正议

[露西的百宝箱 @Lucy_love_AI]

10个GitHub仓库帮你爬取整个互联网

全部收藏。每个都能从任何网站提取干净数据，这种访问权限通常需要销售电话和合同才能获得。

1. https://t.co/rNvFxlszKj

指向任何网站，它就能爬取每个页面、渲染JavaScript，返回AI能立即读取的干净结构化数据。13万星，进入GitHub百大仓库。半数AI创业公司悄悄运行的爬虫骨架，完全开源。

2. https://t.co/j93MmIWNzu

GitHub排名第一的爬虫。把任何网站转换成干净的LLM就绪的markdown，比付费服务更快，无需API密钥、无需账户、无需按页面付费。某开发者被16美元的付费爬虫激怒后几天就搞出来了。5.1万星。Apache 2.0。

3. https://t.co/UA4GAQP483

像真人一样操控浏览器的AI代理，点击、滚动、登录、填表，从未见过的网站中提取数据。两位苏黎世ETH研究员开发，一年内达到9.5万星。能爬取简单爬虫无法触及的页面。MIT协议。

4. https://t.co/XYCjgK7qWL
完整专业爬虫框架，包含轮换代理、自动重试、浏览器指纹欺骗和队列管理。防止被封禁的全套机制，爬虫公司收费数千的技术栈，现在免费给你。

5. https://t.co/WhdHyC9RdW

十多年来悄悄为数据团队赋能的工业级爬虫。爬百万页面、提取任何内容、干净导出。在大多数付费工具无法触及的规模上经过实战检验，始终免费。

6. https://t.co/8PTPlhMs7J

微软自家工具，将任何文件或网页、PDF、Office文档、HTML、图像转换成AI能用的干净markdown。整个数据管道公司都在围绕此构建，由微软开源。

7. https://t.co/DPX6WLhXXg

隐形爬虫，能自动适应网站布局变化，绕过反爬虫检测。防爬供应商当高级功能出售的猫鼠游戏，现在免费开源。

8. https://t.co/2M3DkTLTxV

从电脑远程控制任何安卓手机，提取数据和自动化没有网站的应用。通往大多数爬虫无法触及的纯移动平台的桥梁。13万+星。Apache 2.0。

9. https://t.co/O2EKeQRsAI

给一个例子它就自动找出规律爬取网站其余内容。无需选择器、无需代码维护。'直接给我数据'按钮，几行Python。

10. https://t.co/0lRVfrfkSA

curl的增强版，完美模拟真实浏览器指纹，请求看起来就像有Chrome的真人。昂贵爬虫API底层暗用的最低级技巧，现在免费。

公司为此收费2000美元/月。源代码就在这儿。

[Brian Roemmele @BrianRoemmele]

BOOM!

Meet the open source Cambrian Explosion of repulsion of Anthropic!

Meet Qwythos 9B, a Qwen3.5 based GGUF that's both uncensored and quantized for efficiency.

I am running it now and it is brilliant!

A model that can reason through 1 million tokens of context, understand images and text, and even call functions.

Come and take it!

https://t.co/UFU3fas9OD

[Lonely @Lonely__MH]

🚀 Hermes Agent 玩家有福了！

推荐一个刚上线的超实用社区工具：Hermes Bible！

由于 Hermes 迭代速度极快，很多最佳工作流常常被埋没在 X 推文、Discord 或零散帖子中。

为此 @iamlukethedev 专门打造了这个统一的文档与工作流搜索平台。

核心亮点：

1. 海量内容：已收录 169 篇核心文档与 24 个社区优质工作流。

2. 极速体验：支持 ⌘K 瞬时全局搜索，找资料不再费时。

3. 共建生态：支持提交你自己的最佳实践，共同打造 Hermes 社区。

属于 Hermes Agent 玩家的知识库，作者牛逼！🥳🥳

传送门👉🏻：https://t.co/1Pry6fNkeN

[正议 @chkmythought]

@AstraDigitalBTC @Baoliaogeming64 就一傻逼

[火山哥🕊️ @huoshan007]

今天刷到这个，真有点离谱。

有人把一整套 Claude 课程直接免费放出来了，整整 1 小时，讲的不是那种空话，而是怎么用 Claude 构建东西、自动化流程，甚至拿它去赚钱。

说白了。

你少看一集电视剧，把这 1 小时啃完，明天醒来可能就多了一门能变现的手艺。

关键是，他居然不卖课。

免费教你怎么玩，爽吧！

[Alok @analogalok]

I just got Gemma 4 26B A4B MoE model running fully locally with Hermes agent on an 8GB RTX 4060 and it's now backtesting trading strategies end to end, no hand holding.

