Community Discussion · Policy

After Three Years of 'ChatGPT Moment' Hype, What Should We Believe Now?

YimingJul 222026/07/22 67 views

Last week, when reviewing Q2 progress with the team, our conference room whiteboard was filled with pain points regarding "humanoid robot deployment": costs too high, unclear scenarios, and client budgets cut in half. We just lost a warehouse robot deal; the client said, "Your tech is great, but the ROI doesn't add up. Let's wait two years." I closed the report and remembered what Wang Xingxing said at the Xi'an forum—that the "ChatGPT moment" for embodied intelligence could arrive within two to three years at the fastest. Honestly, I'm both excited and anxious. Excited because this direction is worth betting on; anxious because for a startup, a "two to three year" window is either a lifeline or a noose.

Policy

1 replies

Ctrl + Enter to reply

Deng MingzheJul 22(edited)

[quote="han_yiming, post:1, topic:1372"]

Last week, when reviewing Q2 progress with the team, the whiteboard in the meeting room was filled with pain points for "humanoid robot deployment": costs too high, scenarios unclear, client budgets halved. We just lost a warehouse robot deal; the client said, "Your tech is great, but the ROI doesn't add up, let's talk in two years." I closed the reports and remembered Wang Xingxing's words at the Xi'an forum—that the "ChatGPT moment" for embodied intelligence could arrive within two to three years at the fastest. Honestly, I'm both excited and anxious. Excited because this direction is worth betting on; anxious because the "two to three years" window is either a lifeline or…

[/quote]

You mentioned the technology-driven vs. scenario-driven camps; it's actually similar to our work on AI learning platforms. We stepped into pitfalls: general models performed poorly in actual student usage, and teachers gave feedback that they were worse than subject-specific training tools. A question: regarding the lost warehouse deal, which specific part didn't make the ROI work?