Colin Roberts

We should study how to convert X ego posting into power. That’d push us up Kardashev

@zkbrunny Oh a Thiel video. Probably equally enlightening.

@Autoparallel its one of those things that sounds good and catchy but doesn't hold up to scrutiny (like below) x.com/jawwwn_/status…

This timeline is so tiring sometimes. Now I’m not supposed to introspect? Come on man, that is the most ridiculous thing anyone can come up with. I promise you, whoever you are reading this, if I hadn’t spent as much time introspecting, I wouldn’t be half the person I am today

@zkbrunny I’m not sure he’s joking and I also think people follow what he says no matter what it is. It’s sad.

@Autoparallel its purposeful rage bait

@0xz80 @real_philogy C++20 and beyond are excellent. I just wish I had Cargo. And I wish modules were more than a pipe dream. lol

@Autoparallel @real_philogy modern C++ is so much nicer than what I had to learn

idk who needs to hear this but I promise you your app does not need to be memory safe it’s ok to pick another language that doesn’t make you fight the borrow checker

@real_philogy @0xz80 Rust developer experience is peak. It’s also partially because so much happens at early stages of compilation. I’m finding C++ to be pretty nice given that it’s far lazier (later) with type checking and auto/overloading allows for more flexibility. Never thought I’d say that.

I personally go back and forth. Rust has so many nice things (amazing LSP, great packages, nice stdlib) but the borrow checker is persistently annoying yes. You get better at managing it but it never fully goes away. The main contender I see is Zig but playing with their stuff the stdlib, language, ecosystem & specifically LSP really aren't of the same caliber, it's a non-trivial trade off.

@itsstock They were the OG slop band

@Autoparallel do elaborate...

Did Nickelback pioneer AI slop?

@ZeroDarkDev Classic

@Autoparallel Hard to rewrite stable 20 yr old libraries. Unlike all the web devs that argue over languages. It's pretty much all of 2 or 3 here and we link against all of them. Heh. Started to write my own MPI for Zig years ago. Quickly gave up. It's somewhere on GH. Lol.

Who’s doing scientific computing in Rust? Show me these people, X algorithm

@ZeroDarkDev I've also been thinking about Zig for HPC. Again, I don't think this is common, but maybe it'll grow in popularity. C++ is just... C++.

@ZeroDarkDev I'm not surprised! Just wanted to see if there was anyone out there

Anyone interested in: - Rust / Zig / C++ - Mathematics - Machine Learning - Scientific Computing Come join us here! discord.com/invite/qEwPr3G…

@real_philogy C++ has to go through the C library flow which is just a bit of a headache. I think there’s work on direct C++ support as with C, but it isn’t there yet. I also learned Zig is a whole C++ build system lol.

@Autoparallel Not sure about C++ but Zig has really nice support for binding with and dealing with C libraries

Zig seems like it could be amazing for scientific computing / HPC Especially if you could just pull in C++ with no effort.

If anyone wants to contribute, feel free to just make issues, put up PRs, or whatever. I'm going to try using `joy` more if I can!

With all that said, I'm going to keep it up here and there to see if this is something AI agents can actually do today. Even if this workflow isn't ideal (it isn't; vibe coding isn't), then I'll still surely learn something.

Alright, I've tried actual vibe coding. This is a project that I wouldn't want to build myself and it's also one where we can follow other examples, so it seemed like a good one to try. joy: a sane C++ build tool github.com/harnesslabs/jo…

@gakonst @itsstock

best ML researcher and paper author skills you've seen?

@moodlezoup @eddylazzarin 'bout damn time. Love the work on Jolt, guys ♥️

[1/2] Jolt is now a "zkzkVM" 🤫. A somewhat subtle point that differentiates our approach from most other zkVMs: Jolt proofs are "directly" ZK, while others rely on "wrapping" their zkVM proof with a different proof system (often Groth16) to achieve ZK.