The “many worlds interpretation of quantum mechanics” is loosely that when you do a quantum coin flip, the universe splits into two universes, one for each result.

The reason for this thought is when you work with quantum mechanics, your system has a state that evolves smoothly, but if you “measure” it, the state suddenly snaps to (a random) one of the possible measurement values (when the coin isn’t being observed, it smoothly evolves, but once you measure it, it suddenly takes on a random value). However, if you expand your quantum description of the system to include your measurement device as well as the quantum “coin”, that sudden “snapping” goes away. Instead your whole system smoothly evolves, and it evolves into a “superposition” of the shared state of the state of the overall system in each of the possible measurement outcomes.

Extending this idea, it would seem that whenever you could describe a situation that acts like a “quantum coin flip”, both results happen, and the universe “splits”.

I really want to emphasize that the practical meaning of these “other worlds” is just that things are a lot “fuzzier” when you zoom in than classical statistics would suggest. Not that there’s another universe where you stayed with your ex or took a different career path or whatever.

Also this is an “interpretation” of quantum mechanics for good reason. It doesn’t really have any physical implications. In particular, it’s not possible to go “interact with” those “other universes”.

Most importantly, there are other “interpretations” of quantum mechanics, like that quantum mechanics is really a rethinking of statistics not of physics.