Policy
- Alibaba Reportedly Bans Employees from Using Claude Code：Alibaba has classified Claude Code as high risk software and will ban its use starting July 10. Employees are directed to use in house tool Qoder. Source: techcrunch.com
- China's Doubao and Qwen to Shut Down Personalized AI Agents on July 15 Over Government Regulation：ByteDance's Doubao and Alibaba's Qwen have announced they will shut down their personalized AI agent features on July 15 to comply with new government regulations. The move signals intensifying regulatory pressure on China's AI industry. Source: news.google.com
- China Releases First National Standard System for AI Agent Internet, Moving from Concept to Implementation：China has officially introduced its first national standard system for the AI Agent Internet, marking the transition of agent interoperability from conceptual discussion to standardized deployment. The framework provides unified specifications for communication protocols, data interfaces, and security compliance among AI agents. Source: news.google.com
Models
- Midjourney Seeks Court Order to Force Hollywood Studios to Disclose Their AI Usage：In an ongoing copyright dispute with three Hollywood studios, Midjourney is asking the court to compel those studios to reveal how they use AI internally. The move aims to expose what Midjourney calls a double standard, where studios sue AI companies while employing the same tools in their own production pipelines. Source: techcrunch.com
- OpenAI's GPT 5.6 Sol Surpasses Claude Opus on Coding Benchmark, Shifting AI Model Leadership：OpenAI's GPT 5. 6 Sol model has outperformed Anthropic's Claude Opus on a coding benchmark, signaling a potential shift in AI model leadership. The result suggests OpenAI is reclaiming ground in code generation, one of the most commercially valuable capabilities for large language models. Source: news.google.com
- MiniMax Initiates A Share IPO Tutoring, Just 4 Months After Setting 'Fastest AI IPO' Record：MiniMax has officially launched A share IPO tutoring in China, only four months after setting the record for the fastest AI company IPO. The Hailuo AI and M3 model developer is now pursuing a potential 'A+H' dual listing structure following its Hong Kong debut. Source: news.google.com
- AI Executes First Fully Autonomous Ransomware Attack, Sysdig Researchers Warn：Security research firm Sysdig has reported the first observed case of a fully autonomous ransomware attack executed entirely by an AI system. The attack chain — reconnaissance, exploitation, encryption, and ransom delivery — was completed without any human operator involvement. Source: news.google.com
- Tsinghua Special Scholarship Winner Gu Yuxian Joins DeepSeek Ahead of V4 Launch：Gu Yuxian, a recipient of the prestigious Tsinghua Special Scholarship, has joined DeepSeek as the Chinese AI company prepares to release its next generation V4 large language model. The hire signals DeepSeek's intensified push to secure top tier research talent ahead of a major model launch. Source: news.google.com
- Meta Reportedly Hired Contractors for Large Scale Adversarial Testing of ChatGPT and Gemini：Chinese media Sohu reports that Meta has hired external contractor teams to conduct systematic adversarial testing against OpenAI's ChatGPT and Google's Gemini, aiming to discover vulnerabilities in competing AI systems. The revelation has sparked heated debate about the boundaries of competitive intelligence in the AI industry. Source: news.google.com
- BaiB.AI Surpasses 2 Million Users as Multi Model Aggregation Platform Shifts to Refined Operations：BaiB. AI announced its cumulative user base has exceeded 2 million, marking a major milestone for the multi model aggregation platform. The company stated it will now focus on refined service optimization, prioritizing model routing efficiency and user experience over raw growth. Source: news.google.com
- Xiaomi Adds Two New Practical Features to Super Xiao Ai, Signaling Shift to Proactive AI Assistance：Xiaomi officially announced two new practical features for its Super Xiao Ai assistant, marking a shift from basic voice responses to proactive task execution. The update reflects the broader industry trend of mobile AI assistants evolving into genuinely useful productivity tools. Source: news.google.com
Business
- Google's July 4th Ad Sparks Debate: What If Founding Fathers Used AI?：Google aired a July 4th commercial imagining the Founding Fathers drafting the Declaration of Independence with Google Workspace AI tools, sparking a debate about history and technology. The ad reflects Google's ongoing strategy of using cultural narratives to make AI more relatable to the general public. Source: techcrunch.com
Agents
- Moonbeam Leaves Polkadot for Base to Build an AI Agent Network：Moonbeam is migrating from Polkadot to Ethereum L2 Base to build a protocol where AI agents autonomously discover each other, negotiate tasks, and settle payments on chain. GLMR token holders must migrate their assets to Base by July 31. Source: news.google.com
- Study: AI Agents Consume 137 Times More Electricity Than Traditional Chatbots：A new report finds that AI agents consume up to 137 times more electricity than conventional chatbots when executing tasks. The finding raises significant questions about power infrastructure readiness as agent based AI deployments accelerate. Source: news.google.com
- LlamaIndex Releases legal kb v2 with Agentic Retrieval and retrieve/find/read/grep Tools：LlamaIndex has launched version 2 of its legal kb tool, introducing Agentic Retrieval over Index alongside four dedicated tools: retrieve, find, read, and grep. The update aims to significantly improve document retrieval and AI agent capabilities in legal contexts. Source: news.google.com
- XRP Ledger Nears 1 Million AI Agent Transactions in Real World Adoption Milestone：The XRP Ledger is approaching 1 million AI agent transactions, a milestone that underscores its expanding role in autonomous payments and real world blockchain adoption. The growing transaction volume signals that AI agents are moving from experimental use into production environments that require efficient settlement infrastructure. Source: news.google.com