Lingguang 2.0 Tests the Super-App Path to Personal Agents
品玩 tested Lingguang 2.0 with a half-month pet-supply price-tracking task and a multi-month household-car expense-tracking project. The agent planned recurring comparisons and reports, but could not directly retrieve ETC and charging-app orders and required screenshots or amounts; its bookkeeping service also encountered an initial backend issue. The report argues that, in China, Personal Agents may emerge through super-app ecosystems such as Alipay rather than by directly copying Muse.
The test provides concrete evidence about Lingguang 2.0’s persistent-task capabilities while showing that stable service connections, user authorization and reliable long-term execution remain unresolved.
Personal Agent在海外已经不再只是一个概念。
在北美已有超过1000万人创建自己的Muse，约四分之一每天使用；购物是它的第一大场景，相关订单的平均金额超过200美元。至少从这些数字看，让AI持续参与个人生活、替用户推进任务，已经是一种成立的产品需求。
但同样的产品形态到了中国，却一直没有真正普及。这件事一度让我们很困惑：中国拥有更发达的移动互联网和更丰富的生活服务，为什么反而没有长出一个被广泛使用的Personal Agent？
问题可能不在AI，而在它要进入的是两套完全不同的互联网。海外的信息源开放但分散，订单、账单、预约和通知大量留在邮箱与网站中；一个Agent接入邮箱、日历和浏览器，就有机会把这些碎片重新串起来。中国的信息和服务看上去更集中，却被分隔在电商App、支付工具、出行平台和一个个小程序里。对Agent来说，前者需要整理碎片，后者首先要穿过服务孤岛。
#Lingguang 2.0#Muse#Alipay#DingTalk#Feishu#Mi Home#Personal Agent