

Ultimately, a lot of the ideas you’re talking about are just models. Models are useful insofar as they can explain data and make predictions. “Believing” in a model isn’t something that scientists do. We use models as long as they’re useful, update them with new information, and abandon them if they stop being useful.
One important thing about hypotheses and the models they feed into is that they are falsifiable. If there’s nothing that could prove an idea wrong, it’s not really science. String Theory is a good example of this. It’s totally internally consistent but it’s almost perfectly designed such that any measurement that could be made to prove it wrong is completely impossible. To me, that’s not science anymore, it’s some sort of mathematical philosophy.
The “universe is a black hole” idea fits into this not-quite-science frame. Any data that could prove it wrong would only be gathered from outside the observable universe, which is impossible. So, unfortunately, it’s not quite science.





The trouble with string theory is that plenty can be explained given its axioms and what comes out of the resulting math, but those axioms are completely unproven and unprovable and I haven’t seen anything come out of it that have turned into real world results.