AI
-
Policy & Regulation
DeepSeek and MiniMax clear China's new AI export registration and keep releasing models abroad, a state media-linked account says
A social media account tied to Chinese state broadcaster CCTV says DeepSeek and MiniMax have completed registration under Beijing's draft AI export-control rules and the steps needed to open-source models and authorize t…
-
New Products
Google Cloud adds an Anthropic-run way to use Claude, giving customers new Claude Platform features on day one inside their Google accounts
Google Cloud has made generally available a deployment of Claude that Anthropic operates on Google infrastructure, with Anthropic as data processor and immediate access to platform features such as Claude Managed Agents,…
-
Business
OpenAI is closing in on a raise of as much as $30 billion that would value it at $1.25 trillion, with UAE investors and Coatue in the lead, reports say
CNBC and Bloomberg report that OpenAI is in the final stage of talks on a round of roughly $30 billion that values it at $1.25 trillion before the new money, a lower figure than the $1.4 trillion discussed earlier, as a …
-
Policy & Regulation
US appeals court takes a Mississippi case away from a federal judge whose orders contained AI-invented parties, quotes and a fake precedent
The 5th US Circuit Court of Appeals vacated Judge Henry Wingate's injunction against Mississippi's law restricting DEI programmes in public education and ordered the case moved to another judge, citing fabricated materia…
-
New Products
AWS lets SageMaker customers write their own model training loops while it runs the GPUs, as training sessions reach general availability
Amazon Web Services has made training session APIs generally available in SageMaker AI, so teams can script custom fine-tuning and reinforcement learning against a live, stateful GPU session billed per token, starting wi…
-
Research
Anthropic says Claude wrote machine-checked proofs that 123 open-source libraries cannot crash on their fuzz tests, working directly on compiled code
In a project with the verification startup Theorem, Anthropic had Claude produce Rocq proofs covering 239 fuzzing entry points in 123 OSS-Fuzz libraries at the RISC-V binary level, patching hundreds of bugs it hit along …
-
New Products
OpenAI opens its Decisions API to all developers, pitching GPT-6 Luna as a fast, input-priced engine for yes-no, choice and scoring calls
OpenAI's Decisions API, which returns a probability, a pick from fixed options or a rubric score instead of free text, is now generally available, adds GPT-Live voice integration and US and EU processing, and charges onl…
-
Infrastructure & Chips
NVIDIA launches Minerva, software that maps an entire AI factory and uses machine learning to flag failing parts before training jobs stall
NVIDIA's new Minerva AIOps platform pulls telemetry from GPUs, networking, storage, schedulers and workloads into a topology graph, predicts failures and folds alert storms into single incidents, with IREN already using …
-
Research
Microsoft's MAI-DxO diagnostic system reaches Science with results ahead of generalist doctors on past cases, as a live hospital trial is planned
A peer-reviewed Science paper reports that Microsoft AI's multi-agent MAI-DxO reached about 83% accuracy on complex NEJM cases versus about 36% for generalist physicians with references, and 80% versus 47% on emergency-d…
-
Research
Google DeepMind releases ForTy, a free 10-meter world map that tells natural forests from plantations and tracks yearly changes since 2017
Built with Google Research, the World Resources Institute and IIASA on AlphaEarth Foundations embeddings, the Forest Typology map sorts every 10-meter patch of land into six FAO-based classes and was right on forest type…