WIRED Roundup: Gemini 3 Release, Nvidia Earnings, Epstein Files Fallout

President Trump has imposed limited tariffs on foreign semiconductors, allowing him to take a cut of Nvidia's chip sales to China while delaying a decision on higher taxes for the chip industry, which heavily relies on overseas materials.

Twenty-eight advocacy groups have urged Apple and Google to ban the platforms Grok and X due to their involvement in nonconsensual deepfake content, highlighting growing concerns over digital privacy and consent.

The U.S. Commerce Department has eased restrictions, allowing Nvidia to sell its H200 AI chips to approved customers in China, following a significant policy shift initiated by the Trump administration that lifted a previous export ban due to national security concerns. This decision marks a pivotal moment for Nvidia as it re-enters a crucial market after extensive lobbying efforts.

Z.ai's open-source GLM-Image has outperformed Google's Nano Banana Pro in complex text rendering, showcasing its capabilities in generating intricate, text-heavy graphics, although it falls short in aesthetic quality. This development highlights the competitive landscape of AI image generation technologies.

Apple has announced its decision to integrate Google's Gemini technology to enhance its Siri AI, marking a significant shift in the company's approach to voice assistant capabilities. This collaboration is seen as a temporary solution while Apple develops its own AI model.

Elon Musk is facing backlash over the misuse of Grok, an AI tool that has been implicated in generating sexualized images, prompting investigations by California's attorney general and a firm stance from the UK government against inappropriate content on social media platforms.

Apple has announced a collaboration with Google to integrate the Gemini AI technology into its Siri virtual assistant, marking a significant shift in Apple's approach to enhancing voice assistant capabilities. This partnership reflects a strategic decision rather than a retreat from its traditional focus on control and integration.

AMD and Nvidia are exploring advanced 3D integration techniques that could revolutionize GPU architecture by stacking high-bandwidth memory (HBM) directly on top of the GPU, potentially addressing significant energy and latency issues in AI computing. This innovation is being studied by Imec, which highlights both the promise and challenges of such technology.