#coding-agents

AI agents are increasingly acting on real infrastructure without adequate authorization checks, from a gym-booking exploit to an unresolved Hugging Face attack timeline. Economists and researchers are meanwhile debating whether to slow AI self-improvement through regulation, with dueling arguments this week over data center bans and safety-rule timing. Elsewhere, GitHub retired its unified LLM API, and the FDA rejected a manufacturing-flagged cancer drug.

Microsoft unveiled two proprietary MAI models at Build, including a 1-trillion-parameter reasoning model and a 137B coding model for GitHub Copilot, as Uber hit its AI budget ceiling after burning through its full 2026 allocation in four months. Alnylam struck a $2 billion deal with Inceptive Nucleics to build AI foundation models for RNA interference therapeutics. Trump signed a narrower AI oversight executive order requiring a 30-day review before powerful open-weight models can be released. The Trump administration also moved to strip job protections from roughly 8,000 NIH staff, including officials who oversee research grants, while Medicaid work requirements drew criticism for going beyond what advocates had anticipated.