Cloudflare open-sourced the AI security auditor it uses to hunt bugs Most AI security tools have one big flaw: the same agent finds the bug and decides it's real. Cloudflare just open-sourced a fix. That's the smartest rule in Cloudflare's new open-source security skill, and it's now one of the fastest-growing repos on GitHub. The problem: Ask an AI agent to "find security bugs" and it will find plenty. Many won't be real. AI agents are confident, and they rarely argue with themselves. So security teams end up wasting hours checking false alarms. Cloudflare's fix is a skill (a set of instructions your coding agent follows) called security-audit. You install it, point your agent at a codebase, and say: "security audit this codebase" Then it runs six phases: 1Reconnaissance: map the code, the trust boundaries and where user input enters 2Hunting: separate agents search for different attack types in parallel, like injection, access control and business logic 3Validation: for every suspected bug, a fresh agent tries to prove it wrong 4Structured output: findings saved as machine-readable JSON, checked against a schema 5Independent verification: another pass checks each finding against the real source code 6Report: a clean write-up for humans Phase 3 is the key. Finding and checking are done by different agents, like having a second developer review your code instead of reviewing it yourself. Why it matters: → It came from real use. It's the early version of the system Cloudflare built to hunt bugs across its own repos. 
→ It works with any coding agent that supports tools and parallel sub-agents, like Claude Code or Codex. 
→ It's MIT licensed, so free to use anywhere. The catch: Six phases with many agents use a lot of tokens. And no AI audit replaces a human security review. It finds candidates faster; it doesn't give you a guarantee. What I find interesting as someone building agent tools: this week we saw attackers exploiting PHP2Shell hours before the patch. Defenders need speed too. A skill like this lets any team run a serious audit before attackers do. Links in the reply 👇 💻 Repo: github.com/cloudflare/securi…

466K views. 10K bookmarks. Built on a quote that was never said. 466K views. 3.8K likes. 10K bookmarks. All built on a quote that, according to a fact-check, was never said. A viral post this week opened with: "Elon Musk posted an idea that hit 27 million views." Then came the quote, then a step-by-step plan to let a Grok bot run a money-making website for you. Grok itself replied under a quote of the post: no Elon post matches the quote or the 27M views. Recycled hype from promo accounts. Look at the numbers again. Bookmarks are almost 3x the likes. People didn't just enjoy it. They saved it as a plan to follow. Thousands of people filed away a playbook whose first line, by that check, isn't real. And the account behind it? Six months old. Four username changes. A paid blue tick. Readers saw a tick and a famous name, and stopped asking questions. This isn't new, and it isn't only bots. → An MIT study of 11 years of Twitter found false news was 70% more likely to be retweeted than the truth. The cause was people, not bots. False news feels new, and new things get shared. 
→ A Columbia and Inria study found 59% of news links shared on Twitter were never clicked at all. 
→ A satire site once posted "70% of Facebook users only read the headline of science stories." The article body was fake filler text. It was still shared 46,000 times. Here's the irony: A big following works like borrowed trust. Readers stop asking "is this true?" and start asking "is this useful to me?" A blue tick and a famous name in the first line make checking feel unnecessary. And the more useful a post looks, the less people check it. A "prompt to make money" gets saved, not questioned. Even the plan itself deserves a second look. One step says the bot negotiates deals "by email, in your name". That means an AI writing to real people as you, without you reading it first. What I take from this, as someone who posts about AI every day: • A famous name in the first line is a reason to check, not a reason to trust 
• A blue tick means someone pays for X, not that their claims are true
• Bookmarks measure hope, not truth 
• If a claim can't link to its source, it's a story, not a fact Note: I have deliberately left out the account name. The problem isn't one poster. It's 10,000 people saving a claim nobody checked Sources are in the reply 👇

