TonyFettes

Preview of ds4.mbt: native MoonBit bindings for github.com/antirez/ds4, running DeepSeek V4 Flash locally on Metal. Streaming reasoning, local GGUF inference, no API calls. Thanks @antirez for such an amazing project. @moonbitlang #MoonBit

The best harness engineering is the one shipped with language toolchain. `moon coverage analyze` works amazingly well with GPT-5.5, helping me build great confidence in the 30k LOC of MoonBit it spits in one session. #MoonBit @moonbitlang

Closed enums force you to nail every case upfront. Need a new variant from another package? Fork the source or wrap it. MoonBit's `extenum` is open — declare it once, and any package can plug in new constructors. Perfect for events, messages, plugins. @moonbitlang

xterm.js meets Rabbita. Use a real xterm.js terminal from MoonBit with Rabbita-managed lifecycle, subscriptions, resize/title events, and output. The demo replays an asciinema recording as a proof of concept. moonbit-community.github.io/rabbita_xterm @moonbitlang #MoonBit

Elm-like async IO (process spawning here), where effects, subscriptions are all fully extensible and testable. Made possible with @moonbitlang #MoonBit

Libadwaita actually looks pretty solid on macOS... Built with #MoonBit, btw. @moonbitlang

Made GTK working in MoonBit! @moonbitlang #MoonBit

Prototyping a MoonBit Objective-C bridge... @moonbitlang #MoonBit

"Mom, can we have weak references?" "We have weak references at home." Weak references at home: @moonbitlang #MoonBit

🥳🥳🥳 moonbitlang.com/blog/compiler-…

Quelqu'un m'a envoyé ça et faut que je le partage :

Use a string of mirrored current sources to adjust them all with a single analog potentiometer:

Imagine having this much automation in interactive theorem proving in Agda/Lean

If you're interested in my talk, please checkout the video at youtube.com/watch?v=Bw2bd_…

Arch Linux

Presented our work on how to design a PL for LLM code gen at @llm4code yesterday! If you're interested, please checkout our X account @moonbitlang and play with MoonBit at try.moonbitlang.com

@WenzhangYang 刚刚在 ICSE 听了一个类似的 talk，也是做 code translation，GPT-4 被 c2rust 暴打...

编译器比赛，用LLM的拿了冠军。哈哈哈哈哈哈哈哈哈哈哈哈哈哈

Attending ICSE today, thank @moonbitlang for this opportunity! #ICSE2024

@HKSHAO though I'm not 100% on AGI, I'm pretty sure DSL/constrained LLM output space might be a good way to reach super-human performance...

@Ryauou libgen