China's Kimi AI model splits Trump's AI advisors; Anthropic pays record $1.5B copyright settlement
Moonshot's free Kimi model rivals GPT-4, sparking White House infighting, while Anthropic settles...
Moonshot, a Chinese AI company, launched Kimi, a free open-source model that rivals the intelligence of OpenAI and Anthropic's paid models. This has fractured Trump's AI advisory team: David Sacks called Anthropic's models 'lobotomized,' while Emil Michael insulted OpenAI's leadership. The core issue is that every new smart free model from China reduces incentive for US companies to pay for premium AI, creating economic and political headaches for the president.
Separately, Anthropic's $1.5 billion copyright settlement over using pirated works to train Claude has been approved, marking the largest known copyright payout. While the plaintiffs won compensation, many authors and creators argue it does not address the systemic issues of AI training on copyrighted material without consent. The case highlights growing tensions between AI development and intellectual property rights.
- Moonshot's Kimi model matches GPT-4/Claude performance at zero cost, causing White House advisors to publicly clash over whether to restrict Chinese AI.
- Anthropic's $1.5B copyright settlement over pirated training data is the largest in history, but critics say it fails to protect creators' rights.
- Trump's AI safety chief Chris Fall resigned after just three months, signaling instability in federal AI policy.
Why It Matters
The battle over open-source Chinese AI and copyright payouts will shape the future of AI access, regulation, and creative industries.