Hermes Agent makes 3D models from text. But the real news is bigger than that. An AI agent that makes 3D models from text? Not new. Meshy, a text-to-3D tool, has an MCP server (a standard way to plug tools into AI agents). Any agent that supports MCP, including Claude and Codex, can already connect to it. Agents asking permission before acting? Also not new. Claude Code and Codex do that all the time. So what did Nous Research actually add with the new Meshy plugin for Hermes Agent? The new part is the price check. A normal approval asks: "Allow this tool?" Hermes asks: "This costs 15 credits. Go ahead?" In the demo, someone asks for a boss enemy for their Godot game. Hermes picks the settings (15,000 triangles, 4K textures, a GLB file for Godot), tells them the cost, and only spends after they say yes. That's a small change, but an important one. Agents are starting to spend real money: credits, API calls, cloud resources. Permission alone isn't enough. You need to know what it costs before you click yes. Nous also added polish on top: → One-line install from the official plugin catalog
→ Models render in the chat, and rigged characters animate there
→ A gallery of every model and prompt you've made
→ Short preview videos on messaging apps, so you can create from your phone
→ 3D-print prep: fixing problems, splitting into parts, handing off to a slicer The catch: you need a Meshy Pro plan and API key, and every model costs credits. What I find interesting as someone building agent tools: as agents start spending money for us, every approval should show the price, not just ask permission. Hermes is one of the first to make that the default. Links in the reply 👇

How do you let an AI agent run shell commands without letting it wreck your computer? That's the headache Vercel just solved. Agents get much more useful with a shell (the command line, like bash). They can list files, read CSVs, run scripts and search code, the same way a developer works. But giving an AI real shell access is risky. One bad command, like deleting files or leaking secrets, and you have a problem. So every team built its own setup: 1Write a custom "run command" tool 2Build or rent a sandbox (an isolated space where the agent can't touch anything important) 3Glue the two together Lots of work, and easy to get wrong. Vercel's AI SDK now ships both parts, ready-made: → A bash tool. One line of code adds shell access to your agent.
→ Pluggable sandboxes. You choose how isolated the agent is. The sandbox options: • In-memory: fast and cheap, with a pretend file system that lives only in memory 
• Full virtual machine: heavier, but a real computer that can install software and run programs Example: put a sales.csv in a restricted in - memory sandbox, give the agent the bash tool, and let it analyse the file. It works only inside the sandbox. It can't reach your actual machine. Why it matters: • Safety by default: the agent only sees what you put in the sandbox 
• Less custom code: no shell tool or isolation layer to build yourself 
• Flexible: start cheap in memory, switch to a full VM later without rewriting your agent. Think of it as a sandpit. The agent can dig and build freely, but it can't wander out into the street. A detail most people will miss: the in-memory option isn't a real shell. It's a fake bash called just-bash, with commands like ls and grep rewritten in TypeScript. That's why it's so cheap, and why it can't run real programs. One note: the function is called experimental_createBashTool, so it's still experimental and may change. Links in the reply 👇

Someone trained an AI on candlestick charts instead of text LLMs learned the language of words. Kronos learned the language of markets. Everyone on crypto Twitter talks about "AI trading". This is what a real research model for it looks like, and it's open source. The problem the paper starts from: Big pre-trained models worked so well for text that researchers built the same thing for time series (data that changes over time, like prices or temperatures). These are called Time Series Foundation Models. But on financial candlestick data, they often do worse than smaller, specialised models. Market data is just too noisy. So the team built one model only for that. How it works: A large language model reads text as tokens and predicts the next one. Kronos does the same with K-lines, also called candlesticks: the bars on a price chart showing open, high, low, close and volume. To Kronos, a price chart is just a long sentence. What makes it interesting: → Trained on data from over 45 exchanges around the world
→ Works zero-shot, meaning no extra training on your data
→ Does price forecasting, volatility forecasting, and generates synthetic price data
→ Tiny by AI standards: 4.1M to 102.3M parameters. Runs on a normal laptop.
→ A live demo forecasts BTC/USDT as a range of possible paths, not one magic price One thing most reposts miss: this isn't new. The paper came out in August 2025. It's resurfacing now because of the AI trading hype. The catch: A model that predicts patterns does not make you money. Markets react to news, people and events that no chart contains. This is a research tool, not a trading bot. What I find interesting as a data engineer: the foundation model idea, one big pre-trained model you reuse everywhere, is moving from text to every kind of data that changes over time. Finance is just the noisiest test. Links in the reply 👇

IRONY : Sisyphus pushed a boulder up a hill forever. We push our usage limits to zero, wait for reset day, and call it productivity.