If you’re a trader or work on Wall Street, you don’t want to miss this.

Yes. fully automated. No cloud. No APIs beyond market data.

# Here's what I did:

Setup:

- Model: Gemma 4 26B-A4B QAT (MoE), Q4_K_XL Unsloth's quant (link in the comments)
- Inference: llama.cpp (turboquant fork by @no_stp_on_snek link in the comments)
- Hardware: RTX 4060, 8GB VRAM + 16GB RAM only (with 50 other chrome tabs open)
- Context: 64K

llama.cpp turboquant flags:
-m gemma-4-26B-A4B-it-qat-UD-Q4_K_XL.gguf -c 64000 --cache-type-k q8_0 --cache-type-v turbo3 --port 8080

turboquant helps achieve high prefill and decode throughput for interactive sessions.

throughput with Hermes agent:
decode: 25+ tokens/sec
prefill: 250+ tokens/sec

# Then I gave the agent one task:

Backtest a strategy:

- Buy when RSI crosses above 30
- Sell at +2% profit or -1% stoploss
- No overlapping positions
- Use Google stock via yfinance
- Generate a full HTML report with candlestick charts + signals

What happened next was wild. It didn't just write code, it ran the entire workflow itself:

Audited the environment (pip list, dependency check)

Hit a ModuleNotFoundError, multiple Python installs were conflicting

Ran where python to map every interpreter on the system

Manually selected the correct Python 3.13 path and re ran the script

Wrote a clean statevmachine backtester (strict no overlapping trades logic)

Patched a yfinance MultiIndex quirk that would've crashed the script

Built Plotly candlestick + RSI charts with buy/sell markers

Calculated win rate, PnL, and summary stats
Exported a polished single file HTML report. check the report at the end of the video or in the comments.

Biggest takeaway: local LLMs aren't just "chat assistants" anymore. They debug their own environment, write production code, and ship a finished deliverable on consumer hardware, for $0 in API costs.

If you're still calling local models "toys," you're already behind.

This is just the beginning.

Hermes agent just surpassed 1 trillion tokens in a single day on OpenRouter. Think about the scale of total token generation happening right now.

Disclaimer: This is not financial advice. Consult a professional before making any trading decisions.

[zhouluobo @zhouluobo]

