Researchers from Google propose Simula, a framework that treats synthetic data generation as dataset-level mechanism design to address data scarcity for specialized AI. Simula uses first-principles reasoning to enable fine-grained control over dataset coverage and complexity.

Google Research scientists Gal Elidan and Yael Haramaty report a study demonstrating that generative AI can assess 'future-ready' skills comparably to human experts. Conducted with New York University, the Vantage experiment is now accessible via Google Labs.

Project Glasswing is an initiative by Anthropic aimed at securing critical software in the context of the AI era. The article provides a URL to the project page and a comments section on Hacker News.

Anthropic has signed an agreement with Google and Broadcom for multiple gigawatts of next-generation compute capacity, expected to come online starting in 2027, to power its Claude models and meet customer demand.

Researchers from Google Research propose a framework to systematically evaluate behavioral alignment in large language models (LLMs) by converting established assessments into large-scale situational judgment tests. The work, led by Amir Taubenfeld, Zorik Gekhman, and Lior Nezry, aims to map and und…

Codex has added pay-as-you-go pricing for ChatGPT Business and Enterprise, offering teams a more flexible option to begin and expand usage.

Anthropic and the Australian government signed a Memorandum of Understanding to collaborate on AI safety research and support Australia’s National AI Plan. Anthropic CEO Dario Amodei met with Prime Minister Anthony Albanese to formalize the agreement and announced AUD$3 million in partnerships with…

Researchers propose an evaluation framework for ML models using 'gold' ratings data to balance the number of items and raters per item, aiming to improve reproducibility and capture human disagreement in AI benchmarks.

Google Research introduced TurboQuant, a set of quantization algorithms designed for massive compression of large language models and vector search engines. These algorithms operate on vectors, which are fundamental to how AI models process information.