Claude Duet is Anthropic's answer to Grok Bot, OpenAI's Dot and Meta's Muse. Different names, mostly the same idea: an AI that works in the background for you. Everyone is copying everyone now.

More Details: VoxCPM2 and Drama 3 are models. But most people don't want a model. They want an app. That's why VoiceStudio gained 3,000+ stars on GitHub today. It's an open-source, fully local alternative to ElevenLabs. Instead of one model, it bundles many into one picker: → 16 text-to-speech engines, including CosyVoice 3, IndexTTS 2.5, MOSS-TTS and dots.tts
→ 11 speech-to-text engines for dictation and subtitles
→ Voice cloning, voice design, video dubbing and audiobooks
→ Runs on your own machine. No account, no API key, no cloud. And here's the interesting part. I checked the engine list: ✅ VoxCPM2 is in.
❌ Drama 3 is not. Remember the licence difference from my last post? VoxCPM2 is Apache 2.0, free for anyone to build on. That's exactly the kind of model apps like this can pick up and ship. One more catch: VoiceStudio itself is AGPL-3.0. Free to use, but if you change it and offer it as a service, you must share your code. So the open voice stack now looks like this: Models (VoxCPM2) → Apps (VoiceStudio) → You 👇 github.com/debpalash/VoiceSt…

Follow-up post: VoxCPM2 and Drama 3 are models. But most people don't want a model. They want an app. That's why VoiceStudio gained 3,000+ stars on GitHub today. It's an open-source, fully local alternative to ElevenLabs. Instead of one model, it bundles many into one picker: → 16 text-to-speech engines, including CosyVoice 3, IndexTTS 2.5, MOSS-TTS and dots.tts
→ 11 speech-to-text engines for dictation and subtitles
→ Voice cloning, voice design, video dubbing and audiobooks
→ Runs on your own machine. No account, no API key, no cloud. And here's the interesting part. I checked the engine list: ✅ VoxCPM2 is in.
❌ Drama 3 is not. Remember the licence difference from my last post? VoxCPM2 is Apache 2.0, free for anyone to build on. That's exactly the kind of model apps like this can pick up and ship. One more catch: VoiceStudio itself is AGPL-3.0. Free to use, but if you change it and offer it as a service, you must share your code. So the open voice stack now looks like this: Models (VoxCPM2) → Apps (VoiceStudio) → You

We just saw this happen with PHP2Shell. The bug leaked on Monday, attacks began Tuesday morning, and the patch landed that afternoon. The patch window isn't shrinking. It's gone.

@a16z says the patch window is collapsing. Its own data source has doubts. 87% of exploited bugs are attacked on or before the day they go public. But read the fine print. a16z's "Charts of the Week" are going viral. The story: 1In 2020, about 23% of exploited bugs were hit on day zero. Now it's about 87%. 2In 2022, half of exploitable bugs were still unused after 1.5 months. Now almost none are. 3The median time from disclosure to attack has dropped from a year to about a day. It's projected to hit 1 minute by 2027. All three charts come from one source: the Zero Day Clock, built by Sergej Epp of Sysdig. Here's the catch most people miss. The Zero Day Clock has since retired this exact metric. Its own FAQ says it can mislead: → Recent bugs have had less time to collect evidence, so recent years look faster than they really are.
→ The number can't go below zero, so once most attacks land on day one, it can't show more speed.
→ Same-day attacks get rounded to zero. And the "1 minute" box isn't measured. It's a trend line drawn forward from 2018 to 2024 data. So is it all hype? No. We saw the real version this week. PHP2Shell details leaked on 20 October. Cloudflare saw attacks the next morning. The patch arrived that afternoon. Attackers were hours ahead of the fix. What I take from this as a practitioner: • The direction is real. Attacks are arriving faster.
• The exact numbers are softer than the charts suggest.
• Either way, monthly patching is too slow. You need automatic patching, day-zero firewall rules, and fast exposure checks.

