Similarly to how the dot com bubble DID create a handful of businesses that remain dominant to this day, but also a ton of garbage that never had a future. We're gonna have a handful of companies that actually revolutionize computing, and a whole bunch of crap that we'll forget by the 2030s.

Godspeed to those who think they can predict who the winners will be. Hopefully you don't bankrupt yourself on a bad bet like so many in 1999.

The fact is, if you're only looking at one type of AI in a bubble, sure.

As it stands, AI has already provided a plethora of new opportunities. AI bone-rigging and animation makes these things super fast. AI audio is replicating voices, making audiobooks, games, etc that much more interesting, in places where professional voiceovers couldn't be afforded; the vast majority of middle-management can probably now be replaced by a small (AI assisted) shell script.

It's not a bubble, and will drive a new future of tensor processors and optimizations in the silicon space.

It will cool down, for sure - but it's not a tech bubble like we've seen in things past.

However: It will continue to drive separation between the haves and have-nots. There's a new generation of people who will either adapt or die. If you can adapt, and roll AI generation into your workflow - you're going to THRIVE. If not, you're going to be left by the wayside.

I'm running LLMs locally on my machine through a program called Jan. We're not going to need warehouses worth of datacenters for this stuff. Just like we needed warehouses for a single computer back in the 70s, and have supercomputers in our pockets everywhere we go today; the same will happen with LLMs, Stable-Diffusors, etc.

There were lots of people saying computers were "just a fad" back when I started this journey. AI is just the next thing. We don't have AGI yet (Artificial Intelligence in the generalized sense) - but we may at some point during your lifetime.