The only way that'll happen is via issues of shares at a lower price aka destroying the valuation of OAI compared with today.
Anthropic in comparison will be OK, as they have focused on building a viable business enterprise.
Only in the same sense that Lehman collapsing caused the 2008 crash. Nothing is monocausal, and crashes rarely happen all at once, but I expect it to be a big “uh oh” moment for investors.
Lehman’s market cap was an order of magnitude smaller than the current OpenAI valuation, I think even in inflation-adjusted terms. The intuition that a private company is small and can’t pose systemic risks just isn’t true anymore.
