I’d like to say it is a far, far better video than I have ever seen — the animation and explanation of the light clock, which is the standard explanation of time dilation, are nice. But there is one glaring mistake in it, where it is claimed that time dilation doesn’t happen in an accelerating frame. The GPS satellite constellation would be surprised to hear that, were they not distracted by me anthropomorphizing them.
The key is that time dilation is symmetric in inertial frames, and an acceleration removes that asymmetry. In inertial frames neither twin can say that his measurement is “right” and the other one “wrong,” since there is no absolute reference frame. They both have to be valid measurements. Acceleration removes that symmetry — you can tell if you are accelerating, so you can no longer claim to be at rest — and the clock that accelerates will be slow as compared to one that does not.
Question: Does an orbit constitute an accelerating reference frame? I was told that it doesn’t.
I think you have to be careful of context, because in GR freefall is considered an inertial frame. AFAIK that just changes how you approach the analysis. From the view of SR, though, it’s accelerating and is not inertial. GPS satellites definitely have a kinematic time dilation component, running slow by ~7 microseconds per day.