U.S. federal court dismisses arbitration request for WLFI; Sun Yuchen's lawsuit regarding token control rights will continue to be heard in public.

After large models become agents, the training challenge is no longer just about getting them to answer a question correctly, but about enabling them to learn to observe, trial and error, use tools, and adjust their strategies based on the results in a series of continuous actions. A paper published on August 20th, titled "_EnvHarness: Awakening Static Worlds for Agent Learning", focuses on a often underestimated aspect: many existing environments are only suitable for one-time evaluations and lack the feedback, course difficulty, and controllable perturbations required for training. The approach proposed by the research team, EnvHarness, is not about creating another massive virtual world, but rather about making adjustments without altering the underlying logic of the environment.

Reuters/Ipsos poll shows that most Americans believe it is inappropriate for the Trump family to profit from crypto businesses, and related controversies are affecting discussions on crypto legislation in the United States.

Fidelity argues that the online activity brought about by AI agents may not necessarily translate into value for the public chain and tokens. Closed platforms, stablecoin payments, and regulatory factors could all lead to a diversion of profits.

Wall Street Journal article discusses the world's largest humanoid robots ETF, Yushu IPO, and the configuration of the robotics industry chain, with a focus on the robotics supply chains and commercialization in China and the United States.