the reason devs are waxing is because software has an interesting property: it's testable. testability allows us to clearly see whether ai is making shit up or not: if it's novel and it works, it doesn't matter whether a person or an ai wrote it.

and as exciting as ai writing functioning software is, what's really interesting is that ai is getting measurably better at it, and it doesn't appear to be "just at software".

if you want to feel these vibes, pick something that you know and make a test: can you teach ai how to do it, and can it do something adjacent without you explaining how to. you can do all of that using just a chat interface, but some models will be way better than others. you should at least run the experiment with claude