> Technically every hypothesis that fits our data should get a little weight
This sounds superficially right, but unfortunately there are too infinitely many possibilities to assign more than 0 weight to them just because they're compatible with the data. It's the same fallacy as "the sun either will rise tomorrow, or it won't. Two possibilities mean a 50-50 chance."
A more practical approach will only assign weight to possibilities that might influence your decisions about what to do.
This sounds superficially right, but unfortunately there are too infinitely many possibilities to assign more than 0 weight to them just because they're compatible with the data. It's the same fallacy as "the sun either will rise tomorrow, or it won't. Two possibilities mean a 50-50 chance."
A more practical approach will only assign weight to possibilities that might influence your decisions about what to do.