I've been making the comparison to the first synthesizer in 1897. It was so huge, that it took up the basement of an entire city block in NYC. It was enormous, and relatively useless but it worked...technically. But 60 years of development later, and it can be put in a suitcase and carried around.

I see data centers and AI the same way. Sure, we can technically do it, but it's big, unwieldy, and wasteful. It clearly isn't ready for prime time.

Go back to the drawing board, address the real problems, including regulations, and get back to us in a decade or two, when they've figured out how to do this properly. Because right now it's a monstrosity that's more of a curiosity than an useful product.

I agree with this.

That lines up on the technical side, too, with how LLM “intelligence” is plateauing, cost of cheaper models is decreasing, where open weights are going and such. Q3 2027 seems about when an OpenAI coding subscription makes no sense.

That being said, I’d be wary of the “Facebook effect.”

Once a service gains a huge foothold, it can deteriorate for a long time without going away. Especially with regulatory capture. And for many smartphone users, OpenAI is the only AI they know.