Here's what's fascinating about this topic of scope creep:
When building personnel carriers (or tanks) you're making a physical object. So the decisions you make on what is in or what is out have huge repercussions. That's why you keep seeing these plans being printed to validate the look of the object. But what you did not see are any of the trade offs communicated in any real way.
When working with digital, the handcuffs of creating physical items are unshackled. There are practically an infinite amount of ways to deal with a request of a customer. But it is critical to set out some objectives up front about what you'd like it to do and then check back when the scope creeps. Scope creep on its own is a psychological way to validate that what you're doing is "getting better." It's not a bad thing in and of itself.
It's only bad when it goes off target of your objectives. When you wanted to hold 11 soldiers and now you're holding 6 (almost half!) then are you achieving your target? What if you ended up holding 10? That seems within reason. What's the least number of men you want to carry before it's not a speedy personnel carrier but a tank? How do you determine this number? That magic of working towards business objectives and collaborating around whether those objectives are being met is the stuff of great project managers (not just good ones). The colonel in the video starts out well, but he's never really identified what the objectives were and never re-validated them at all. He never asked "forget about the 'look and feel' and 'munitions', but rather how many personnel do you want this thing to carry?" Getting buy in on the number would create a pretty clear scope and be able to easily loop back when that's not being met.
Constraints are actually quite beautiful. Constraints eat scope creep for breakfast.