Mildly horrifying observation here at neverworkintheory.org - ML models relying on existing large corpuses of data mean that future new languages have a huge ergonomics hill to climb as far as tooling is concerned, because that corpus doesn't exist.
@mhoye Yeah. And that applies to new frameworks and platforms as well. Not sure how ML copilots are going to work with programming’s innate novelty-seeking in the long run.
