I would be utterly unsurprised to see an AI crash in the next 24 months, leading to another AI Winter. I lived through 1999 and Petfood.com and the Internet bubble pop. And I can pattern match.
But the Internet crash didn’t last long. Google and Amazon survived just fine, Ruby on Rails was big within half a decade, and soon enough we were doing Web 2.0 and AJAX and all that fun stuff.
It’s possible that current generation LLMs might hit a wall soon, for various architectural reasons that are obvious to many people but that I’m superstitiously averse to amplifying. If they do, that increases the chance of an AI Winter until the underlying research gets done.
But I have trouble imagining any series of events that buys us 10 more years. Bubble pops in tech are usually an early correction that wipes out a Precambrian Explosion of dumb money, and that ultimately concentrates resources into a few successful players.
I would be utterly unsurprised to see an AI crash in the next 24 months, leading to another AI Winter. I lived through 1999 and Petfood.com and the Internet bubble pop. And I can pattern match.
But the Internet crash didn’t last long. Google and Amazon survived just fine, Ruby on Rails was big within half a decade, and soon enough we were doing Web 2.0 and AJAX and all that fun stuff.
It’s possible that current generation LLMs might hit a wall soon, for various architectural reasons that are obvious to many people but that I’m superstitiously averse to amplifying. If they do, that increases the chance of an AI Winter until the underlying research gets done.
But I have trouble imagining any series of events that buys us 10 more years. Bubble pops in tech are usually an early correction that wipes out a Precambrian Explosion of dumb money, and that ultimately concentrates resources into a few successful players.