W&B Inference is now on @OpenRouterAI! You can use @AIatMeta's Llama 3.3 70B Instruct today, with more models rolling out in the coming weeks. https://t.co/rqT6MCTsn1

Built a game in <60 seconds? ✅ Created a repo with one sentence? ✅ This is what happens when you combine GPT-5 in GitHub Copilot with the new GitHub MCP Server. 👀 Watch the full live demo and learn how to use these tools for your project. 👇 https://t.co/P2Xk4lS64c

Plan and track your work right next to your code. 🚀 Start using GitHub Projects to: ➡️ Create custom boards and explore project insights. ➡️ Automate status updates. ➡️ Manage your issues and PRs all in one place. Ditch the context switching. Learn how. 💡 https://t.co/o1C4SC3PIQ

Britannia hotel seacroft leeds, tonight. https://t.co/YqrX0eTxgD

Grok Code is taking over the dev space on the Cline! It’s only been 3 days and it’s already a clear standout choice for many It's not just faster but much more capable with very promising results ✨ “Feels 10x better and faster than Claude” 💸 Cost savings are huge: - Up to 11x cheaper than Sonnet 4 or Gemini 2.5 Pro - Delivers 85–91% savings per session, or 93–97% with caching, especially for high-frequency tasks like prototyping and debugging >level with Sonnet-4 in diff edits, and improving Grok Code costs lowest with highest performance yet✅ Plus, xAI is extending the free trial for another week, don’t miss out! (Ends Sept 10)

America spends trillions on NATO to protect democracies including Germany. But it is has just prevented an opposition candidate from running for mayor for frivolous reasons including his praise for "Lord of the Rings." This is a clear violation of NATO's Charter. https://t.co/KdTB4S1GX6

The VS Code Insiders Podcast now has video support on Spotify! New episodes out every Monday - makes sure to subscribe: https://t.co/rjJQwVl3s3 https://t.co/O7ajF4nvnE

The cleanest butcher in Texas tested every step of their supply chain to see where plastics are entering American beef @eatRadius Whole Foods grass fed ribeye tested off-the-charts for toxic phthalates earlier this year, but no one could figure out where it was coming from... @natfriedman @plasticlistorg Initial theories included: - Plastic wrap on hay bails making it into cow feed - Plastic cutting boards - Butcher paper These all turned out to be false. Radius' levels of the main offending phthalate DEHP were 10x less than the Whole Foods average thresholds, and their in-house butcher process did not increase phthalate levels. But why is their plastic in beef at all? Radius' leading theory is the acidic washing and shrink-wrapping process carried out by many large US meat processors. Meat is sprayed down with an acidic disinfectant to eliminate surface pathogens, then shrink wrapped. The acidity breaks down the plastic, leeching into meat. Radius meat comes wrapped in loose plastic, leading to far less phthalate exposure. Debugging our supply chains takes serious work, but the health effects of even tiny amounts of these chemicals are real. Shout out the @eatRadius team for this study, you can read the full study at eat radius dot com slash blog 🫡

Fun to look at property taxes in absolute dollar terms (left) and as a percentage of home prices (right) https://t.co/sJR2eHb0YU

meta's new chief scientist threatened to quit and return to openai before being given the role, while zuck hunted for other candidates https://t.co/IB8r6eyNVw

you can use DINOv3 for keypoint matching 😏 I got asked a lot so I built an app where you can try multiple DINOv3 models for matching features between images 🦖 for instance here's DINOv3 matching street view vs map ⤵️ https://t.co/bgBBhGV7uL

next year, medicare introduces ai to transform coverage decision-making for innovative entrepreneurs Source: https://t.co/mROCDfsS0X

hacker unearths crucial data tesla lacked in deadly crash, sparking ai business innovation conversations. Source: https://t.co/Yil0evpr22

Bullish Social Heat Signals for: $SGI 97.8 $GOOGL 96.5 $MSFT 95.2 $SHOP 94.8 $DBX 94.6 $NCLH 94.3 Bearish Signals: $GSAT 1.7 $CPB 2.0 $BIGGQ 2.6 $PRSU 3.3 $CHDN 3.7 I really enjoy social heat investment analytics by @AndySwan @landonswan and @mebrantley2 https://t.co/f4IWqdQ2ya

🎓 Interested in a #PhD in machine learning or #AI? The ELLIS PhD Program connects top students with leading researchers across Europe. The application portal opens on Oct 1st. Curious? Join our info session on the same day. Get all the info 👉 https://t.co/0Tq58uexHk #ELLISPhD

Steve Jobs on the most important job of a CEO: “The greatest people are self-managing. They don’t need to be managed. Once they know what to do, they’ll go figure out how to do it…" https://t.co/QFGZzc6Y7K

AI now outperforms doctors in diagnosis. AI was right 80% of the time. Doctors, 30%. The study tests OpenAI’s o1-preview model on clinical reasoning, not multiple choice medical exams. Uses five complex tasks: differential diagnosis, triage, reasoning, probabilistic estimation, and management decisions.