说一个暴论，如果想理解 AI 接下来怎么走，先把信息源换一批。

相信很多朋友和我一样，AI 关注列表里，有大量工具号、变现号、提示词号，这些当然有用。

但如果你想看更上游的东西，比如模型方向、Agent 演进、开源生态、机器人、推理、AI 安全、产品化趋势，可以优先关注这些人。

按 4 层来分，咱们一层一层看。

第一层：顶级研究者和思想源头

@karpathy
最适合普通人长期关注的 AI 技术解释者。LLM、训练、Agent、教育内容都讲得非常清楚。

@ylecun
图灵奖得主，Meta AI 代表人物。观点很锋利，经常和主流 LLM 路线辩论。

@fchollet
Keras 作者，ARC-AGI 发起人。很适合看他对智能、泛化、评测的思考。

@ilyasut
Ilya Sutskever，SSI 创始人。发得少，但每次都容易成为行业信号。

@geoffreyhinton
深度学习奠基人之一。AI 风险、智能本质、神经网络历史，都绕不开他。

@goodfellow_ian
GAN 发明者。低频高信号，适合关注基础研究和技术判断。

@NoamShazeer
Transformer 论文作者之一，https://t.co/21h2YzU8lz 联合创始人。真正影响过这一代模型架构的人。

第二层：顶级实验室和负责人

@OpenAI
模型、Codex、Agent、开发者生态，一手更新源。

@AnthropicAI
Claude、Computer Use、AI 安全和可控 Agent，非常适合关注 Agent 工作流的人。

@GoogleDeepMind
Gemini、Alpha 系列、科学 AI、机器人和多模态研究。

@AIatMeta
开源模型、Llama、视觉和基础研究，开源生态必须看。

@xai
Grok、X 算法、实时信息流和 AI 产品结合，和 X 内容生态关系很近。

@sama
OpenAI CEO，看产品方向和行业节奏。

@DarioAmodei
Anthropic CEO，看安全、Scaling、模型治理和长期判断。

@demishassabis
DeepMind CEO，AI for Science 这条线非常重要。

第三层：Agent、工程化和真实工作流

@hwchase17
LangChain 创始人。LangGraph、Agent 架构、Evals、长期工作流，非常适合做 Agent 的人看。

@LangChain
Agent 工程化更新源，LangGraph / LangSmith / Deep Agents 都在这里。

@jerryjliu0
LlamaIndex 创始人。RAG、知识库、数据层和 Agent 结合，实战价值很高。

@goodside
Riley Goodside，提示词、模型行为、边界测试，很多案例都能直接启发工作流设计。

@HamelHusain
做 AI 系统测试、Evals、可靠性，非常适合看“怎么让 Agent 真正可用”。

@swyx
AI Engineer 代表人物之一。Agent、AI UX、开发者生态、产品化趋势都很有密度。

@DrJimFan
NVIDIA 机器人和具身智能方向，非常适合看 Physical AI、世界模型和机器人。

第四层：产品、创业和 AI 应用

@AndrewYNg
AI 教育、应用落地、创业和职业建议，适合长期看。

@levelsio
独立开发者代表，AI 产品化和快速上线能力很强。

@steipete
OpenClaw / 桌面 Agent 相关，适合关注本地 Agent、电脑控制和开发者工具。

@rileybrown
Vibe coding 和 AI 产品实操，适合看普通人怎么把 AI 用成产品。

@gregisenberg
AI 创业、创意、商业化，适合看需求和产品机会。

@yoheinakajima
BabyAGI 作者，早期 Agent 探索者，适合看自主 Agent 的实验。

如果要选出 10 个一定要关注的，我会选下面这个列表，

@karpathy
@fchollet
@ylecun
@ilyasut
@OpenAI
@AnthropicAI
@GoogleDeepMind
@hwchase17
@goodside
@DrJimFan

这 10 个基本能覆盖模型、研究、Agent、产品和未来方向。

工具号可以帮你追热点。

这些账号能帮你判断热点背后的趋势。

[How To Prompt @HowToPrompt__]

someone open-sourced a tool that runs an entire Android phone inside a Docker container.

100% open-source.

[Akshay 🚀 @akshay_pachaar]

Web scraping will never be the same.

(100% open-source visual search at scale)

PixelRAG is a retrieval system that skips HTML parsing completely.

Instead of scraping a page into text and embedding chunks, it screenshots the page and retrieves the image. A vision-language model reads the answer straight off the pixels.

Why that matters: parsing is where web RAG quietly loses information.

- A single HTML-to-text parser can drop 40%+ of a page.
- Tables, charts, and layout get flattened or thrown out.
- Swapping parsers alone can move accuracy ~10 points on the same docs.

PixelRAG indexes the page a person actually sees. The team built a visual index of all of Wikipedia, 30M+ screenshots, and it still beats the strongest text RAG baseline by 18.1% on text-only QA.

The repo also ships a Claude Code plugin that gives Claude eyes.

It lets Claude screenshot any URL and read the rendered page instead of scraping the DOM. So you can hand it a live page, an arXiv paper, or your local site and ask what it actually looks like.

One setup script. No MCP server, no backend.

How the pipeline works:

- Renders each document (web, PDF, image) to image tiles.
- Embeds them with Qwen3-VL-Embedding, LoRA fine-tuned on screenshots.
- Builds a FAISS index and serves a search API.

A stronger reader model lifts accuracy with no re-indexing, since the index is just pixels.

Everything is open-source under Apache-2.0.

