当我们什么都能做，还知道自己想做什么吗？When We Can Do Anything, Do We Still Know What We Want? — Ryo Lu, ex-Cursor

Ask an AI to build a journal and it might give you mood scores, streaks, daily reminders, and weekly reports. Which would you keep? Ryo Lu, a former founding designer at Cursor, uses a working journal demo to show how design decisions slip into an AI-generated answer and how a good tool lets you change them. Recorded at AI Engineer Shanghai on November 5, 2026. Choose Settings → Audio track to hear the original Mandarin or the AI-generated English dub, produced with Ryo’s permission. The slides and captions appear in Chinese on the left and English on the right. 点击播放器的“设置”→“音轨”，可切换中文原声和英文配音。画面左侧为中文，右侧为英文。 00:00 Introducing Ryo Lu 00:38 Stronger tools, greater unease 03:04 Before making things became a way to prove ourselves 05:47 The journal AI built for me 07:32 Which feature would you remove? 09:13 A service counter or a workbench? 12:29 Three principles for designing tools 13:07 When to put the tool down 16:20 What a jazz ensemble teaches us about collaboration 19:45 Finding out what you want by making things 21:35 An exercise for the next AI-generated answer Ryo’s post: https://x.com/ryolu_/status/2118207967728451760 Conference program: https://ai.engineer/shanghai/2026/program