"一句也听不懂，你们能听懂吗？"。 1月26日上午，当理想汽车的员工们被临时通知参加全员会时，许多人期待的是李想能像一位船长，在 经历了2025年的风浪后，清晰地指出2026年的航向——如何应对日益"内卷"的市场，如何反思MEGA失 利后的产品策略，以及最关键的，如何保证大家的年终奖不再打折。 然而，他们等到的是一场近两个小时、几乎完全聚焦于AI的"未来学讲座"。在这场"AI布道"中，李想以 前所未有的激昂姿态，描绘了一幅宏大甚至有些遥远的未来图景。 他断言"2026年是所有希望成为AI头部公司的'上车'最后一年"，并承诺"最晚2028年实现L4级自动驾驶 的落地"，甚至称理想在汽车之外，"一定会做人形机器人，并会尽快落地亮相"。 据一见Auto报道，为了支撑这一宏伟蓝图，理想汽车的研发组织将迎来一场剧变，被重构成"基座模型 团队"、"软件本体团队"和"硬件本体团队"三大支柱，而汽车与机器人，都被归入了"硬件本体"的范 畴。 然而，这位"产品经理"出身的CEO精心准备的未来盛宴，并未赢得满堂喝彩。鞭牛士消息称，在理想汽 车内部的员工社群"理想汽车同事圈"里，吐槽和困惑的情绪迅速蔓延。 这戏剧性的一幕，将一个矛 ...

Core Viewpoint - The article discusses the importance of user feedback in product development at Li Auto, highlighting a case where initial skepticism about user needs was overturned by data-driven insights [2][3]. User Feedback and Product Development - A Li Auto employee and L series car owner identified a strong desire among users to maximize electric usage during high-speed driving, driven by cost savings and smoother performance [2]. - Initial feedback from Li Auto's founder, Li Xiang, dismissed this as a "pseudo-demand," believing most users preferred to use gasoline at high speeds [2][3]. - Subsequent data analysis revealed that approximately one-third of users primarily use gasoline at high speeds, while two-thirds expressed a desire to maximize electric usage, leading to a change in perspective from Li Xiang [3]. Implementation and Future Plans - After recognizing the genuine user demand, the feature aimed at optimizing electric usage during high-speed driving was approved for implementation in the upcoming OTA 8.0 update scheduled for September [3]. - Future plans include integrating large models to tailor charging strategies based on individual user data and preferences, enhancing the overall user experience [3]. Risks and Strategic Focus - There is a potential risk associated with Li Xiang's focus on AI, which may detract from attention to hardware and user experience, potentially impacting short-term sales [4]. - The company values operational efficiency and may hesitate to invest in features unless there is substantial user demand supported by data [3][4].