Some nuance: try transpiling to a garbage collected rust like language with fast compilation until you have millions of users.
Also use a combination of neural and deterministic methods to transpile depending on the complexity.
I don't know what languages you might have in mind. "Rust-like" in what sense?
If going to complain about some of those being slow, remeber that they have various options between interpreter, bytecode, REPL, JIT and AOT.
V-lang is the one I'm tinkering with. It's like rust in terms of pattern matching as an expression, sum types, ?T instead of exceptions.
Like golang, it has shorter compile times.
I try to keep my argument abstract (that you need to lower python to something intermediate before rust) for that reason.
I don't have a whole lot of experience hand writing v-lang. Mostly machine generated from static python.
But I find it convenient for what it does. Golang that is less verbose, single binary for distribution and fewer tokens if you're using an agent.
GitHub.com/LadybugDB/ladybug-vlang has a CLI I recently wrote with an agent for a database I maintain.
Static python with design by contract can be a stronger specification than natural language. @antirez was discussing this on his social media today.
Nonetheless, I would not bet anything too serious on such a small language.
Arguably, only the explosion of AI slop (maybe it's more profitable), has slowed down the outrageous bombardment of Zig and Rust peddling. Languages which are not even in the top 10, but one would not have known that, by the number of previous HN headlines.
https://x.com/arundsharma/status/2033635906532638776
You should judge the language based on its license and what it does, not on the morality of people who wrote it and sponsored it.
Either produce an alternative language which does the same thing written by more "unshady" people and supported by a large corporation (talking about the "small language" comment here).
These human emotions won't matter in 6 months as code will be increasingly written by machines which have no such considerations.
The "AI" makes a few changes of the free content it sucked up, then gives it back based on correct prompts, for a fee. They got it for free, but you will pay a fee. If it is not something simple, it's often riddled with errors or takes a huge numbers of prompts to get it right. So some time was saved in typing, but then lost in code review and fixing errors. Humans that can understand what they are doing, troubleshoot, and architect are still needed.
People got interested in the language over wildl claims that are simply CS-wise impossible. Later, these were removed from the site and you are left with a.. not too interesting language. I'm judging it on this late phase, with a bad taste in my mouth from the former shenanigans.
> in 6 months as code will be increasingly written by machines
Come on..
So what I do now (since Claude Code) is write really bare bones (and slow) pure python implementation (like I used to do for numba, pypy or cython ready code), with minimal dependencies. Then I use the REPL, notebooks and nice plotting tools to get a real understanding of the problem space and the intricacies of my algorithm/problem at hand. When done, I let Claude add tests and I ask it to transpile to equivalent Rust and boom! a flawless 1000x speed upgrade in a minutes.
The great thing is I don't need to do the mental gymnastics to vectorize code in a write only mode like I've had to do since my Matlab days. Instead I can write simple to read for loops that follow my intent much better, and result in much more legible code. So refreshing!
And with pyO3 i can still expose the Rust lib to python, and continue to use Python for glue and plotting
I wish someone writes a stdlib without using it. My attempt from a few months ago in a repo under the py2many org.
Quite hard to lose the #1 reason people use the language for.
Most people don't even know what C-API is or why it slows things down.
Compatibility is made into a bigger deal than it is. That's the COBOL argument.
I wish the python community focused more on why openclaw and opencode are getting written in typescript, not python.
Why aren't agents more efficient at translating python code into shippable end user binaries using fast interpreted -> compiled agentic loops and attempt memory safety only for binaries/libraries with a large distribution.