MoonBit's new native backend will soon support **x86-64-linux-gnu**! It's now available for public testing on MacOS Apple Silicon, and the Linux x86-64 gnu version is coming very soon 😁 Ready for blazing fast native compilation? #MoonBit #moonbitlang

This is the request_id middleware in #moonback. I like the new copy-on-write Headers/MutHeaders design. It's simple and elegant. @moonbitlang

Using AI agents to write MoonBit? Use the official MoonBit Agent Skill It is actively maintained by the MoonBit team and kept up to date with the language and toolchain, so agents can generate and modify MoonBit code with better context. github.com/moonbitlang/moonb… @moonbitlang

看到一群人不理解为什么 bun 从 zig 切到 rust，宁愿合并未被充分 review 的 AI 生成的代码。 我来解释一下： bun 被 anthropic 收购了，claude-code native 跑在 bun 上面，出了很多内存泄露问题。bun 团队一直疲于应付内存泄露问题，顶不住 anthropic 这边的压力，切到 rust 是必然的选择

The MoonBit language design is mostly done in next month,and we have three months to clean up to reach 1.0. What an intense 4 years hard work , build the programming language for AI in the next twenty years!

Language features compose! See the beautiful diff by using list comprehension pattern match constructor in @moonbitlang !

中国语and日本语are同一but文法and副词完全different。 so借用some 副词from英语，中国人and日本人can相互交流

以前是：需求太多，程序员不够用，产品开发不过来 现在是：产品太多，用户不够用了

Thanks to github.com/moonbitlang/async…, moonback now have graceful shutdown support with ease. It will handle in-flight requests gracefully with timeout if it receives SIGINT/SIGTERM while serving.

🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length. 🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice. Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today! 📄 Tech Report: huggingface.co/deepseek-ai/D… 🤗 Open Weights: huggingface.co/collections/d… 1/n

Made GTK working in MoonBit! @moonbitlang #MoonBit

新时代开源，不是开源代码而是开源对话。 code is cheap, show me the talk. All chat is open source. github.com/Innei/Yohaku

A browser written in MoonBit. Unlike Cursor’s 3-million-line AI slop dead-end, this is actually maintainable, high-quality software — roughly 400k lines. There’s a bit of a paradox here: in the future, people may not care what language a piece of software is written in, because all they see is the product itself. But the language matters enormously, because the choice of stack directly determines long-term maintainability and verifiability. Some languages, like MoonBit, are designed to be provable. Others are slow by design — no matter how much you optimize, their ceiling is limited. GitHub: github.com/mizchi/crater #MoonBit #browser

Can you imagine this in other languages? @moonbitlang

I heard people are saying that formal verification is more than important in the AI era. So here comes the proof-carrying MoonBit code where all the proof obligations are discharged by SMT solvers. @moonbitlang

说个我最近在搞 Agent 时候发现的野路子。。。 大家设计 Agent 的时候是不是都在纠结一个事——这个 Agent 的人格到底怎么写？ 写了一堆 prompt 描述什么"你要有批判性思维""你要注重细节"。。。写完自己都觉得空 直到试了一个极其简单的方法—— 直接告诉它：你是某个真实的历史人物 比如我现在的 Agent 团队是这么配的： 调研官 → 查理·芒格 逆向思考天然适合筛机会，多元思维模型拿来做跨赛道分析简直绝配 试错官 → Eric Ries 精益创业那套 Build-Measure-Learn，本身就是试错官该干的事 产品官 → 稻盛和夫 极致流程、阿米巴经营，对细节的执念，做标准化产品太合适了 增长官 → Naval Ravikant 杠杆思维、金句密集，写内容天然就有穿透力 营销官 → Alex Hormozi $100M Offers 的作者，定价/赠品/紧迫性，信手拈来 你发现没有，这比你自己憋半天写的 prompt 强太多了。。。因为这些人的思维方式、表达风格、决策逻辑，AI 模型本身就学过大量语料，你只要唤醒它就行了 而且这个匹配过程你都可以让 AI 帮你做—— 把你的岗位职责丢给它，让它推荐合适的历史人物 这招真的太香了

One thing I am particularly excited about @moonbitlang is iso-morphic app, build front end using MoonBit JS backend while backend using native backend for optimal performance, the types and validation can be shared, and native SSR, checkout the repo below for more details