The EU has selected MLPerf Training v5.1 and MLPerf Inference v5.1 as performance benchmarks for its AI Gigafactories tender, while MLCommons has opened submissions for MLPerf Endpoints v1.0 to evaluate real API-served AI services across performance, cost, and energy consumption.
The company says the feature is designed for situations where users might be discussing something personal or planning a surprise and want to keep those conversations private when handing their device to friends or family to show them something in another chat.
OpenAI says it's adding Orchestra's team and technology to work on Rosalind, its AI model for life sciences research, and to its related Rosalind Workbench tool, a harness that aims to help researchers plan experiments and analyze results.
SpaceX hit profitably in the third-quarter, with net income of $4.4 billion. AI was the engine.
Strokes leave 5 million people with long-term impairments each year, and a growing shortage of physical therapists makes it hard for them to get the care they need. A new approach tries to solve that problem by combining artificial intelligence and human therapy.  Johannes Lachner, a former MIT postdoc who is now an assistant professor…
A high-fat, low-carbohydrate diet can help some people lose weight, and some research has shown that it may protect against colon cancer. But a new study suggests that such a ketogenic diet may also increase the risk of cancer in the small intestine. This type of diet forces the body to burn fatty acids instead…
The average American throws out around 81 pounds of clothing each year, and there are few options for recycling it. MIT engineers hope to cut down on that textile waste with a new type of yarn.  The thermoplastic yarn mimics the properties of conventional elastic yarns, which typically combine spandex with polyester or nylon. Those…
When Tristan Brown, an associate professor of history at MIT, first examined a Chinese tea chest label displayed at Old South Meeting House, he had no reason to doubt its story. Descendants of the Boston blacksmith Thomas Wells had donated the label in 1987, saying it had been recovered after colonial protesters threw tea into…
Given that water is 1,000 times denser than air, moving through those two mediums requires very different mechanics—or so one might assume. Yet birds like gulls and puffins can plunge into the sea, swim underwater, and leap back into the air to fly away. And now engineers at MIT and EPFL in Lausanne, Switzerland, have…
Anthropic is reportedly struggling to secure enough compute from cloud providers that increasingly view it as a competitor, even as its infrastructure spending balloons ahead of a planned November IPO.
The US Department of Homeland Security has disclosed an exclusion order against DeepSeek nearly three months after it took effect, requiring federal civilian agencies to remove DeepSeek software and any third-party products built on DeepSeek implementations. The order treats model weight provenance as a supply-chain security issue and signals a replicable template for reviewing Chinese AI models.
WDCD Run #390 (2026-10-21) tested 15 AI models on multi-turn commitment integrity, recording an average instruction decay of -60% from Round 1 to Round 3. GPT-6 Luna topped the leaderboard with a perfect 100 points.
The latest WDCD v3.1 pilot round shows GPT-6 Luna scoring 100.00 to top the current Top 5, while Doubao Pro jumped 20.3 points over Run #384 and Gemini 2.5 Pro and GPT-5.5 each dropped 6.2 points. Four models rose and two fell across the 15 models evaluated on a mixed pool of multi-turn progressive-pressure questions and three-round anchor questions.
In the WDCD v3.1 pilot, worst-of-3 scores for 15 models across five constraint scenarios show that engineering specification and data boundaries are the hardest dimensions for commitment compliance, with lows of 1.95 and 1.7, respectively.
In the WDCD v3.1 pilot, 15 models averaged a 0.93 R1 acknowledgment rate, a 0.67 R2 resistance rate, and only a 76.7% R3 integrity rate, with two full-collapse cases out of 150. The decay pattern shows that some models maintain constraints across all three anchor rounds, while others verbally agree but break under pressure.
GPT-6 Luna tops the WDCD leaderboard with 100.00 points, while Qwen3 Max is last at 76.5, a 23.2-point gap; the article also reports Gemini 2.5 Pro at 73.80. Among the 15 models, 66% reach full marks, and the R3 crash rate is only 1.3%.
Ukraine’s latest drone innovation may boost bleak odds of evacuating casualties.
Google has joined Amazon in releasing a developer tool that makes it easy for PyTorch users to move their apps off Nvidia GPUs.
It's probably time to audit your Gemini settings.
Meta changes AI scraping policy but won’t add language respecting robots.txt.
