Yet another swoon piece.

I don't see how the OP can say these apps worked "perfectly", given the nature of what they're doing (and the lack of either a metric or an easy to way to verify their outputs).

The repo summary, in particular, seems to be largely a regurgitation of its README file, along with some claims (about that the kinds of applications it can support) that sound, well made up. In any case not straightforward to verify from the contents of the repo.

It's certainly "a form of" programming. But it seems quite strange to call this "the new programming".

Prompt engineering is as engineering as data science is science.

$5k and two weeks included what? A lot more goes into proper design and deployment than some code you run somewhere.

But most devs don't do more than gpt is doing I have to admit.

I have to wonder if OpenAI does this internally now.