Keep pulling the thread on Sam Altman.
At the G7 meeting, Anthropic CEO Dario Amodei called for international cooperation on AI risk, including structured access to frontier models and chip trade deals that exclude China.
French President Emmanuel Macron warned that the Trump administration holds an "AI kill switch" and made a plea for the US not to restrict allied access to its frontier AI models.
The United Kingdom's request for a carve-out for its nationals and companies from US AI export control restrictions was denied.
The European Commission has unveiled a plan to build up to five "AI gigafactories" with a €20 billion commitment, expecting to deploy around 100,000 GPUs.
The US government ordered Anthropic to revoke access to its Mythos model for the Korean telecom company SK Telecom due to concerns about its ties to China.
AI researcher Noam Shazir has left Google to join OpenAI, where he will work on creating new architectures for AI models.
In 2024, Google spent $2.7 billion to license technology from Character AI as part of an acquihire deal to bring its founder, Noam Shazir, back to the company as the technical lead for the Gemini project.
ZAI's GLM-5.2 model is ranked number one on BridgeBench for reasoning, reportedly beating Fable-5 at one-tenth of the cost and achieving 300 tokens per second.
Microsoft is reportedly considering using a locally hosted, fine-tuned version of the Chinese model DeepSeek V4 to power its Copilot Co-Work product as a lower-cost alternative.
In an experiment with Fireworks, Harvey created a worker-advisor agent system where an open-weight model (GLM 5.1) delegated complex tasks to a frontier model (Opus 4.7), resulting in both lower costs and increased performance.
AI leaders including Sam Altman of OpenAI, Demis Hassabis of Google DeepMind, Alexander Wang of Meta, and Dario Amodei of Anthropic attended the G7 meeting as part of the US contingent.
Canadian Prime Minister Mark Carney stated that the United States could lead an international AI coalition.