This is good advice. The breakdown really depends on how big of a gap there is between the reality of your product and your "live-looking" demo. Also, the stakes matter here as well. You can end up in Magic Leap territory pretty quickly, and it is telling that many people might not even understand this reference. In general, I totally agree with the OP, especially for a talk. However, the Meta demo likely failed because the technology was simply not fully there yet. Add in thoughtless executives and a marketing team, and you can be doing a live looking demo of something that absolutely does not exist. You will then be ripped apart by the press and your users taking a massive reputational hit.

You offer a clever and practical solution to the live demo dilemma, where excitement and risk often collide. The staged live demo is a great way to balance engagement with reliability, but it raises an interesting question: How can we ensure that this "pseudo-live" approach doesn't erode trust with the audience over time? Will audiences start to feel misled if they realize too many demos are staged, even if the outcome is impressive?