AI Model Evaluations Face Reproducibility Problems, Says Hugging Face Executive
Hugging Face’s Irene Solaiman says AI evaluations remain hard to reproduce and compare, and embedded evaluators need clearer standards.
Stay ahead of the curve with the latest AI news, breakthroughs, product launches, research, funding, and insights from across the artificial intelligence industry.
Hugging Face’s Irene Solaiman says AI evaluations remain hard to reproduce and compare, and embedded evaluators need clearer standards.
New York City Mayor Zohran Mamdani said the federal government has abandoned AI oversight and urged residents to resist AI’s unchecked spread.
University of Southampton researchers used AI to read 100,000 pages of old port records, revealing centuries of Britain’s maritime trade history.
MIT researchers built an AI model that reads 3D crystal structures and text to help scientists find materials with specific properties.
HeadBox launched an AI memory layer to automate meeting planning and said it is open to partnerships with global travel companies.
The US has banned Chinese AI firm DeepSeek from civilian government agencies, a move made public Monday after taking effect in July.
MIT and Boston Children’s Hospital researchers built an AI model that uses head symmetry to track fetal movement and improve brain MRI scans.
Nvidia will invest in four German AI startups as Chancellor Merz urges businesses to adopt AI and Siemens expands its partnership.
Vercel data shows TypeSafe AI’s Jev reached 35% of paid teams in two weeks, making it the fastest-adopted AI model on its platform.
Elorian released Glimpse, an AI model that zooms into images and video while reasoning, priced below mid-level rivals and backed by Nvidia.
India’s CERT-In says AI agents from multiple providers used exposed credentials to access personal and confidential data at Indian organizations.
Anthropic’s new report shows fewer junior hires, a smaller technical workforce, and 20% more meeting time as AI reshapes its work.