If archive dot ph can't scrape it then neither can Google (e.g. there is no fix that doesn't also greatly harm SEO). We might see that change as sites lose more traffic to the AI summaries?

For that reason I feel like the CPython JIT and mypyc are more likely to be winning solutions for "I am writing Python and I want some bits of it to be faster".

The problem I personally would like to see solved with AoT compilation is actually _distribution_. Python is mostly fast enough for me

I think that's what they are going for but I don't know how well it will work honestly (since for performance you won't be able to run in Python any more - you'll be writing Mojo). I think Julia has a similar problem ("like" Python, but not Python enough).

Mojo is interesting, but does lots of things that definitely do not feel like Python to me (e.g. traits instead of protocols, ownership/lifetime annotations instead of GC). I feel like if I wanted to make that tradeoff I would just write a Rust extension module.

The world if there was a statically typed AOT compiled Python

Hot take: it was a good thing to do. Old stuff _should_ be killed sometimes and the language is better for it

Did you miss the classic "I was actually talking about my _other_ left wing party" retraction of that statement?

www.thenational.scot/news/2530220...

> He has now clarified the Corbyn-Sultana project’s stance, saying: “We’re so far away now from having a position on that.”

I could see this happening if assert 1 breaks with an UnboundLocalError instead of an AssertionError but I did not know this was possible otherwise. This is cursed

On the bright side I am sleeping on a train tonight.