kaidon

[mousepotato @iluciddreaming]

我真心建议，如果你目前没有工作的话，

可以下载 Docker，自己学习安装 n8n + openclaw，然后把 webhook 设置好，让它们自己能够互相沟通。

之后你可以在简历上写，除了原本的专业，你还会 AI Agent，可以帮公司做办公室自动化流程，

再把你做的小作品展示一下。现在很多小公司或微型公司都很需要这样的人才。

[阿西_出海 @axichuhai]

这个 AI 搜索skill 已经登顶GitHub Trending ，斩获 4w star

它把 Reddit、X、YouTube、HN 的近30天热点全部进行抓取，会按照真实互动数据加权排序

搜之前会先弄清楚你在找什么，把人名、公司名解析成对应的账号和社区，再决定去哪搜。

如果同一件事在 Reddit、X、YouTube 都有人讨论，它会自动合并成一条，不给你三份重复内容

生成的简报是 HTML 文件，直接丢进 Notion 或邮件就能

很适合竞品追踪、话题研究、内容选题

[— 𝙰𝚕𝚊𝚜𝚔𝚊 @alaskyne]

Cara lain nonton pildun gratis tanpa link illegal adalah pakai aplikasi VLC terus masukin link IPTV Public dan hasilnya bakal muncul semua channel public tinggal pilih aja mana nyiarin pildun. Tanpa iklan mengganggu 👍🏻

[铁锤人 @lxfater]

现在任何都能永久免费用上Claude code！！

下面视频教你配置👇

[Nav Toor @heynavtoor]

10 GitHub repos that defined 2026 so far.

Bookmark this list.

1. OpenClaw
Peter Steinberger went from 9,000 to over 300,000 stars in months. Personal AI assistant that runs entirely on your devices. Steinberger joined OpenAI shortly after.
Repo → https://t.co/9B3jxI7HwY

2. anthropics/skills
135,000 stars. The patterns Anthropic uses internally to extend Claude. The repo that defined the skills ecosystem of 2026.
Repo → https://t.co/6VBUdubDG0

3. affaan-m/everything-claude-code
141,000 stars. The complete library of Claude Code skills, agents, and commands. The aggregator every serious builder forks.
Repo → https://t.co/PaXno1QqDi

4. andrej-karpathy-skills
Forrest Chang built a single CLAUDE md file in January. 109,000 stars. The most starred single-file repo in GitHub history.
Repo → https://t.co/unItpr073y

5. Hermes Agent
Nous Research released this in February. 105,000 stars. The self-evolving AI agent that gets smarter the more you use it.
Repo → https://t.co/OMgRfKAts4

6. obra/superpowers
94,000 stars. Officially accepted into Anthropic's skills marketplace. Multi-agent orchestration without the boilerplate.
Repo → https://t.co/Z7i3fzf4s4

7. claude-task-master
Multi-agent orchestration on top of Claude Code. Turn one prompt into a coordinated team of specialists shipping a feature while you sleep.
Repo → https://t.co/0xYzJpSX4z

8. MemPalace
Milla Jovovich, the actress from Resident Evil, co-built this with Ben Sigman using Claude Code. Near-perfect score on the LongMemEval benchmark.
Repo → https://t.co/o8xKSTz60D

9. karpathy/autoresearch
Andrej Karpathy released his own research automation framework. 23,000 stars in three days. The closest thing to having Karpathy as your research partner.
Repo → https://t.co/YURNnYJJN3

10. karpathy/nanochat
Karpathy released a complete pipeline to train a ChatGPT from scratch for $100. 54,800 stars. The best ChatGPT money can buy is the one you train yourself.
Repo → https://t.co/BZYF0qXE9z

Save this. Share it with the developer in your life who deserves to know what they missed.

100% free. 100% open source.

[stellar protocol @stellarprtcol]

Anthropic mungkin nggak akan suka melihat project yang satu ini. 👀

Ada developer yang baru saja membuat alternatif Claude yang bisa dijalankan 100% secara lokal.

Dan yang bikin menarik, project ini bukan sekadar chatbot biasa.

