saljamMar 15

this ai bubble is going to burst any minute now, and we all know what that means: cheap data centre rack space in huge quantities.

can we use server racks to grow salad instead of swindle? i set out to find out: https://sa.lj.am/rack-mount-hydroponics/

Show threadBradley M. KühnMar 15

It *will* burst,but studying past bubbles (incl. industrial revolution),it's years until burst after an obviously ridiculously overvalued market (eg: pundits predicted a DotCom bubble burst in 1996)!

IMO this'll last *longer* than DotCom—in part—b/c most wealthy tech bros invest heavily in #LLM-backed #AI.

Analogy: ∃ enough money to keep music playing. Musical chairs doesn't stop until music ends.

If only we *could* predict burst date: we'd get so much 💰 shorting that we'd fund #FOSS forever!

Show threadJessie HealdMar 15
@bkuhn Do you think that it _could_ burst in the near term, though? All this compute is absurdly expensive for something that doesn't provide the value it claims.
Show threadBradley M. Kühn
I played #poker for a living (part-time) when I worked for #FSF & SFLC as I was paid so little I couldn't afford {Boston/NYC}-area w/out 2ⁿᵈ job.
I'm reasonably good at making odds for real-world situations after all that poker. I have followed #LLM-backed #AI closely enough, & Wall Street is just a big casino really, so here's my predictions:
IMO, AI bubble bursts (≥15% drop in 90+% AI-related stocks in ≤3months):
≤1yr: 30-to-1 against
≤3yr: 3-to-1 "
≤5yr: 1-to-10 favorite
≤10yr: 1-to-25 "
Mar 15 at 5:51pmWeb
Show threadBradley M. KühnMar 15

Just in case it needs to be said: I am not a qualified investment advisor and the post that I'm replying to is not investment advice.

These are just predictions based on the information I have because I've followed carefully the rise of LLM-backed AI and related technologies, and my experience studying the policy and market implications of past tech bubbles.

Please do not invest based on my advice. I could easily be very wrong.