The source of his confusion is believing that all observers must share a single reality. This is not the case: as an observer of event A=a, you only share the same reality as all other observers who also measure A=a (or anything downstream of A=a). If some observer comes along and measures A=b, they split away from your reality. Only the version of that observer that saw A=a stays with you.

There is no "remote synchronization" mechanism between observers. All observations are independent, and when an observation is made, the other outcomes are not discarded, they continue "running in parallel" until another observer comes along. That is to say, from the perspective of other observers, you and your measures are also an observation they have to make (and thereby collapse).