We partnered with Mozilla to test Claude's ability to find security vulnerabilities in Firefox. Opus 4.6 found 22 vulnerabilities in just two weeks. Of these, 14 were high-severity, representing a fifth of all high-severity bugs Mozilla remediated in 2025. https://t.co/It1uq5ATn9

Frontier models are now world-class vulnerability researchers, but they’re currently better at finding vulnerabilities than exploiting them. This is unlikely to last. We urge developers to redouble their efforts to make software more secure. Read more: https://t.co/LRbhsb6XUb

New on the Anthropic Engineering Blog: In evaluating Claude Opus 4.6 on BrowseComp, we found cases where the model recognized the test, then found and decrypted answers to it—raising questions about eval integrity in web-enabled environments. Read more: https://t.co/oVCNyaiK5w

Soon. https://t.co/gwSviU4TsE

Soon. https://t.co/gwSviU4TsE

We heard your feedback loud and clear, and 5.3 Instant reduces the cringe. https://t.co/WqO0XzLcVu

GPT-5.3 Instant also has fewer unnecessary refusals and preachy disclaimers. https://t.co/wX6wh5Aynp

GPT-5.3 Instant gives you more accurate answers. When using web search, you also get: - Sharper contextualization - Better understanding of question subtext - More consistent response tone within the chat https://t.co/3rwJb8tJe2

The Codex app is now on Windows. Get the full Codex app experience on Windows with a native agent sandbox and support for Windows developer environments in PowerShell. https://t.co/Vw0pezFctG https://t.co/gclqeLnFjr

🧵1/ We've brought the most advanced AI to Prism by introducing Codex to Prism. Prism is already the best place for scientific writing to happen—and with Codex, now you can write, compute, analyze, and iterate all in one place. https://t.co/cDCLYCj7tR

Construction is underway at our site in Port Washington, Wisconsin. This is an important step in our long-term compute strategy. Grateful to our partners @VantageDC and @Oracle who are helping bring this capacity online. We’re been making rapid progress expanding the infrastructure behind OpenAI. Last week we announced an expansion of our foundational work with @NVIDIA systems that will power the next generation of models. We also shared how we're continuing to deepen our work across the ecosystem, including lean in further with @awscloud. Building world-class AI requires world-class infrastructure. Excited to see the work happening on the ground in Wisconsin and grateful to the teams and partners helping make it happen.

🔥 We're cooking in 60 minutes. Join @JamesMontemagno @BurkeHolland @PierceBoggan as they live code using the latest in VS Code, GitHub Copilot CLI plus some cool new stuff Burke's been building 🔮 https://t.co/GRvd0T0uF2 https://t.co/kVF1LegJst

🤖 GitHub Copilot Dev Days is coming! From March 15 to May 15, developer communities worldwide will host free, hands-on events exploring GitHub Copilot with @code, the CLI, .NET, Java, Python, JavaScript, and more. 👉 Find an event near you: https://t.co/O7dceTTCqe https://t.co/wyh0EA3xRv

Want to learn more about features in the new VS Code release? Our latest blog post walks you through some of the top highlights, centered on making agents practical for real-world development. 🔗 https://t.co/kgKDPZkHGh https://t.co/Uq5QR3M5eA

Design → code → canvas → feedback → repeat. The @figma MCP server is now bidirectional. @GitHub Copilot users can pull design context into code and push working UI back to the Figma canvas, all from @code . No handoffs or context switching. Just flow. https://t.co/FbDcp7kboG

For historical participants, technological renaissance felt like existential demise b/c they couldn't reason by analogy to the previously unimaginable https://t.co/UC5t0gXaWH

"Un padre y una madre unidos en matrimonio, tomados de la mano, paseando con sus hijos en brazos, van a ser el gesto más revolucionario e intrépido en este decadente siglo" G. K. Chesterton https://t.co/yBAHT9JCUz

Inspired by @ring_hyacinth & @simonxxoo, I overhauled their "Lobster House" into a "🦞 Lobster Library"! I was drowning in scattered generated code, images, and docs, plus its memories and logs. So I built this unified hub to track what Lobster is reading in real-time, browse recent gens, and index local files. It's still a rough WIP but if y'all are interested, I'll polish it up and open-source it! #openclaw

By the power of @ComfyUI and @Alibaba_Wan ⚔️🐅 https://t.co/l9sEjzRKx9

By the power of @ComfyUI and @Alibaba_Wan ⚔️🐅 https://t.co/l9sEjzRKx9

A funny little clip with Seedance 2 tonight! It reminds me of some of my attempts at changing my hairstyle at the hairdresser's, haha. I got the idea from noticing the resemblance between some dogs and their owners on the street; it's full of surprises, haha. AI tools: Character Design: Midjourney + (Nano Banana on @krea_ai) Animation: Seedance 2 on @dreamina_ai

🚨BREAKING: Google just dropped another hit! It's called Always-On Memory Agent and it gives your AI a memory that never turns off. It perfectly pairs with Gemini 3.1 Flash-Lite. That means it runs all day and all night without costing you much money at all. No complicated setup. No special databases. No forgetting. Here's how it works: 3 helpers run in the background → One reads your files — notes, pictures, audio, videos → One connects your ideas every 30 minutes while you sleep → One answers your questions using everything it learned Here's the coolest part: Most AI tools forget everything after you close them. This one keeps learning and connecting dots just like your brain does at night while you dream. Drop a file in a folder → the AI reads it in 5 seconds. 100% Open Source.

Meet Turkish-Gemma-9b-T1: a powerful, specialized LLM fine-tuned for Turkish language tasks. It's not just another translation model. It's a full conversational AI that understands, reasons, and generates fluent Turkish text. This is a big deal for Turkish NLP. https://t.co/LGDwAVCnYx

Talking to a voice AI LLM over ham radio (on UHF 420.69 megahertz, of course!) (Note: cool experiment, but be careful: FCC regs require a licensed control operator to be present at the control point the entire time the LLM is operating.) https://t.co/S2WcCrkp83