DeepMind Speeds R&D with Startup Mode, Court Keeps Anthropic Supply-Chain Risk Tag
Audio in Mandarin Chinese · English transcript below
⚡ DeepMind goes startup, White House rewrites rules, court flags Anthropic, China bets DeepSeek V4 vs chip ban, JD/Meituan pull plugs, Gemini drops notebook, OpenClaw 150k stars ignites SaaS battle royale, finance AI posts bond, geometry data war, ModelScope one-click jump in—miss a day, lose the race!
Global AI development is increasingly deeply intertwined with geopolitical and national security considerations, with technological autonomy and supply chain security becoming core to national competition. US courts upholding Anthropic's "supply chain risk" label underscores the strict scrutiny governments apply to the national security attributes of AI companies. Meanwhile, the potential reliance of China's DeepSeek V4 model on domestic chips reveals the decisive role critical hardware plays in geopolitical competition, prompting companies to accelerate internal R&D or tighten the use of external tools to navigate the increasingly complex AI ecosystem.
Today's Top 3 Headlines
- 科技
联邦法院驳回Anthropic（AI公司）撤销“供应链风险”标签的申请
周三，美国哥伦比亚特区联邦上诉法院由三名法官组成的合议庭在四页裁决中驳回Anthropic（AI公司）要求阻止国防部将其列为安全风险的动议。法院认为，Anthropic未能满足暂停该标签的严格要求，称“公平权衡偏向政府”。该裁决并非最终决定，案件将继续审理。此前，国防部与Anthropic就一份价值2亿美元的合同条款产生分歧，国防部长皮特·赫格塞斯将该公司列为“供应链风险”。该标签通常用于存在国家安全担忧的外国公司，一旦被列入，企业基本被禁止与美国政府实体开展业务。
Source ↗
+10 more headlines
- 中国寄望AI突破，静待DeepSeek V4登场
- 京东据称已禁止外部AI工具，美团则收紧了对阿里巴巴Qwen模型的使用。
- David Sacks推动白宫推进亲行业AI政策
- Gemini推出“笔记本”功能，集中管理项目资料
- ModelScope 全面实施指南发布
- AI代理时代来临：Claude（AI助手）与OpenClaw带来机遇与混乱
- AI代理金融幻觉风险催生代理风险标准
- 大语言模型带来的流量转化率可达30%–40%，而大多数企业尚未对此进行优化。
- 研究显示，半数Z世代使用AI，但好感度正在下降
- Google AI优势成隐藏王牌，Alphabet股价看涨