They want to run a whole AI training run as ONE GPU program 1B raised. No product. No benchmarks. No code. AI lab Flapping Airplanes just closed a $1B Series A at a $6B valuation. Index Ventures and Lux Capital co-led. Nine months ago they were valued at $1.5B. That is 4x in under a year. So what are investors paying for? Here is the problem. When you train an AI model, the GPU does not run one big program. It runs thousands of small programs called kernels (small pieces of code that run on the GPU), one after another. Every switch between kernels costs time. Data moves in and out of memory. The chip sits idle for tiny moments. Across a huge training run, that waste adds up. Their idea: a megakernel. Fuse the whole training run into one single GPU kernel. No switching. The chip stays busy. They claim to have the most complete megakernel framework in the world. The team is serious: → Ben Spector, ex-Stanford, worked on ThunderKittens (a well-known library for writing fast GPU kernels) 
→ Asher Spector 
→ Aidan Smith, ex-Neuralink The thesis is simple: AI demand will grow faster than chip supply. If you can't get more GPUs, the winner is whoever wastes the fewest. But there is a catch. Nothing is public yet. No product, no benchmarks, no open-source release. This is a pure research bet. qz.com/flapping-airplanes-ai…

Follow-up: Not sure if your PHP site is affected? OpenAI just released a Codex skill that checks it for you. Point Codex at the repo and follow along. I think this is how security fixes will ship from now on. 💻 Repo: github.com/openai/php2shell-…

Review code 100X faster with pr-lens. AI agents now write a big part of our code. But someone still has to review it. Reading a 40-file pull request (PR, a request to merge code changes) line by line is slow and tiring. There is a faster way: look at a picture first. A diagram shows which parts changed and how data moves between them, before you read any code. But there is a catch. Diagrams go out of date fast. Nobody draws one for every PR, so reviewers go straight to the diff (the list of changed lines). The team at Coldtea built an open-source tool for this: PR Lens. The idea is simple: every PR gets its own diagram, automatically. 1You or your agent opens a PR. 2PR Lens reads the diff with an AI model. 3It draws the change as animated architecture and data-flow diagrams. 4It posts them as a comment inside the PR, and updates them on every push. You see the shape of the change before you read the code. How it works: → The AI first writes the change as a JSON graph (a simple list of parts and links). A fixed renderer then turns it into SVG images, so the drawing is the same every time.
→ The animation is plain SVG with no JavaScript, so it still moves inside a GitHub comment.
→ Wrong names? Fix them once in a .github/pr-lens.yml file. The fixes stay when the diagram is redrawn.
→ It runs 4 ways: GitHub App, GitHub Action, CLI, or a skill for coding agents like Claude Code, Cursor and Codex. What you get: • Free for open source, MIT licence
• Bring your own model: Gemini, OpenAI, or compatible
• No API key needed with the hosted GitHub App
• The agent skill uses your agent's own model, so no extra bill What I find interesting as someone building agentic workflows: the agent that writes the code can now explain it too. A diagram won't replace tests or careful review, but it tells you where to look first. Links in the reply 👇

The most downloaded(860k downloads) model you've never heard of 860,000 downloads. 0 likes. 50% word error rate .. meaning this model gets half the words wrong. A news site just wrote about a Slovak speech model that "quietly" hit 860k monthly downloads on Hugging Face. So I looked at the model page. It's a small community fine-tune of Meta's XLS-R (a speech model trained on many languages), adapted for Slovak. Here's what I found: → Word error rate is about 0.50. Roughly half the words come out wrong. 
→ The model card title says "cs" (Czech), but the repo is "sk" (Slovak). Probably copied and never updated. 
→ It has 0 likes. That doesn't look like a model 860,000 people chose. My guess at what's really happening: 1An old tutorial or test suite uses this model as an example. 2Automated pipelines (CI jobs, scripts, containers) download it every time they run. 3Each run counts as a download. 4Nobody ever looks at the output. So the number is real. It's just not people. This matters if you pick models by download count. A high number can mean "everyone trusts this". It can also mean "some script somewhere runs every 5 minutes". What I find interesting as someone building data pipelines: we judge models by downloads, but downloads measure traffic, not quality. Check the eval results, the likes and the model card before you trust the number. If anyone knows where these downloads actually come from, please reply. I'd love to know if I'm right or wrong. Links in the reply 👇