Kemampuannya cukup serius:
✅ Bisa digunakan dengan berbagai model AI
(Claude, GPT, Gemini, vLLM, dan lainnya)
✅ Dirancang untuk workflow deep research
✅ Punya kemampuan mirip fitur Cowork
✅ Sudah dilengkapi 50+ connector bawaan
✅ Deploy hanya dengan satu perintah

Yang lebih gila lagi...
Semuanya open-source dengan lisensi MIT.

Dan saat gue melihat repo-nya, project ini sudah mengumpulkan 28.000+ stars di GitHub. 🤯
Menurut gue, tren yang mulai terlihat sekarang cukup jelas.
Dulu orang mencari model AI terbaik.
Sekarang orang mulai mencari sistem terbaik.

Karena model bisa diganti.
Hari ini Claude.
Besok GPT.
Lusa Gemini.

Tapi workflow, memory, connector, dan kemampuan agent itulah yang mulai menjadi pembeda sebenarnya.

Dan project seperti ini menunjukkan bagaimana komunitas open-source bergerak sangat cepat mengejar fitur-fitur yang sebelumnya hanya tersedia di platform komersial. 🚀

Kalau penasaran, repo GitHub-nya ada di bawah. 👇
https://t.co/AlTFZe4Hn2

[Ascenso Infinitx. @AscensoInfinitx]

EL MÉTODO DE CARL JUNG PARA HACER QUE LOS DESEOS SE HAGAN REALIDAD.

Funciona Incluso Para Escépticos.

Practícalo Durante 10 Días, Y Algo Extraño Comienza a Suceder:

HILO🧵

[Han🍀 @0xhanyfa]

KALAU LO LAGI NATA ULANG HIDUP DI USIA 30-35+, LO HARUS TAU INI DAN MULAI DARI 12 HAL INI.

1. Rapihin gigi, rawat kulit, dan benerin postur badan.

[Tom Dörr @tom_doerr]

Wi-Fi motion detection without cameras

https://t.co/4t8bf1jnk1

[Nix0n @UnTalNixon_exe]

🛠️ Deja de pagar suscripciones mensuales: 10 repositorios Open Source que reemplazan software premium
No gastes más dinero en licencias corporativas. Estos proyectos de código abierto son alternativas legítimas y gratuitas que puedes ejecutar localmente en tu propia máquina.

El arsenal definitivo:

📈 TradingAgents: Un equipo completo de analistas de IA (técnico, fundamental, sentimiento y gestor de riesgos) debatiendo y operando en tu computadora 24/7.
🔗https://t.co/jGZx3H3U59

🧠 LibreChat: Una interfaz única que unifica Claude, ChatGPT, Gemini y DeepSeek en tu propio servidor con soporte nativo de MCP.
🔗 https://t.co/qjZxGqiLeK

🎬 HyperFrames: El motor de video de código abierto de HeyGen. Genera MP4s de nivel de producción usando HTML, GSAP, Lottie y Three.js.
🔗 https://t.co/RrUVQtblgH

💼 Fincept Terminal: El clon de una terminal Bloomberg ejecutándose en tu laptop con agentes de inversión inspirados en traders legendarios.
🔗 https://t.co/Ie0QtAW2u1

📹 MoneyPrinterTurbo: Automatización de contenido. Introduces una palabra clave y genera guiones, imágenes, subtítulos, música y el video final editado.}
🔗 https://t.co/7ZjS0jJXWF

📬 Agentic Inbox: El cliente de correo inteligente de Cloudflare. Una IA lee tu bandeja de entrada y redacta respuestas sin que tus datos salgan de tu entorno.
🔗 https://t.co/HBircfkrBF

🗣️ VoxCPM: Clonación de voz avanzada. Duplica cualquier voz con solo segundos de audio o crea una desde cero con una descripción de texto.
🔗 https://t.co/ZyNbqzkOK1

🛰️ Flowsint: Una bestia para investigaciones privadas y OSINT local. Introduces un dominio y mapea IPs, subdominios, correos y carteras cripto vinculadas.
🔗 https://t.co/ZyNbqzkOK1

💻 agent-skills: La librería de habilidades de producción para Claude Code liberada por ingenieros de Google. Cubre diseño de APIs, debugging y CI/CD.
🔗 https://t.co/SI8ZaPDcz8

🔌 Nango: La capa de integración por la que las empresas pagan miles de dólares. Cientos de APIs preconstruidas y gestión de OAuth integrada.
🔗 https://t.co/HVeU0cJ363

Entornos estables, 100% gratuitos y listos para integrarse en tu flujo de trabajo diario.

