>but it’s not necessarily reasonable to spend decades building layers and layers of models that assume X at the most basic level.
If the only option for finding out better evidence for or against X is by building those models and watching them either keep matching observations or finding a contradiction that can lead to the downfall of X as the suspect, then it is if you want to progress science any further.
Maybe there is another area that will give results faster, but much of the easy and fast science has already been science. And if someone finds a better option we missed, which does happen from time to time, add some rigor to it, verify it with testing, and they'll likely have themselves a Nobel prize.
If the only option for finding out better evidence for or against X is by building those models and watching them either keep matching observations or finding a contradiction that can lead to the downfall of X as the suspect, then it is if you want to progress science any further.
Maybe there is another area that will give results faster, but much of the easy and fast science has already been science. And if someone finds a better option we missed, which does happen from time to time, add some rigor to it, verify it with testing, and they'll likely have themselves a Nobel prize.