l!ᴧǝ◖sǝ!ɹ∀ @hash_tree

if i rest i rust Joined January 2015

Tweets

19K
Followers

144
Following

2K
Likes

60K

l!ᴧǝ◖sǝ!ɹ∀ @hash_tree

a month ago

@suwakopro @mylifcc 其实没啥大用，代码量大了无解，跟go这种比编译速度是没法比的，再怎么优化都不行

1 0 1 30 0

View Details

Berryxia.AI @berryxia

3 months ago

A14Team 宣布 LabClaw 开源——这是斯坦福-普林斯顿团队为 LabOS 开发的技能操作系统层。一条命令即可将任何 OpenClaw 智能体变成一名完整的 AI 合作科学家。演示： labclaw-ai.github.io 龙虾军团报到🦞

AI4Science Catalyst @AI4S_Catalyst

3 months ago

We’re thrilled to open-source LabClaw — the Skill Operating Layer for LabOS by Stanford-Princeton Team One command turns any OpenClaw agent into a full AI Co-Scientist. Demo: labclaw-ai.github.io Dragon Shrimp Army reporting for duty 🦞🔬 #AIforScience #OpenClaw

56 260 2K 445K 1K

7 25 121 29K 174

View Details

@abdimoalim.bsky.social @abdimoalim_

3 months ago

Introduction to non-lexical lifetimes (by Niko Matsakis): smallcultfollowing.com/babysteps/blog…

0 2 12 954 11

View Details

Vivek Galatage @vivekgalatage

3 months ago

I came across two excellent articles by Russ Cox on memory models - one from the hardware perspective, the other from the programming language. Explains how concurrency works under the hood and why it matters for real-world code research.swtch.com/hwmm research.swtch.com/plmm

1 13 131 6K 145

View Details

Peter Boncz @peterabcz

3 months ago

We just released a @duckdb community extension that automatically privatizes complex queries. It is based on MIT’s new PAC framework (news.mit.edu/2025/new-metho…), and in collaboration with them (🙏 🙏). We are 👀 and eager for your feedback! paper: arxiv.org/pdf/2603.15023…

CWI DA @cwi_da

3 months ago

PAC is now available as a DuckDB community extension! github.com/cwida/pac

0 0 10 11K 6

1 4 69 10K 42

View Details

DinoDeer @xDinoDeer

3 months ago

YC 的这个CEO 有几把刷子呀，github.com/garrytan/gstack 最好用的是/plan-ceo-review，在讨论一个新的 idea，以及针对这个 idea 做 10x 的商业模式设计时候，还是震撼到了我。在这个上面比 Superpowers 专业。

13 89 510 42K 867

View Details

Vivek Galatage @vivekgalatage

3 months ago

The Little Book of Semaphores by Allen B. Downey greenteapress.com/wp/semaphores/

Vivek Galatage @vivekgalatage

3 months ago

One of the best coverage on concurrency primitives ostep.org

5 84 748 59K 680

4 83 709 39K 808

View Details

被减数 @Lessnoise365

3 months ago

Obsidian Dashboard GitHub：github.com/InlitX/Obsidia…

3 20 422 29K 495

View Details

Tom Dörr @tom_doerr

3 months ago

IDA Pro plugin for LLM-powered reverse engineering github.com/symgraph/IDAss…

1 65 287 17K 224

View Details

Martin Fowler @martinfowler

3 months ago

NEW POST Conversations with AI are ephemeral, decisions made early lose attention as the conversation continues, and disappear entirely with a new session. @techygarg explains how to externalize the decision context into a living document. martinfowler.com/articles/reduc…

15 47 407 84K 534

View Details

Rohan Paul @rohanpaul_ai

3 months ago

Alibaba just open-sourced OpenSandbox ( a general-purpose execution environment ) to give AI agents an isolated environment to run code safely. 8k+ Github stars ⭐️ This stops your AI Agent based applications from accessing your actual host infrastructure. By removing the hardest security roadblock, this release will massively accelerate how fast developers can build autonomous Agent based tools. OpenSandbox puts the agent inside isolated runtimes like gVisor or Firecracker. You can run it locally using Docker or scale it up using Kubernetes. The system includes a code interpreter and a file system that the agent uses to complete tasks. It also manages network traffic so you control exactly what the agent accesses online. I think this will become the standard infrastructure for autonomous systems because building custom sandboxes is too dangerous for most teams.

45 111 746 57K 950

View Details

Kshitij Mishra | AI & Tech @DAIEvolutionHub

3 months ago

This tool turns JSON into visual diagrams. Perfect for debugging APIs and large datasets. Project: github.com/AykutSarac/jso…

43 283 2K 82K 2K

View Details

日拱一卒王小楼 @wang_xiaolou

3 months ago

发现一个宝藏：Github 上 196K 星的 OSSU 项目，集齐了哈佛、MIT、普林斯顿等名校的 CS 公开课，从编程到算法到系统安全，该有的都有。项目方说：每周学 20 小时，2 年拿下完整本科 CS 知识体系。Discord 社区随时答疑，课程全免费，实在要钱也能申请助学金。我寻思这不就是「用爱发电」的终极版本吗？每周 20 小时是什么概念？相当于你得有一份朝九晚五的工作，然后下班回家继续上第二份「班」。周末？不存在的，继续学。所以能完成的都是什么人？要么财务自由提前退休，要么还在上学有大把时间挥霍，要么就是传说中那种「我可以」的卷王。 2.4 万人 fork 了这个项目，但我猜真正学完的比例...大概和健身卡的使用率差不多😂 github.com/ossu/computer-…

2 90 297 22K 392

View Details

Tom Dörr @tom_doerr

3 months ago

Go struct terminal output dumper github.com/goforj/godump

1 23 155 7K 69

View Details

NK @Nikaiw

3 months ago

meet VMkatz, github.com/nikaiw/VMkatz

12 245 937 69K 721

View Details

l!ᴧǝ◖sǝ!ɹ∀ @hash_tree

3 months ago

cool

mem7 @mem7ai

3 months ago

Just open-sourced mem7 — Long-term memory for LLM agents powered by Rust. •Dual-path retrieval: vector + graph in parallel •Ebbinghaus forgetting curve: decay + reinforcement •First OpenClaw memory plugin Inspired by mem0, with time dimension. github.com/mem7ai/mem7

0 0 1 681 1

0 0 0 20 0

View Details

Christos Tzamos @ChristosTzamos

3 months ago

1/4 LLMs solve research grade math problems but struggle with basic calculations. We bridge this gap by turning them to computers. We built a computer INSIDE a transformer that can run programs for millions of steps in seconds solving even the hardest Sudokus with 100% accuracy

248 806 6K 1.8M 5K

View Details

Rainier @mtrainier2020

3 months ago

这个哥们牛逼plus啊。把一个 WebAssembly（WASM）解释器 hard-coded的方式嵌入到了 Transformer 模型的权重里，而且是无损的（losslessly）。这相当于在 LLM 内部运行着一台真正的计算机。这台计算机可以实际执行计算（run computations），而不是像现在大多数模型那样，只是通过infer来给出计算结果。这个思路有点类似于TI的dsp芯片， arm负责逻辑思维，然后dsp专门负责高速数据计算。各取所长。那么类比一下，LLM 老算不清3.11 与3.8哪个大的问题可以通过构建一种混合架构（hybrid architectures）： 1. 神经网络负责推理和理解（reasoning） 2. 嵌入式解释器 / 计算引擎负责高精度计算这样就能同时兼顾智能推理能力和确定性计算精度。这对，数值计算，物理模拟，金融建模，密码学运算有很大好处。