tl;dr: Lattner: "I didn't realize Julia aims to be a new Python. Our language Mojo also aims to be a new Python, but Mojo's approach is 'We won't try to get people to learn a new language. We'll instead make Python better.'"
Interesting to see him first try to make Swift for ML a thing and then go back with a new approach of "Actually, let's instead start with Python and make that better. Meet people where they are." Which is a big part of making a successful and impactful product. I think people making new languages often neglect this aspect.
I dunno man, I just started using Scheme (finally!) and now I'm mad all over again at all of us programmers for just not using Lisp. Like, WTF? It's older than I am! It's like, how much time and energy should we waste on syntax and "languages"? Hasn't there been enough suffering and woe?
I know I sound like a ninny, but I'm seriously y'all.
We should just use lisp.
I don't even like lisp. I'm a Python fan. But now that I've seen for myself what it is, what it's been since half a century already, I just see (almost) all the other programming languages as waste.
Maybe it had to be this way, but frankly it just feels like the human mind has an unfortunate tendency to fixate on abstractions and then refuse (likely for valid caloric reasons) to climb out of a local maxima.
Really, the situation reminds me of the Cambrian explosion. Maybe we really do have to try out all these variations to eventually settle on bilateral symmetry and four legs...
Interesting to see him first try to make Swift for ML a thing and then go back with a new approach of "Actually, let's instead start with Python and make that better. Meet people where they are." Which is a big part of making a successful and impactful product. I think people making new languages often neglect this aspect.