Riviera Beach gunman: "I shot every one of those people" Lawyer: "Shh… this is being recorded" Gunman: "I shot 'em in they head" https://t.co/KTPGoi5Oj7

LATEST: 🚨 OpenClaw developers are being targeted by a phishing campaign on GitHub, offering fake tokens to trick them into connecting their wallets, according to OX Security. https://t.co/yBkfmyWyzQ

Programming Massively Parallel Processors is the gold standard GPU programming textbook and the 5th edition just dropped. Mojo 🔥 solutions to every exercise are in today's nightly release. Check out the repo and see what GPU programming looks like in Mojo ⬇️ https://t.co/Nsnui1LcQU

Waiting minutes for AI video generation is over. Runway's new model on NVIDIA Vera Rubin hits under 100ms to first frame. Minds are blown. https://t.co/PEz85QTE5l

✏️ Introducing Vectorize on Lovart → One-click PNG/JPG to SVG conversion → Auto-traced paths and contours → Download and drop into Illustrator or Figma From AI output to scalable design asset. Like + repost + follow + reply – 10 lucky winners get a free month of Pro

You as agent: super cool Gemini demo by @FilipHavlena👤🍎🖥️ We've been experimenting with agentic interfaces at Creative Lab and this one was so fun: he turned himself into a 3d agent with spatial awareness + reasoning loop + tools (docs/emails with @GoogleWorkspace CLI). Made in @threejs. Real-time screencap ▶️

New safety data from @Waymo today: Compared to human drivers where Waymo operates, 170M miles of fully autonomous driving resulted in: - 92% fewer serious injury or worse crashes - 92% fewer injury crashes involving pedestrians - 85% fewer injury crashes involving cyclists - 83% fewer crashes resulting in any airbags deploying - 82% fewer crashes involving any injury at all At current scale, the data suggests Waymo is preventing 1 serious injury crash *every 8 days.*

Nvidia CEO Jensen Huang in new interview on orbital datacenters: "The challenge of course is that cooling, you can't take advantage of conduction and convection, so you can only use radiation, and radiation requires very large surfaces, but that's not an impossible things to solve. There's a lot of space in space. We're going to go explore it. We're already radiation hardened. We have Cuda in satellites around the world. In the meantime, we're going to explore what is the architecture of datacenters look like in space. It'll take years, but that's ok. I got time." via @theallinpod

Edit: trying it: the poor AI has to use Playwright! I guess auth would be hard, but I was imagining some rich API for editing and running jupyter notebooks on cloud runtimes 😂 Well, still nice to have some GPU access for anyone with a google account... https://t.co/MgTYn0gWTu

When one brain isn't enough, switch to Grok 4.20. Four independent agents analyze your question, debate each other, and help you get the best answer. Available now to SuperGrok and Premium+ subscribers globally. https://t.co/fW7637fMZ6

I think Google's new Stitch tool is a really great example of bringing "vibework" to an area outside of coding with an interface built around design & prototyping. There are rough edges, but (a) the results are very impressive and (b) it will feel more natural for many non-coders https://t.co/kN5gFKgOLD

It also shows how general purpose these models are right now, capable of doing many things with different harnesses, so I think we will see similar applications from other labs soon, as more knowledge work becomes the focus of AI. Anyhow, its free for now: https://t.co/BV3hkZWc9S

The CIA doesn't want you to find this GitHub repo 👀 It's called Shadowbroker and it aggregates every open-source military signal on Earth into one dashboard. → US Navy carrier strike groups tracked live → Spy satellites color-coded by mission type → GPS jamming zones with severity overlays → 25,000+ ships. 2,000+ CCTV feeds globally Right-click any point on Earth. Get a full intelligence dossier. 100% Opensource. Link in comments.

https://t.co/VNZ7pMGLou

https://t.co/VNZ7pMGLou

OpenMed is the #1 most referenced organization powering open-source AI research on Hugging Face. This is the official State of Open Source report, Spring 2026. I started this 8 months ago.

Building AI apps in TypeScript just got easier. ⚡️ Tomorrow on Open Source Friday, learn all about @mastra, a TypeScript-first framework for building AI applications, directly from CTO @AbhiAiyer. Set a reminder and join the stream. 🔔👇 https://t.co/jZGtRr6Hjk

The state of robotics. With the Head of Robotics and Edge Computing Ecosystem @ Nvidia. Amit Goel. Just one of dozens of interviews I've done this week for https://t.co/2BjIVhqvR4 who is having me cover the change that AI is bringing to all businesses. Thanks to NVIDIA for the studio and access.

take a shot everytime I say Nexus... knowing when and why your agent is silently failing or your users are frustrated is the make or break of keeping users happy that's why we Nexus proactively monitors and automates all that for you for your agents in prod https://t.co/cYIMBL1Cq2

https://t.co/olor3fE2on

https://t.co/olor3fE2on

We’re rolling out summaries for Articles now. Just tap the Summarize button if you want to know if it’s worth your time to read it (or if your attention span is 12 seconds). https://t.co/n40VnoBDeR

