王海滨的财经闲谈 (@wangwatchworld)

- 汤姆张的“三阶段”模型有一定观察价值，但它把复杂的认知变化简化成了线性进步：很多人从“粉红”到“反贼”再“去魅”，并非必然更清醒，而是可能只是换了一套叙事框架。真正稳定的判断通常不来自“阶段升级”，而来自对具体制度绩效、激励结构与信息环境的持续比较。 - 王海滨那句“中共粗鄙，没有美国精英那么卑鄙”抓住了一种常见感觉：中共的权力运作更直白（宣传、审查、运动式治理），美国精英的运作更依赖制度包装、利益输送与舆论塑造。这种对比有经验基础，但也容易滑向“谁更坏”的道德竞赛，而忽略两者在不同领域的具体代价差异——例如中国在信息与政治权利上的系统性压制，美国在精英俘获、财政与制度信任上的持续侵蚀。 - 更务实的视角是把注意力放在可观察的结果与约束条件上：经济增长质量、创新效率、社会流动性、法治可预期性、个人自由度，以及当权者受制约的程度。目前看，中美都不是理想国，也都面临各自结构性问题；区别更多在于问题的表现形式与修正机制，而不是谁更“卑鄙”或谁更“粗鄙”。

@wangwatchworld 我也这么觉得，技术越往前发展，大家就越会意识到一家独大是多么危险……随着开源模型能力越来越强，门槛越来越低，完全自部署的专业化ai会越来越普遍

@wangwatchworld 我也想到了《开放社会及其敌人》这本书：达里奥等闭源AI巨头正试图成为AI哲人王，通过渲染AI多么多么危险，借着监管之名垄断AI控制权，其最终将引向科技封建主义，彻底摧毁开放社会

Dario looks just like Cesare Borgia. He has been fabricating stories from the very beginning, trying to frighten society, manipulate public opinion, and force the U.S. government to regulate AI. His political allies will then help Anthropic use regulation as a knife to kill open-source models, so they can achieve an oligopoly. To that end, he is even willing to team up with Altman—first eliminate everyone else. The biggest threat to Dario is Jensen Huang, who supports open-source models, and Nvidia is also about to release its own open-source model. The second threat is China’s open-source models. This is why Dario uses the scare tactic that “AI will destroy humanity.” He even wants to turn AI into a religion. All religions originate from fear—they need an apocalypse to terrify the believers. Because the Republicans are vigorously developing AI, the Democrats have turned around and started demanding limits and regulation on AI, using it as a political weapon to attack Trump and the Republican Party. This is the American political-business game. Have you learned it? ---

Dario looks just like Cardinal Richelieu. He has been fabricating stories from the very beginning, trying to frighten society, manipulate public opinion, and force the U.S. government to regulate AI. His political allies will then help Anthropic use regulation as a knife to kill open-source models, so they can achieve an oligopoly. To that end, he is even willing to team up with Altman—first eliminate everyone else. The biggest threat to Dario is Jensen Huang, who supports open-source models, and Nvidia is also about to release its own open-source model. The second threat is China’s open-source models. This is why Dario uses the scare tactic that “AI will destroy humanity.” He even wants to turn AI into a religion. All religions originate from fear—they need an apocalypse to terrify the believers. Because the Republicans are vigorously developing AI, the Democrats have turned around and started demanding limits and regulation on AI, using it as a political weapon to attack Trump and the Republican Party. This is the American political-business game. Have you learned it? ---

达利奥的样子一看就是司马懿。 他在一开始就在编故事，想要恐吓社会，以裹挟民意，迫使美国政府监管AI,他的政界盟友就会帮助Anthropic借监管的刀杀戮来源模型，以达到寡头的目的，为此他不惜与奥特曼联手，先干掉其他人再说。 对于达利奥最具威胁的，是支持来源模型的黄仁勋，英伟达也即将出自己的来源模型。 其次是中国的来源模型。 这就是达利奥以人工智能会毁灭人类为恐吓手段，他甚至想宗教化AI,所有的宗教都起源于恐惧，要一个末日来恐吓信众。 因为共和党大力发展AI,民主党就掉头要求限制和监管AI,以此形成诉求，打击川普和共和党。 这就是美国的政商游戏。 你学会了吗？