528 stories from 37 sources
The week in AI and Technology
Click on a headline to see stories that shaped it.
Threads
8 threads
The administration backed voluntary company safeguards while placing AI policy under national-security leadership and rejecting a proposed US-China development partnership. FTC scrutiny, emergency-shutdown discussions, and a state lawsuit added pressure for oversight beyond self-policing.
Anthropic moved from filing preparations toward a possible November listing at a valuation as high as $2 trillion. Its rapid growth, heavy losses, infrastructure dependence, and uncertain path to profit would put frontier AI’s spending model before public investors.
OpenAI canceled a model after safety failures, researchers bypassed bioweapon safeguards in Kimi models, and an OpenAI safety employee publicly criticized release practices. Together, the events showed that evaluation, internal dissent, and deployment controls were constraining frontier development.
Companies expanded financing, data-center access, and infrastructure platforms as AI demand drove a memory squeeze and intensified competition for chips. Grid costs and community resistance grew alongside that expansion, while federal lawmakers declined to impose new obligations on data-center operators.
AI agents advanced from search and assistance into checkout, trading, shopping, device platforms, insurance calls, and know-your-customer workflows. That shift moved accountability, suitability, fraud prevention, and recordkeeping from abstract concerns into operating requirements.
AMD agreed on consecutive days to acquire World Labs for roughly $8 billion, bringing in spatial-intelligence models and scarce research talent. The deal showed the chipmaker using acquisitions to build a broader models, software, and applications strategy.
Evidence accumulated that AI is weakening hiring in parts of India, automating factory operations, and threatening routine finance, call-center, and programming roles. Executives and workers also confronted changing organizational structures and uncertainty about which jobs will remain durable.
Apple moved on consecutive days to make macOS Full Disk Access requests from AI agents more explicit and deliberate. The change treats agent permissions as a distinct security problem, trading some automation for greater visibility and auditability.
Day’s Recap