Yes, we’ve been hearing that since 2010 or so. It will be right (again) eventually, probably before we hit AGI. But it’s hard to plan around an event that could come tomorrow or 10 more years from now.
It seems like a silly meme. There was an “AI Winter” with a catchy name once so everyone is always waiting for another one. It’s exactly the kind of stochastic parrot thing that makes LLMs seem so human.
It's not about the efficacy of the techniques we're seeing deployed, it's about the gap between capabilities and promises and the use of AI as a magic keyword for fundraising and relevance. For example, github is now branding itself as "AI Powered".
I've been making a living with many of the tools and techniques from prior to the original "AI Winter" -- but the companies that produced them and that were promising expert systems and other forms of knowledge engineering have very few survivors.
When Meta, which is basically dying advertising company, is promising AI as it's core innovation center, we know that it's a shibboleth for signaling some participation in a notional future economy.