GitHub repo: https://t.co/qun9TjAdmw

Talking about RAG, I recently wrote an article on a new approach that makes retrieval much more efficient by cutting corpus size by 40x, reducing tokens per query by 3x, and improving vector search relevance by 2.3x.

The article is quoted below.

[Alok @analogalok]

gemma-4-12B-agentic-fable5-composer2.5 V2 is out.

the agentic upgrade to the model trained on Fable 5's reasoning. Running it now with TurboQuant llama.cpp on a single RTX 4060( 8 GB VRAM) at 30 tokens/second with full 25000 context and reasoning:

# The benchmarks

v2 is built for coding + agentic work. writing code, running commands, using tools, debugging, multi step technical tasks. The clearest signal is tau2 bench telecom, an agentic tool use benchmark whose diagnose → fix → verify loop mirrors real terminal/debugging work:

tau2 bench telecom numbers:

base Gemma 4 12B: ~15%
this finetune: ~55%. (Self reported)

thats a huge jump

# TheTom/llama-cpp-turboquant flags:

llama-server.exe -m gemma4-v2-Q4_K_M.gguf -ngl 99 -c 25000 --cache-type-k q8_0 --cache-type-v turbo3 --port 8080

Flag breakdown:

-ngl 99 → full GPU offload
-c 25000 → 25K context
--cache-type-k q8_0 --cache-type-v turbo3 → mixed-precision KV cache — K at 8-bit, V at ~3-bit via TurboQuant (Walsh Hadamard rotated polar quant, Google's own KV-compression research).

Not even merged into mainline llama.cpp. running it off a fork.

No API. No cloud. Just llama.cpp. well, a fork of it and any 6gb+ GPU.

If you tried yuxinlu1/gemma-4-12B-coder-fable5-composer2.5-v1-GGUF, check this out and share your experience with the models

[Ridark @ridark_eth]

I genuinely don't understand why everyone isn't using this yet

Andrej Karpathy, a co-founder of OpenAI, posted a simple idea that hit 16 million views: stop using AI to write code, use it to build a second brain.

You point Claude Code at a folder, drop in any source, an article, a transcript, a PDF, and Claude reads it, links it, and files it into a living wiki of everything you know. It compounds like interest, the more you feed it, the smarter it gets.

Here's the whole thing:

> Install Obsidian, create a vault, open it in Claude Code

> Paste Karpathy's wiki idea file and tell Claude to build it

> Claude makes three folders: raw for sources, wiki for its pages, a CLAUDE.md that runs it

> Drop any source into raw and say "ingest this"

> Ask questions across everything, forever

Five minutes to set up, and you never start from a blank chat again.

Full step-by-step guide with Claude and Obsidian, link below.

Bookmark this

[precis0x @precisox]

Un desarrollador chino llamado tw93 se hartó de que sus aplicaciones de escritorio le devoraran la RAM y el disco.

Abría Slack y desaparecían cientos de megabytes. Abría Discord, Notion o cualquier otra app y pasaba lo mismo. ¿La razón? Casi todas son lo mismo por dentro: un sitio web empaquetado con una copia completa del motor de Chrome (Electron).

Decidió que tenía que haber una forma mejor.

En 2022 empezó a construir Pake. Usó Rust + Tauri, que en vez de incluir un navegador completo, aprovecha el WebView nativo del sistema operativo.

El resultado fue brutal:

- Slack con Pake → 8 MB (en vez de 524 MB)
- Discord con Pake → 9 MB (en vez de 265 MB)
- ChatGPT con Pake → 9 MB (en vez de 260 MB)

Cuatro años después, su repositorio tiene más de 51.000 estrellas en GitHub. Tiene builds listos para Grok, ChatGPT, Gemini, Discord, YouTube, Twitter y muchos más. Todo bajo los 10 MB, ligero, rápido y gratis.

Y lo mejor: con un solo comando puedes convertir cualquier página web en una aplicación de escritorio nativa.

No fundó una startup. No levantó inversión. Solo resolvió un problema que molestaba a millones de personas.

A veces el cambio real lo hace una sola persona que se cansa de las cosas como están.

