NVIDIA unveiled Ising, a family of open-source AI models designed to accelerate the path to useful quantum computers. The suite includes a 35-billion-parameter vision-language model for quantum processor calibration and convolutional neural networks for error-correction decoding that are 2.5x faster and 3x more accurate than existing tools. Major quantum labs including Fermilab, Harvard, and the UK National Physical Laboratory are already adopting the models, and the announcement sparked a rally in quantum computing stocks.
The Danish pharma giant and OpenAI have formed an enterprise-wide partnership to apply advanced AI across drug discovery, manufacturing, supply chain, and operations. The deal is molecule-agnostic — not tied to any single pipeline asset — and aims to compress timelines from research to patient. Pilot programs launch immediately with full integration targeted by end of 2026.
European cybersecurity agencies have been largely excluded from early access to Anthropic's Claude Mythos, a model reportedly capable of identifying and exploiting technical vulnerabilities at superhuman levels. Access under "Project Glasswing" has been limited to select US tech giants and the UK AI Safety Institute, while only Germany has opened a dialogue within the EU. The European Commission has publicly backed Anthropic's cautious staged rollout, but the access gap is fueling transatlantic AI governance tensions.
A Nature report drawing on the 2026 Stanford AI Index found that the best autonomous AI agents perform only half as well as PhD-level human experts on complex scientific workflows. Despite these limitations, researcher adoption of AI tools continues to surge, creating a paradox where enthusiasm is outpacing actual capability on the hardest problems.
After losing nearly 40% of its workforce, the General Services Administration is deploying its internal AI tool USAi to automate up to one million work hours — roughly a full year of 500 employees. GSA says it has already identified 400,000 automatable hours and is nearly halfway to its goal, following an "eliminate, optimize, automate" playbook that could expand government-wide if successful.
UChicago's Polsky Center announced a new partnership with AI Research Commons, Microsoft, and NVIDIA to identify and accelerate AI startups in the Midwest. Selected teams will be announced in early summer, with the program providing compute resources, mentorship, and enterprise connections to founders building outside the traditional coastal AI hubs.
