Dario Amodei, co-founder of Anthropic, predicts the emergence of superintelligent AI could happen as soon as next year and stresses the need for societal frameworks, including universal basic income to address potential job losses. In a related context, Chinese officials advised AI experts to avoid travel to the U.S. due to fears of detention and security risks, highlighting rising global tech tensions.

In California, Senator Scott Wiener has reintroduced a bill aimed at protecting whistleblowers reporting harmful AI technologies, proposing the establishment of “CalCompute,” a cloud computing resource for AI research across the University of California system. Meanwhile, the U.S. government is contemplating broader export restrictions on AI chips to China, potentially impacting companies like Nvidia, which could face revenue declines amid deteriorating U.S.-China relations.

OpenAI announced the release of the research preview for its most advanced AI model, GPT-4.5, marking a significant milestone in AI development. Collaboratively, MIT and GlobalFoundries will work on enhancing AI chip efficiency, focusing on power gains for both data centers and low-power edge devices. Microsoft expanded its Phi model lineup, introducing the Phi-4-mini for mobile and Phi-4-multimodal for processing diverse data types.

On the performance front, Cerebras partnered with Mistral AI, achieving remarkable AI processing speeds with Mistral's Le Chat application. Meanwhile, the AI Action Summit in Paris brought together over 1,000 participants, co-chaired by Presidents Macron and Modi, to explore economic opportunities driven by AI, emphasizing collaborative investments. Lastly, independent developers have successfully revived the Humane AI Pin after its official shutdown, repurposing it through community-driven coding efforts.

Learn more about your ad choices. Visit megaphone.fm/adchoices