https://t.co/jVJcUfyjhF

We're taking the next big step with Researcher. With Computer Use, it can now securely browse the open and gated web to find hard-to-locate information—even across hundreds of sites—and handle multi-step tasks to uncover insights, take action, and create richer reports.

In engineering, you should first solve your problem within a relaxed design space, and only *then* should you determine the minimal constraints required to implement that solution. Don't settle on the hardware before you know what software you'll need to run. Don't design the

Watching Claude Code execute in an infinite loop with defined objectives while autonomously handling its own git commits and https://t.co/SGbauPrlvn is genuinely mind-blowing.

don’t take it personally ‼️

Love the new ARC-AGI-3 games. Having to discover the rules is very interesting. I think there are some priors about playing games required to be able to efficiently learn the rules. I doubt any human can pass them in a reasonable time even if they never played games.

Excited to share two advances that bring us closer to real-world impact in healthcare AI: SDBench introduces a new benchmark that transforms 304 NEJM cases into interactive diagnostic simulations. AI must ask questions, order tests, and weigh costs, mirroring the complexity of

The world’s first multimodal, bilingual radiology dataset could reshape the way radiologists and AI systems make sense of X-rays. PadChest-GR, developed by the University of Alicante with Microsoft Research, has the potential to advance research across the field for years to

🔍 Why it matters: - 4,555 chest X-ray studies - 10,459 sentence-level clinical findings in Spanish & English - Precise spatial annotations - Designed specifically to ground AI in multimodal data and improve clinical verification and interpretability

Learn more from the blogpost:

NEJM-AI: Introducing PadChest-GR, the world's first bilingual, grounded radiology reporting benchmark for chest X-rays! https://t.co/7PMiIf1Uts

There is a lot of hyped up excitement about MCP, but I think most people miss the essence of why MCP is a such a BIG thing. As the fundamental theorem of computer science states "All problems in computer science can be solved by another level of indirection". That is exactly

Today I did my first PR in GitHub without writing a single line of code, just asking questions and reviewing. Imagine a team of copilot doctors working with you.

A big day – we are making the largest dataset of chest X-rays available to the world. This is the first dataset to cross 100k patients. https://t.co/dPSSpFHkae

Three Microsoft CEOs walk into a room on Microsoft’s 50th anniversary … and are interviewed by Copilot!

Lee Iacocca understood the power of truly listening to your team. His leadership philosophy, which helped create the Ford Mustang and save Chrysler from bankruptcy, wasn't taught in business school:

GPT-4.5 got better at drawing

Maybe pretraining still works after all??

Livestream in 4.5 hours.

GitHub Copilot is all-in on agents. Check out Agent Mode, and a first look at our Autonomous SWE agent.