https://index.zshipu.com/gpt/tags/AI%E9%9B%86%E6%88%90/ Recent content in AI集成 on GPT资讯 -- 知识铺 Hugo -- gohugo.io zh-CN Mon, 31 Aug 2026 00:00:00 +0000 https://index.zshipu.com/gpt/post/20260831/%E9%A2%84%E7%AE%97500%E9%80%81%E5%A5%B3%E5%8F%8B300%E9%80%81%E7%88%B8%E7%88%B8%E8%93%9D%E8%80%98MaaS%E5%A6%82%E4%BD%95%E5%9C%A8%E9%B8%BF%E8%92%99App%E9%87%8C%E8%A3%85%E4%B8%8A%E5%A4%9A%E8%BD%AE%E5%AF%B9%E8%AF%9D%E9%80%89%E7%A4%BC%E5%A4%A7%E8%84%91/ Mon, 31 Aug 2026 00:00:00 +0000 https://index.zshipu.com/gpt/post/20260831/%E9%A2%84%E7%AE%97500%E9%80%81%E5%A5%B3%E5%8F%8B300%E9%80%81%E7%88%B8%E7%88%B8%E8%93%9D%E8%80%98MaaS%E5%A6%82%E4%BD%95%E5%9C%A8%E9%B8%BF%E8%92%99App%E9%87%8C%E8%A3%85%E4%B8%8A%E5%A4%9A%E8%BD%AE%E5%AF%B9%E8%AF%9D%E9%80%89%E7%A4%BC%E5%A4%A7%E8%84%91/ 预算500送女友、300送爸爸、800送朋友结婚，这些具体数字让礼物App的推荐逻辑失效。蓝耘MaaS正被用来在鸿蒙原生应用里构建多轮对话选礼大脑，直接处理关系、预算和偏好的层层约束。 礼物App的推荐失效源于预算与关系的多重约束 传统礼物App大多依赖标签匹配或热门榜单推荐。当用户