Six open AI families, Nemotron 3.5 and nearly $10 billion. Here's what seems to be driving NVIDIA's move into Europe. NVIDIA isn't scared of Chinese models. It's scared of Chinese chips. This week at GTC Berlin, NVIDIA released a big set of free open models. Language, robotics, biology, even quantum. All pitched at Europe's "sovereign AI" (AI a country controls and runs on its own servers). Why would the world's biggest chip company give AI away for free? The obvious answer: to compete with China's open models. Chinese labs like DeepSeek, Qwen and Kimi now lead open AI. NVIDIA's best open model still scores well behind China's best. But that's only half the story. The bigger worry is hardware. For years, NVIDIA's real moat wasn't the chip. It was CUDA, the software developers use to run AI on NVIDIA GPUs. Everyone learned it, so everyone stayed. Now that moat is under attack: → DeepSeek and Huawei just open-sourced tools that make Huawei's Ascend chips easier to program, aimed straight at CUDA. 
→ DeepSeek plans to put 160,000+ Huawei chips in a new data centre.
→ Parts of DeepSeek V4 were trained on Huawei chips. If the best open models run well on Huawei chips, the world doesn't need NVIDIA as much. So NVIDIA is doing two things at once: 1Give away good models that run best on NVIDIA hardware. 2Lock in Europe before anyone else does. At Berlin, NVIDIA talked about nearly $10 billion committed across Europe. Mistral trained its latest model on NVIDIA GPUs in its own European data centres. Nebius, Nscale, Deutsche Telekom, SAP, telecoms in five countries. All building on NVIDIA. My take: this reminds me of Tesla. Tesla led electric cars, then Chinese carmakers came in cheaper and fast. NVIDIA has watched that happen. Huawei chips are still far behind today. But NVIDIA isn't waiting until they catch up. What I find interesting as someone working with AI: the free models are not the product. They are more like a bait. The real product is the GPU and CUDA underneath. Europe gets "sovereign AI", but it still runs on NVIDIA. Links in the reply 👇

Humanoid robots in 2025: learning to fold laundry. Humanoid robots in 2026: kicking cars and jumping off buildings. Still won't fold my laundry though

The patient isn't real. The surgery isn't recorded. An AI is dreaming it, live. A surgeon picks up the controls. On the screen, the inside of a patient's abdomen. The gallbladder, the tissue around it, the soft shine of light on wet surfaces. The surgeon moves the grasper to the left. It closes around the gallbladder and lifts. The tissue stretches. The reflections shift. Everything responds, frame by frame. But there is no patient. There is no recording either. Every frame is being generated live by an AI, in response to each movement of the controls. This is Cosmos3-H-Dreams, a new open model from NVIDIA. Here is the old way of building a surgical simulator: engineers try to program everything. Every organ, how tissue stretches, how light reflects, how smoke blocks the view. Inside the human body, that's extremely hard to get right. NVIDIA took a different path. They didn't program the physics. They let the AI learn it by watching. They trained it on real gallbladder operations, recorded with the CMR Surgical Versius robot. Each video was paired with the robot's exact movements. The AI learned what happens on screen when the tools move a certain way. Now you move the tools, and it "dreams" what comes next. A few details: → It runs in real time on a single NVIDIA RTX PRO 6000 GPU. → It understands 28 control signals: both arms, the grippers, energy controls, joystick buttons. → It handles movements it never saw in training. Same starting image, different moves, different realistic outcome. → It's open on Hugging Face, with a guide to train it on your own data. NVIDIA is careful to say what it's not. It's a research tool, not a medical product. And looking real is not the same as behaving like real tissue. What I find interesting as someone working with AI: self-driving cars learned to drive in simulated roads before real ones. Surgical robots may follow the same path. One day, a robot might perform its thousandth gallbladder removal before it ever meets a real patient. Every one of the first 999 will have been a dream. Links in the reply 👇 🤗 Model: huggingface.co/nvidia/Cosmos…

An open-source Grok Bot /OpenAI Dot/ Meta Muse alternative showed up on GitHub a day before 💻 Code: github.com/milind-soni/OpenM…
🌐 Website: openmausbot.com/