
Letters are welcome(欢迎) via email to letters@economist.com Find out more about how we process your letter
欢迎通过电子邮件将信件发送至 letters@economist.com。了解更多关于我们如何处理您信件的详情。
Decentralise AI technology
去中心化AI技术
The Trump administration’s blocking of Anthropic’s Fable and Mythos models was more than(不仅仅是) chaotic, it was dystopian (“America’s AI power grab”, June 20th). Just one directive(指令) took a leading AI model offline overnight(一夜之间), which worked because the model has an owner, Anthropic, that could be told what to do. This is the actual problem. Centralisation is a liability(责任,不利条件) for anyone working with AI, and it will happen again whether by this government or another.
特朗普政府封杀Anthropic的Fable和Mythos模型不仅仅是混乱,更是反乌托邦的(《美国AI权力攫取》,6月20日)。仅仅一道指令就让一个领先的AI模型一夜之间下线,这之所以成功是因为该模型有所有者Anthropic,可以对其发号施令。这才是真正的问题。集中化对任何从事AI工作的人来说都是一个隐患,而且这种情况无论由本届政府还是下届政府,都会再次发生。
The solution is to build the only economically viable counter-balance(抗衡力量) to the hyperscalers: decentralised AI. Contrary to what the large AI companies would have you believe, it is possible to match their business models with open, public, decentralised alternatives(替代方案). They just have a vested interest in gaslighting us into thinking otherwise.
解决方案是建立唯一经济可行的、与超大规模公司相抗衡的力量:去中心化AI。与大型AI公司让你相信的相反,用开放、公共、去中心化的替代方案来匹配它们的商业模式是可能的。它们只是在通过心理操控让我们以为不可能,从而维护既得利益。
Island paradise startups
岛国天堂初创企业
Sharing companies’ riches
分享公司财富

He’s knocked them for six
他彻底击败了他们
Shaoxing ex-pats
绍兴海外游子
Speed reading
速读