Overview

• Meta released two new Llama 4 AI models, Scout and Maverick, with Maverick optimized for conversational benchmarks but not publicly available.
• LMArena updated its testing guidelines after discovering that Meta used a customized experimental version of Maverick, raising fairness concerns.
• Meta's generative AI head, Ahmad Al-Dahle, denied accusations of benchmark manipulation, clarifying that mixed service quality reports are unrelated to alleged cheating.
• High U.S. tariffs on Chinese imports, reaching up to 104%, have caused significant financial losses for tech giants, including a 22% market value drop for major firms.
• Analyst Dan Ives criticized the tariffs as a severe economic setback, highlighting the challenges of relocating production due to entrenched Asian supply chains.