This is almost certainly the answer.

Most people are not using Python because it’s the “best” language but because there’s significant mindshare and a huge cost to reproduce complex libraries. If anything, this is what’s slowing alternative language development, not LLMs.

I’m not arguing language development will or should stop, but the set of people developing new languages is a tiny proportion of coders. I’m not sure why people using Claude, etc. should make a difference there?

I can ask Claude to read language docs and try to code something basic. But I acknowledged in my reply that there would have to be some work done de novo.

If you don’t opt out, it’s using your interactions to train the next model. And translation of code across languages seems to require much less training data. But yes, seems like someone needs to write the initial code bases. That said, if we all start using Claude, less incentive for new langs?

An animated graphic showing wingbeats of different birds. This one is so beautiful.

Image: Eleanor Lutz

tabletopwhale.com/2014/09/29/f...

👆