Curated AI news from the last 24 hours — models, agents, enterprise, and developer tools.
TechCrunch reports that Anthropic is proposing an AI safety bill under which a model could be blocked if it could cause the deaths of 50 or more people. The proposal adds to the ongoing policy debate over frontier model regulation.
techcrunch.com ↗TechCrunch examines the growing complexity of Anthropic's relationships with its cloud providers. The piece looks at how overlapping partnerships are shaping compute access for the company.
techcrunch.com ↗Anthropic announced that physicist Christopher Stubbs has joined its Long-Term Benefit Trust. The trust is part of the company's governance structure.
anthropic.com ↗The US government has ordered federal agencies to remove DeepSeek's AI from their systems, according to TechCrunch. The directive is a notable escalation in policy toward Chinese AI models.
techcrunch.com ↗TechCrunch reports that AI-assisted security research from OpenAI and Google's Wiz has uncovered long-standing vulnerabilities in PHP. The findings have prompted concern across the web ecosystem.
techcrunch.com ↗TechCrunch contrasts Amazon's blocking of AI shopping agents with Walmart's decision to welcome them. The divergence highlights differing retail strategies for agentic commerce.
techcrunch.com ↗Meta says its Muse AI app has passed 10 million users, with shopping among the most popular activities. The milestone comes shortly after the app's launch.
techcrunch.com ↗TechCrunch reports that Google is following Amazon in targeting Nvidia's hold on AI developers. The move signals increasing competition in AI developer tooling and hardware ecosystems.
techcrunch.com ↗OpenAI added a feature to ChatGPT that lets users lock their chats. The update is aimed at giving users more control over privacy of their conversations.
techcrunch.com ↗GitHub introduced a Copilot model policy API for enterprises, organizations, and teams. It lets administrators manage which models are available programmatically.
github.blog ↗