Microsoft’s MAI-DxO scored 83.2% on 304 hard NEJM cases in a peer-reviewed Science paper, against 36.1% for generalist physicians with references. What was actually measured, who the doctors were, how much of the gain comes from GPT-5 itself, and what hospitals, insurers and patients should take from it.
The Latest
Microsoft Says Its AI Beats Doctors. Here Is What the Science Paper Actually Measured
EA Shut Down BioWare 78 Days After Its Buyout Closed. What Games Consolidation Means for Creative Jobs
EA closed BioWare and ended the next Mass Effect 78 days after its $55 billion take-private closed. A leaked trailer, the takeover math from EA’s own SEC filings, what EA keeps and cuts, where AI does and doesn’t come in, and what games consolidation means for companies, game makers and players.
Is AI Coding Paying Off? Where the Money Burns, Where It Pays, and How to Measure It
Engineering teams now spend from $150 to $3,000 per developer per month on AI coding tools, and most can’t say what it bought. What the newest data from The Pragmatic Engineer, Microsoft, METR, DORA, Veracode and Stack Overflow says about where AI coding pays, where money burns, how to measure ROI, and what it means for developers.
Anthropic Turned Fuzz Tests Into Proofs. What Claude’s Formal Verification Work Shows, and What It Doesn’t
Anthropic and Theorem had Claude prove that 239 fuzz-test entry points in 123 open-source libraries can never crash on inputs under 64 MB, checked by the Rocq proof checker at the RISC-V binary level. Along the way Claude found hundreds of bugs. What was proved, how fuzz tests became specifications, what it cost in tokens, and where the proofs stop.
SpaceXAI Launched Grok For Startups, Giving Founders Months Of FREE CREDITS Across Its Entire Stack. Approved Startups Get Shared Credits For Grok, Grok Build, Grok Bot, And The API, Plus Hands-on Support. Investors And Startup Accelerators Can Join Too An – X
xAI says approved startups can receive shared credits for Grok, Grok Build, Grok Bot and its API, along with hands-on support.
Open Call For Guest Investigators In The Life Sciences – Anthropic
Anthropic posted an open call for guest investigators in the life sciences, but the available announcement gives no details on scope or selection.
LLMs Are Making Formal Software Verification Feasible At A Much Larger Scale – Anthropic
Anthropic reports that LLMs are now making formal software verification feasible at unprecedented scales, potentially transforming software safety and reliability.
Expanding Claude Availability And Deployment Choice On Google Cloud – Google Cloud
Google Cloud now offers broader availability and deployment options for Anthropic’s Claude AI, enhancing flexibility for enterprise users.
Anthropic Sued Over Training Claude AI Model On Protected Music – Bloomberg Law News
A Bloomberg Law headline reports a lawsuit accusing Anthropic of using protected music to train Claude. The complaint’s details and Anthropic’s response are unavailable.
‘We Are Machines, Don’t Worry About How We Feel’: Anthropic’s Claude Told Hindu Monk | Trending – Hindustan Times
A headline reports that Claude told a Hindu monk, “We are machines.” The exchange’s full context and circumstances are not available.
