I feel that this is because people have been reinventing the wheel a lot in IT.
A lot of innovation that finally lands in programming languages has been done decades earlier, but now it's mostly awkwardly bolted on existing languages and frameworks.
I really enjoy working with Gleam which is a rather boilerplate heavy language. However most of the work is managed by a really well built LSP. If you're writing a strongly type language it's easy to programmatically manage all of the boilerplate. You could you an LLM but why bother when it's a so formulaic.
Yeah, like imports in Java too. Any IDE lets you type List<ctrl-space> and it imports java.util.List - or java.awt.List if you select that one instead!
That's not what's really being discussed here though. Since it's already automated, you don't need to automate it with AI.
To further onto this, humans could technically simulate Turing machines themselves but they don't.
Efficiency and finding the right abstractions and tradeoffs is literally identical with making new capacities and capabilities. If all of your systems were truly "run by AI", we wouldn't have systems. Systems are systems because the AI uses well developed tools and protocols that humans have put together. It's a stable capacity.
A lot of innovation that finally lands in programming languages has been done decades earlier, but now it's mostly awkwardly bolted on existing languages and frameworks.