Maybe, but how longs is eventually? I love GPT, use it every day but it is pretty flawed. 10 years?

The machine building itself is end game. At that point you're suggesting singularity. I don't see how any profession survives that.

There's so much hype around AI right now, it's absolutely unhinged. Yes, we have semi-conversational AI. Yes, image detection is pretty good. It's all supervised.

Can we please touch grass?

who knew what chatGPT was 2 years ago? things are moving fast, and gaining speed. will they plateau before getting to being ubiquitously useful?

LLMs have been around since the 50s. Chomsky has plenty to say on them [1]. It's not half as rosy as the current hype cycle.

1. https://www.nytimes.com/2023/03/08/opinion/noam-chomsky-chat...

Thing is, GPT is just another programming language. It no doubt brings a leap forward in approachability, like Javascript was a leap forward from assembly, but at the end of the day it still relies on programmers to use it. No matter how advanced it becomes, as long as it exists to serve the interests of humans, and not other machines, it is going to require programmers to describe what humans want to see carried out.

It is possible that everyone will become a programmer, like everyone has become an elevator operator, but unlike an elevator button press that takes milliseconds and you're done, does it behoove "the boss" to spend their day describing their ideas to GPT? I suspect, like today, they will still want to hire other humans to bridge that gap so that they can spend time doing more important things.

At that point, the same will likely be true of most doctors, with DoctorGPTs making most of the cognitive decisions and a team of lower knowledge technicians and nurses doing most of the hands on work.

Other than surgeons, a ton of what we would traditionally think of as doctoring has already been abstracted away and work specialized and divvied up to technicians, with MDs pulling strings in the background.