"Actually beating cancer is going to require systemic changes, both technical and economic, in how we develop and test drugs, as Javorsky convincingly argues, but also deeper forms of artificial intelligence that are better able to reason about chemistry, physics and biology." @GaryMarcus covered @Emilia_Javorsky's new "AI vs. Cancer" essay - check out the essay, and Gary's newsletter, linked below:

Video editing just got more dynamic! 🚀 Thrilled to share DynaEdit: a training-free, text-based method for non-rigid video editing. Work done during my internship at @GoogleDeepMind with @Roni_Paiss, @kusichan, @inbar_mosseri, @talidekel, @t_michaeli https://t.co/l5uKGwL7cT https://t.co/u6OoPBIHu3

🚀Excited to share our work, Prompt Relay , an inference-time method for temporal control in multi-event video generation ! 🎉 Our method has been integrated into Wan’s official repo: https://t.co/ZWTjUBWFGE 🌐 Webpage: https://t.co/bacWSZuJbJ ⌨️ Code: https://t.co/am2MzrN8HU https://t.co/gbqOb7bnJC

+1 more

Introducing M²RNN: Non-Linear RNNs with Matrix-Valued States for Scalable Language Modeling We bring back non-linear recurrence to language modeling and show it's been held back by small state sizes, not by non-linearity itself. 📄 Paper: https://t.co/AS8e2tNrRa 💻 Code: https://t.co/LMvBcI22Du 🤗 Models: https://t.co/NCmjrpNriq

(1/5) 🚀 Excited to share our CVPR paper MATCH! Website: https://t.co/pVkFpdk9Og Given calibrated multi-view images, MATCH infers static Gaussian splat textures in correspondence across subjects and expressions. This speeds up avatar optimization by a factor of 10x.🧵👇 https://t.co/1yoBrBRNA9

The future of home cleaning just landed in Shenzhen and it is walking right into your living room. 🤖🏠 @XSquareRobot and https://t.co/gWVMLX0Lfa officially launched China’s first robot home service, moving embodied AI from the lab to your front door. When you book a cleaning on the https://t.co/gWVMLX0Lfa app, a professional cleaner now shows up with an X Square robot partner to tag team the house. The human handles the tricky stuff that needs real judgment while the robot takes over repetitive tasks like wiping tables and tidying up surfaces. X Square is using an end to end foundation model which means the robot actually perceives and plans its own moves instead of just following a script. By testing in the messy reality of a real home, they are proving that if a robot can master a living room, it can handle almost any physical space. This pilot is part of a massive push to turn these machines into reliable partners that can actually assist in our daily lives.

vibe coding in AI Studio just got a major upgrade 🚀 • multiplayer: build real-time games & tools • real services: connect live data • persistent builds: close the tab, it keeps working • pro UI: shadcn, Framer Motion & npm support we can't wait to see what you build! https://t.co/4X524pWIlM

@_igorshilov Have you seen our paper https://t.co/2HPagxsOhY

Kilopower, the little Martian reactor that could: A miracle of simplicity & sodium. NASA was given the brief to design a super-simple micro-reactor for future Martian settlement, and it did so brilliantly by doing it unconventionally. A normal Earthbound nuclear reactor is complex, with banks of pumps, water & steam injection pressurizers, control rod actuation, frequent movement of fuel during refuelling cycles, complex systems of valves, back-up emergency coolant and deluge systems, boron injection… it goes on & on & on. A terrestrial nuclear plant is a triumph of safety-conscious system engineering, but it’s a busy one, and Kilopower dispensed with all of this. The intention for the system is that you could set it up, get it started and just… forget about it, pretty much. It has a small, solid core with no moving pieces except for the control rod which is designed to be moved on startup and then kept in place, unmonitored, for a decade at a time. The small size & low power means that reactivity effects are dominated by thermal expansion and are extremely stable. Low power designs (less than 100 kiloWatts thermal energy) make thermal management and irradiation damage a non-issue that does not complicate system design. Very low power level designs (less than 10kW thermal) feature such low burnup reactivity that movement of the control rod becomes a ten-year interval. A literal ‘tin’ wedding anniversary, marked out by tweaks to the control rod. There’s a quirky Martian anniversary gift for you! Even at much higher power levels, the stability is such, and burnup so low, that control rod movement would be something you’d only have to do occasionally, every few months or annually. In essence this is a nuclear battery. In all cases, the reactor is designed to handle worst-case transients such as coolant loss passively without any need for control inputs, further simplifying the system. Truly, this is a zen ideal: Feng Shui engineering, where less is more and the system simplicity brings reliability and safety, all in one. Perfect for missions to space.

Reasoning models are growing fast, and running them efficiently requires distributing workloads across multiple GPU nodes. NVIDIA Dynamo 1.0 delivers low-latency, high-throughput distributed inference for production AI deployments—while boosting NVIDIA Blackwell inference performance by up to 7x, lowering token cost, and expanding opportunities with free #OSS. Built for production. Key features: 🔹 Disaggregated serving 🔹 Agentic-aware routing 🔹 Multimodal inference 🔹 Topology-aware Kubernetes scaling 🔹 SLA-ready quick deployments Available now with native support for SGLang (@lmsysorg), TensorRT-LLM, and @vllm_project 👉 https://t.co/c5s9zWx6o3