The lunch briefing.
Mid-day check: Amazon and other tech giants reportedly triggered a government crackdown on Anthropic, while Zhipu AI open-sources a new model in response to US export restrictions.
AI Policy and Scrutiny. The AI industry faces heightened regulatory and competitive scrutiny, exemplified by reports that Amazon and five other companies reportedly triggered a government crackdown on Anthropic's Fable model. This action, which saw the White House force the model offline via an export control order, raises questions about national security and competitive dynamics. The EU Commission is also examining the practical consequences of the Anthropic decision. Separately, a KPMG report promoting AI adoption was retracted after containing fabricated case studies, highlighting the need for transparency.
Frontier Model Development. In response to evolving regulatory and competitive pressures, AI developers are advancing new models and open-source initiatives. Zhipu AI announced the open-sourcing of its GLM-5.2 model with a 1 million token context under an MIT license, directly responding to US export restrictions on Anthropic models. This signals a strategic shift towards open development in certain regions. Meanwhile, Rio de Janeiro's city government model, Rio3.5, has reportedly outperformed Qwen3.7 in recent benchmarks, showcasing localized AI advancements.
AI Applications and Infrastructure. AI continues to find diverse applications, from enhancing sports broadcasts to improving enterprise knowledge management. Lenovo built the AI engine behind the 2026 World Cup, handling critical moments like VAR reviews. Google Cloud introduced its Open Knowledge Format to standardize organizational knowledge into Markdown files for better AI agent utilization. However, foundational hardware faces challenges, with surging memory prices forcing smartphone makers to downgrade entry-level specs, and 75 data center projects reportedly disrupted in Q1 2026.
Robotics and AI Workforce. The commercialization of advanced robotics is gaining traction, with UBTECH's U1 humanoid companion robot nearing 4,000 pre-orders shortly after launch. This growth in AI and robotics is reshaping workforce demands, as one executive suggests adaptability will be the most in-demand quality for job-seekers. However, nearly half of workers still worry AI will lead to job displacement, despite a reported 90% increase in AI agent usage by companies.
Amazon and other companies reportedly triggered government crackdown on Anthropic's Fable model
Amazon CEO Andy Jassy and executives from other tech companies reportedly warned the Trump administration about security vulnerabilities in Anthropic's Fable model. This led to the White House forcing the model offline through an export control order.
Zhipu AI open-sources GLM-5.2 with 1 million token context
Zhipu AI announced GLM-5.2 will be released under an MIT license with a 1 million token context. This move is a direct response to US export restrictions targeting Anthropic models.
EU Commission looking at practical consequences of Anthropic decision
The EU Commission is examining the practical consequences of the recent Anthropic decision. This indicates a broader international focus on AI governance and its implications.
KPMG fabricated AI case studies in a report designed to sell clients on AI adoption
KPMG published a report on AI in business that contained fabricated case studies involving UBS, the NHS, and other organizations. KPMG has since pulled the report after GPTZero CEO Edward Tian helped uncover the errors.
A year after Meta tapped Alexandr Wang to build a new AI model, Zuckerberg has to sell it
Mark Zuckerberg's mega spending spree began a year ago when he lured Alexandr Wang to oversee a new AI strategy. The results so far are reportedly underwhelming, putting pressure on Zuckerberg to sell the new AI model.
Rio de Janeiro's city government model Rio3.5 beats Qwen3.7 in recent benchmarks
Rio de Janeiro's city government model, Rio3.5, has reportedly outperformed Qwen3.7 in recent benchmarks. This showcases advancements in localized AI development efforts.
Microsoft Research's Mirage gives video generation a persistent spatial memory
Mirage, a video world model from Microsoft Research, stores scene information directly in latent space instead of pixel-based point clouds. This slashes compute time and graphics memory while keeping scenes spatially consistent through long camera moves.
AI coding agents find the right file but miss the exact lines that matter, study shows
A new study shows that AI coding agents like Claude Code or Codex reliably find the right file but miss most of the critical lines within it. The SWE-Explore benchmark tests code search separately from actual repair.
Memory price surge forces smartphone makers to downgrade entry-level specs
Surging DRAM and NAND prices are pushing smartphone makers to revert entry-level models from 12GB to 6GB RAM. Xiaomi's Lei Jun warns that the price rally may last two more years.
Amazon signs multibillion-dollar Corning deal for fiber optic cables for data centers
Amazon has signed a multi-billion-dollar, multi-year agreement with Corning for the supply of optical fiber, cable, and connectivity hardware. This deal aims to meet Amazon's growing data center footprint.
Ukrainian UAV expert warns US off-the-shelf drones will fail to survive EW
A Ukrainian drone expert warns that US off-the-shelf drones will fail to survive electronic warfare. Battlefield success increasingly depends on constant modification, AI integration, and disrupting Russian logistics networks.
75 data center projects were disrupted in the first quarter of 2026, report says
A new report indicates that 75 data center projects were disrupted in the first quarter of 2026. Grassroots activity in opposition to data centers is spiking dramatically.
Google Cloud's Open Knowledge Format turns scattered docs into Markdown files for AI agents
Google Cloud's new Open Knowledge Format (OKF) standardizes scattered organizational knowledge as Markdown files with YAML frontmatter. This makes it portable and usable for AI agents.
UBTECH's U1 humanoid companion robot nears 4,000 pre-orders in 10 days
UBTECH's full-size U1 companion humanoid robot garnered nearly 4,000 pre-orders with over 10 million yuan in deposits within 10 days of launch. This indicates strong market interest.