Enlaces a los repositorios en los comentarios. Guarda este post en marcadores para recortar tus gastos de software hoy mismo 🔖

[PA13L0 @Fluyeporlaweb]

https://t.co/n1qPtvRTbD

Más repos así en https://t.co/ANlD2Ch1y9

[jessald @jessald]

Experimenting with graph layouts.
https://t.co/c8QnkD5fGJ

[Xudong Han @Xudong07452910]

这篇论文让我感觉，我们对「AI 会取代程序员」这件事的讨论方向可能全错了。

核心观点：AI Agent 的出现不是让软件工程师工作效率更高，而是让「把决策逻辑永久编码进软件」这件事本身变得越来越不必要。

作者说的是一个更根本的范式变化：传统软件工程的本质是，人类把判断逻辑「固化」成代码——if-else、状态机、算法，这些都是把人类决策「提前写死」的方式。但在以 LLM 为核心推理引擎的 Agent 系统里，代码变成了「临时生成、用完即扔的工具」，每次任务，Agent 动态生成需要的代码，执行完就不需要了。决策不再被预先编码，而是在运行时由 LLM 推理循环动态产生。这不是增量改进，而是软件生产范式的结构性替换。

我觉得这个观点里最值得注意的细节是：这不只是生产力工具的升级，而是「软件」这个概念本身的角色在变。以前代码是「系统的中心」，Agent 框架是外壳。现在 LLM 推理循环是中心，代码变成了外壳里的临时辅助。如果这个趋势持续，软件工程的核心能力可能不再是「写出好代码」，而是「设计出可靠的推理约束边界」。

以前我们关注写出可维护的代码，以后可能更要关注设计出可靠的推理边界。

https://t.co/28rZe5qvBP

[Rahul @sairahul1]

This is the best site on the internet to learn harness engineering.

Free. Completely.

Most AI engineers have never heard the term.

https://t.co/bwDbTTYsjM

Bookmark this site.

Then read this setup ↓

[Leonard Rodman @RodmanAi]

SAM ALTMAN HAS A NEW PROBLEM. 🤯

Google just shrunk 31GB of AI memory down to 4GB.

The tool is called TurboVec.

It uses up to 16x less memory, searches faster than FAISS, runs fully offline, and works on a regular Mac.

No expensive GPU cluster.
No cloud dependency.
No compromise on speed.

→ 16x lower memory usage
→ Faster vector search
→ Works with LangChain & LlamaIndex
→ 100% open source

The race to build bigger AI models is loud.

The race to make them dramatically cheaper just got a lot more interesting.

Repo: https://t.co/08TFGtHL6K

[0x_Miko @Mikocrypto11]

Claude Code 的创造者 Boris 讲了一个很关键的转变：

“我现在不再给 Claude 写 prompts。
我写 loops
然后让 loops 去完成工作
我的工作，就是写 loops。”
这其实也是很多人用 Claude Code 卡住的地方

他们还在把它当聊天工具用：
输入一个需求
等一次回答
然后手动继续改

但 Boris 在 30 分钟里展示的是他真实的 daily setup：
Claude Code
loops
dynamic workflow
核心不是让 Claude 一次性给出完美答案

而是把任务拆进一个可以反复执行、检查、修正的循环里
模型不只是回答你
它会在 workflow 里持续推进任务
这类内容，可能比很多 $500 的 vibe-coding 课程更有价值
因为它讲的不是“怎么写一句更好的 prompt”
而是怎么把 Claude Code 变成一个能持续工作的系统

AI coding 的关键，是 prompt 写得更漂亮，还是你能不能设计出让模型自己推进的 loop？

[Harman @itsharmanjot]

10 LEARNING WEBSITES THAT WILL REPLACE COLLEGE.

Bookmark every single one. The most valuable skills are not on a transcript.