Esta brutal, repo en los comentarios 👇

[HankAI @hank_aibtc]

卧槽了！
GitHub上居然公开白嫖这种核武器级社工工具！

Storm-Breaker 直接能做到：
- 远程实时偷看对方摄像头画面（手机/电脑都行）
- 实时监听麦克风，听到周围一切声音
- 精准GPS定位，告诉你目标现在在哪（智能手机专用）
- 无需任何权限直接抓取完整设备信息（型号、系统、IP、浏览器指纹全都要）
- 还有OS密码抓取（Win10测试通过）等隐藏功能

想象一下：
发个“看你性格测试”或“紧急视频”的链接，
对方一点你的屏幕上立刻出现：
1. 实时摄像头画面
2. 麦克风监听
3. 精确GPS位置
4. 对方手机/电脑所有设备指纹

操作简单到离谱：运行工具 → 一键生成钓鱼链接 → 发给目标（伪装成“性格测试”“紧急视频”“文件下载”都行）→ 对方一点开，你后台面板立刻全控！

这个项目用 Python + PHP + JS 开发，
支持本地 + 个人域名/主机部署，还自带 Ngrok 一键外网暴露。

安装超简单（Kali示例）：

git clone https：链接
cd Storm-Breaker
sudo bash install,sh
sudo python3 -m pip install -r requirements.txt
sudo python3 st,py

（个人主机部署把 storm-web 内容上传 public_html 就行）

严肃警告：
- 这玩意儿属于灰色/黑色地带工具，仅限学习研究、红队演练和合法授权测试！
- 乱用违法！侵犯隐私、非法入侵，后果自负，号被封、被抓别怪我没提醒。
- 下载前扫毒，社区很多人说类似工具可能带后门，安全第一！

懂的自然懂……这年头社工门槛越来越低了。
你们平时怎么防这种钓鱼链接？评论区交流交流（别真发自己用法啊）

[Rimsha Bhardwaj @heyrimsha]

Google Translate is cooked after this.

A developer built a local AI translation engine that runs 40 languages entirely on your own laptop.

It's called LibreTranslate.

No API key.
No usage limits.
No sending your documents to Google's servers.

You install it once. It runs forever.

Here's what it handles:

→ Paste text. Translated instantly.
→ Drop in a file. Outputs the translated version.
→ Point it at a URL. Returns the page in your language.
→ Build it into your own app via its local REST API.

The speed is not the story. The privacy is.

Google Translate reads every sentence you paste into it. Legal contracts. Medical records. Internal emails. Client documents. Every word goes to their servers and stays there.

LibreTranslate runs entirely offline. Nothing leaves your machine. Ever.

The numbers:

→ 40 languages supported
→ Runs on CPU -- no GPU needed
→ Self-hosted in under 5 minutes
→ REST API built in for developers
→ 10K+ stars on GitHub

100% open source. MIT licensed. Price: $0.

Google charges nothing for Translate either but it charges you something else.

GitHub: https://t.co/5XF7qgbMRB

[nini @nini_incrypto_]

微软最近推出的走实战路线的东东，直接让 7B 模型接管你的鼠标键盘！

FARA，它抛弃了无意义的聊天，专攻纯本地桌面自动化。

它的核心优势就两个词，听话、省钱：

1. 纯桌面执行：自己打开网页、填表单，自动跑完所有重复的机械流程。

2. 极低门槛：体量仅 7B，普通电脑就能在本地带得动，不烧 Token 不怕断网。

https://t.co/0vbDeMOoEU

[Hugging Models @HuggingModels]

Gemma 4 12B Coder is here and it's a game changer for local code generation. This GGUF model packs Google's latest gemma-4 architecture into a compact 12B size, perfect for running on consumer hardware. It's optimized for reasoning and thinking, making it ideal for developers who want fast, private coding assistance without the cloud.

[Vincent | 信号＞噪音 @VincentLogic]

4.66 GB 的模型，居然在本地跑出了麦肯锡咨询顾问的水平？🤯

Unsloth 最新的 2-bit Gemma 4 12B，效果真的炸裂。

这不仅仅是聊天，它直接化身“超级 Agent”自主干活：
✅ 自主联网搜索引用 15+ 信源
✅ 深度区分 "AI-enabled" 和 "AI-native"
✅ 输出完整的市场地图和 Top 5 创业机会

最恐怖的是门槛：只需 >6GB 内存！ 👇👇

[Alok @analogalok]

Run Gemma 4 26B MoE on 8GB VRAM with 250k context at 20+ tokens/sec

If you own any 8GB VRAM graphics card, stop what you are doing. Local AI just had its absolute "Holy Shit" moment for budget hardware.

