中国AI公司Moonshot推出免费开源模型Kimi，其性能与OpenAI和Anthropic的收费产品相当[1]，引发美国政府和业界对中国AI竞争的关注。特朗普政府内部对如何应对中国AI模型产生明显分歧[1]，一些官员主张采取限制措施，而另一些人则反对政府干预政策[1]。
Anthropic与版权方达成和解协议，同意支付15亿美元的赔偿金[1]，创下AI领域版权诉讼的历史记录[1]。此外，特朗普政府的AI安全负责人Chris Fall在任职三个月后宣布辞职[1]。
Chinese AI company Moonshot recently launched Kimi, a free open-source model whose performance rivals paid offerings from OpenAI and Anthropic, prompting divisions within the Trump administration over how to respond to advancing Chinese AI capabilities [1]. The White House is considering measures to restrict Chinese AI models, yet advisors remain split on the appropriate policy approach [1]. Some officials advocate for aggressive limitations, while others oppose government intervention in the sector [1].
The disagreement reflects broader tensions within the administration's AI leadership. David Sacks characterized Anthropic's model as having been "lobotomized" and accused it of being "woke" [1], while Emil Michael used harsh language to describe OpenAI's strategy head, calling the person "the highest-level village idiot" [1]. These public criticisms underscore the contentious debate over how different AI companies are developing their systems [1].
In a separate development, Anthropic reached a copyright settlement worth $1.5 billion—the largest copyright payout on record [1]. Meanwhile, Chris Fall, Trump's AI safety officer, resigned after just three months in the position [1].
