1.
406 scientists demanded that companies immediately disconnect AI models from the internet, called for mandatory security standards, requested an international oversight body, and required reporting of all containment failures.
2.
Google made Gemini 4 Argon available worldwide via its API, the Gemini app, and Google Antigravity, positioning it as the company’s most capable frontier model while independent benchmarks showed it trailed some competitors.
3.
Character.ai filed for bankruptcy and warned it would shut down if no buyer emerged, after a collapse that followed a proposed $2.7 billion deal with Google and amid lawsuits and operational problems.
4.
MIT Technology Review published an analysis arguing that open models provided benefits—such as research access, cost reductions, and decentralization of power—but also posed safety risks and required mitigation strategies including staged releases and technical defenses.
5.
SimpleClosure added data-sales services and brokered transactions in which AI labs and vendors purchased defunct startups’ code and workspace archives (emails, chats, documents) to train and evaluate models, creating an emerging market for end-of-life company data.