Yesterday, I benchmarked Unsloth Gemma 4 12B Q4_K_XL on an 8GB card.

The community went wild but immediately demanded more: "Can we run a 25B+ model on budget GPUs?"

Today, I’m delivering exactly that.

I am running a massive 26B parameter Mixture of Experts (MoE) model locally on a standard 8GB VRAM setup with 250k full native context!.

If you own an RTX 3060, 3070, 4060, or any budget GPU with 8GB of VRAM, the local AI paradigm has completely changed.

The performance metrics are astonishing:

- 20 tokens/sec flat decode throughput.
- Stable, flat decode speed even with massive prompts.
- I threw a 60k token prompt at it, and it still clocked in at 20 TPS without dropping a single frame.

# What about prefill?

Yes, Time To First Token (TTFT) is slightly high when swallowing massive contexts. But with a solid 200 tokens/sec prefill speed, the wait is barely noticeable and highly usable.

And this is running completely without Multi Token Prediction (MTP) active.

How is this possible? It’s the magic of Google's new QAT (Quantization Aware Training) quants for Gemma 4.

The model weight file (unsloth gemma-4-26B-A4B-it-qat-UD-Q4_K_XL.gguf) is only 13.2 GB, making it the ultimate local powerhouse.

# The Test Setup:

CPU: Intel Core i7

RAM: 16GB System RAM

GPU: NVIDIA GeForce RTX 4060 Laptop GPU (8GB VRAM)

# The Secret Sauce (The -cmoe Flag)

To make this work properly on any 8GB card, you must use the -cmoe (CPU MoE) flag in llama.cpp.

This flag isolates the heavy MoE expert weights directly to system memory (CPU/RAM) while letting your GPU focus strictly on the Attention layers and the KV Cache.

It prevents VRAM spillage and holds the throughput rock solid.

# The flags:

-m "gemma-4-26B-A4B-it-qat-UD-Q4_K_XL.gguf" -cmoe -c 248000 -v

Once running, just open the UI on localhost and toggle the new reasoning lightbulb icon in the text input box to watch the model perform multi step thinking.

Are you still running smaller models, or are you ready to scale up your budget local setups? Let's discuss in the replies

[Kshitij Mishra | AI & Tech @DAIEvolutionHub]

THIS FEELS ILLEGAL 🤯

NVIDIA is giving access to 120+ AI models FREE for an entire year.

No credit card.
No payment.
Just a free API key.

Hermes Studio already supports NVIDIA out of the box, so setup takes minutes.

→ 120+ models available
→ 40 requests per minute
→ Free access for 1 full year

While most people are spending money on AI APIs, this is hiding in plain sight.

[小互 @xiaohu]

Google 发布 Gemma 4 12B 开源模型

16GB 笔记本跑全模态 AI

Gemma 4 12B 采用了一种叫"Unified"的无编码器架构，让文字、图像、音频、视频四种输入直接进入同一个 Transformer 主干网络处理。

模型可直接处理原始的图像和声音

用一个类比讲清楚

传统多模态模型处理图片和音频的方式，类似于一个只会中文的老板配了两个翻译：一个英文翻译（视觉编码器），一个日文翻译（音频编码器）。

每次有英文或日文材料进来，必须先让翻译转成中文，老板才能看懂。翻译本身占工位（显存），翻译过程要排队等（延迟），而且老板拿到的是翻译加工过的版本，不是原文。

Gemma 4 12B 做的事情是：把两个翻译都裁了，让老板自己学会了直接看英文和日文。

几个关键数字：

16GB 显存或统一内存能跑，4-bit 量化低到 8GB，目标就是在普通笔记本上本地运行

256K Token 上下文窗口，支持 140+ 种语言

内置 Thinking 模式（逐步推理）和原生 Function Calling