Ask HN: When the AI bubble is going to crash?

It comes down to how much money is available to invest in new AI companies. Eventually, the money will dry out, it always does, and the whole house of cards will fall apart. My guess is that the FED will need to tighten interest rates at some point because of inflation and everyone will panic. The ones that are leveraged will close their positions and others will follow. The whole bubble will collapse. It won't happen this year but next year is open.

The whole IPO craze has not gotten into gear so we have some time. The current AI darlings will go public at some crazy multiple to earnings and people will be amazed. We should all worry after that.

For techs, people are saying that it's too late to get AI skills but I disagree. There are still a few more years to go. But don't wait too long that too will collapse.

I am not an expert on AI but if I had to make an uneducated guess it will probably plateau at some point once business run out of ways to extract value from it. I think the sign this has occurred is when the same businesses that laid off large percentages of their staff start to re-hire them and / or start to create their own in-house AI that leverage their own big-data platforms they did not previously realize they needed. I am not sure how to define a crash in this context as not everyone will be able to create their own in-house solutions.

The best predictor of a crash is seeing money start to flow into another investment. Eventually the masses notice, get "I want that instead!" fever, at which point the crash occurs.

If next in line was obvious the crash would have already occurred, so I cannot say for sure, but I suspect that AI is still at this point the "must have" investment. When a crash is imminent you start to hear rumblings about the next big thing and I'm not hearing them right now.

The promise of replacing labor costs with bots is too tempting for Wall Street. The money will keep on rolling in.

Alas, if I could predict that with even mediocre accuracy, I could make some serious coin.

when china publicize they have fabs better than Taiwan (or they get the ones in Taiwan).

the last bubbles had no purpose other than limiting china and Russia access to chips so they can't pump modern arms and flood the markets.

it's not a bubble to me. I just finished a frontend project in VUE in three days, which would take me 3 months without ai.

A cheap person like me, who never bought bitcoins, VR/AR, just paid $200 for claude code subscription.