RAG is not dead! However, we are in an interesting phase of exploring unique ways to index and retrieve information. This vectorless RAG framework uses a tree structure index in place of vectors. Reasoning models will enable methods that mimic human-like search. Early days! https://t.co/pQDgTWmsQO

Auto-generate extraction schemas without manual setup using LlamaExtract 🔄 Automatically create schemas for data extraction, making it easier to pull structured information from unstructured documents without needing to define extraction patterns upfront. Simply start off with a prompt and/or example files and: 🤖 Automatically infer the structure of your data for extraction 📊 Skip manual schema definition and let AI handle the heavy lifting 🎯 Focus on using extracted data rather than configuring extraction rules Check out the auto schema generation guide: https://t.co/WeTN9Z3JbK

Google Street View fiction! https://t.co/7CcsRft3f8

In a true accomplishment of context engineering, the team managed to increase the prompt caching percentage up to 92%! https://t.co/FssDV6VnKO

Grok Auto (unified mode) has insanely Improved a lot since the release Now works absolutely the best for all queries.... It picks out the best model according to the complexity of the query The unified mode ensures seamless performance, no matter the complexity... You don't need to pick one, It does it for you

"A government against its people." Elon Musk has summed up Starmer's government in one phrase. Starmer is not for the British. https://t.co/fJUFcqKw2R

groundbreaking 1-step method by us-china collaboration converts plastic to fuel with 95% efficiency, paving the way for innovative business solutions in ai. Source: https://t.co/kh734Zt2YB

Step 2: Move onto your design editor. Create your design (title, dates, logos). This ensures crisp and clean typography. You also have more control of the final composition. > Voilà! ✅ Success *I used ✨@canva✨ https://t.co/QnxgTtfN8I

Original hand-crafted design. https://t.co/kGi1uAkiNN

Results without using an Editor (see typography quality and general loss of detail). https://t.co/WOzC8tblLw

Day #4 of Image Editing with nano banana: Avatar edition Today focus “quality” 1) Separate design from image editing 2) Start with the highest quality image > Let’s go: 🧵 Step 1: Iterate until you are satisfied with the results. 2.5 Flash > Image https://t.co/fdne0Z40we https://t.co/EO6kTLXaiD

The next level of drone shows: Entertaining Can express a lot of topics and scenes 11000+ drones to break the world record @mvollmer1 @morgfair @ChuckDBrooks @Nicochan33 @enricomolinari @NancySinatra @Ronald_vanLoon @alvinfoo @KirkDBorne @Hana_ElSayyed @JimHarris @MikeQuindazzi @Shi4Tech @mhcommunicate @ipfconline1 @kashthefuturist @rwang0 @HeinzVHoenen @YuHelenYu @BetaMoroney @antgrasso @kuriharan @PawlowskiMario @EvanKirstel @HaroldSinnott @terence_mills @FrRonconi @TamaraMcCleary @UrsBolt @pascal_bornet @HeinzVHoenen @SpirosMargaris @richardturrin @Xbond49 @psb_dc @rshevlin @JimMarous @IanLJones98 @Khulood_Almani @enilev @GlenGilmore @DeepLearn007 @KamLardi @debashis_dutta @sallyeaves @EstelaMandela @NevilleGaunt @IngridVasiliu @Eli_Krumova @baski_LA

SmartV 3D Modeller: Interactive Holographic Visualization by @HYPERVSN #AugmentedReality #VR #VirtualReality cc: @spirosmargaris @ianljones98 @pierrepinna https://t.co/i9c8uzMXdy

We recently had the opportunity to discuss the current state of AI with Jensen Huang from Nvidia. This presentation summarizes our discussion. Jensen has talked about building the "AI factory" of the future. What took 144 years during the Industrial Revolution might happen in just a few years with AI. Here's why: rapid specialization. Beyond a certain scale, complex systems need both general-purpose and highly specialized components to mature. Today, we're living through the "Cognitive Revolution": → 1999: First GPU (our "steam engine") → 2016: First AI factory putting all components together → 20XX: The cognitive assembly line The question isn't if this transformation will happen. It's who will be build the Standard Oil, Carnegie Steel and Ford of the AI era. At @sequoia, we believe it's the startups being built right now—the ones carrying out this specialization imperative across every industry. The cognitive revolution is here. And it's moving faster than anything we've seen before. Watch the full presentation on YouTube: https://t.co/faHjhJHARM

Charlie Munger once said: “I love everything about Costco. I’m never going to sell a share.” Let’s look at why Costco was one of Munger’s greatest loves, and what investors can learn https://t.co/odnEvNA9RR