1. https://t.co/Y72s2yrfwV
3,000 hours of free coding courses. JavaScript, Python, React, machine learning, all certified.

2. https://t.co/5yvmVy6f0H
The most respected free full stack web dev curriculum. Built by working engineers.

3. https://t.co/9hyU4IHyRK
Harvard's intro to computer science. Free. Includes graded assignments and a certificate.

4. https://t.co/w36kM7zjUf
K-12 through college math, science, and economics. The teacher you wish you had.

5. https://t.co/uuDbFwxEmk
Every MIT lecture, problem set, and exam. Free since 2002.

6. https://t.co/6BdGCxVTqW
Math and science taught through interactive puzzles. Builds intuition instead of memorization.

7. https://t.co/8OU64RkJ7f
The Barbara Oakley course taken by 4 million people. The science of how to actually study.

8. https://t.co/TEWErJLn6p
3,000 coding exercises across 60 languages with free mentor reviews.

9. https://t.co/xlyfUgvYIb
A search engine for every free university course online. 100,000 indexed from MIT, Stanford, and Yale.

10. https://t.co/viMfFnfn9B
The most underrated learning resource on the internet. Still free. Still ad free. Still neutral.

The students who win the next decade do not need permission to learn anything.

[Hamza Khalid @humzaakhalid]

THIS CHEATSHEET WILL SAVE YOU 6 MONTHS OF FIGURING OUT OBSIDIAN THE HARD WAY.

I just mapped out the entire second brain system, note types, vault structure, PARA method, zettelkasten flow, best plugins, Claude prompts,

AI mode setup, all in one place.

🔖 bookmark this before you scroll past it.

→ 15 note types (fleeting to evergreen to hub notes)
→ full vault folder structure with file naming
→ the PARA method explained (projects, areas, resources, archive)
→ best plugins of 2026, must-haves + AI integration picks
→ 10 claude prompts built specifically for vault work
→ zettelkasten flow: capture → process → connect
→ how to avoid information overload (the mistakes + fixes table)
→ role-playing modes for claude inside your vault
→ linked prompting workflow to run Claude directly from Obsidian

if i had this when i started building my second brain, I would've skipped.

the full guide in the article below ↓

[Charly Wargnier @DataChaz]

🚨 ANTHROPIC JUST PUBLISHED A 36-PAGE SECURITY GUIDE THAT BASICALLY TELLS YOU TO STOP TRUSTING YOUR OWN AI AGENTS.

If you run agents on Claude Code, MCP servers, or automation tools, pay attention.

The attack timeline has collapsed.

AI models compress the gap between a vulnerability and a working exploit from months to hours, for mere dollars.

Agents introduce new autonomous risks, from tool poisoning to context memory manipulation.

The most useful idea in the guide is Anthropic's new security test:

Does a control make an attack impossible, or just tedious?

Automated attackers have unlimited patience. They will grind straight through friction like rate limits and 2FA. To defend at the speed of AI, you need hard barriers and automated defensive operations.

Here is how Anthropic says you should lock down agents:
→ Treat static API keys as compromised. Use short-lived tokens that expire in minutes.
→ Apply "Least Agency": explicitly limit what each tool can DO.
→ Sandbox agents that process untrusted inputs like emails and web pages.
→ Scope permissions dynamically per task, not permanently.

I've added the link to the guide in the 🧵↓

[Kanika @KanikaBK]

A SENIOR GOOGLE ENGINEER DROPPED A 421-PAGE DOC THAT NO ONE IS TALKING ABOUT.

It is called Agentic Design Patterns. 100% FREE.

Every AI builder paying $200/month for courses just got obsoleted.

This is the most comprehensive AI systems guide I have seen in 2026.

Code-backed and production-ready.👇