String theory? Read Smolin, who points out that for decades, string theorists dominated physics, despite the absence of experimental evidence.

Expert systems? I saw this back in the mid 1980s at Stanford, when the CS department had gone all-in on expert systems. Claims were made that strong AI was just around the corner. This from a technology that doesn't really do all that much. You write simple statements about something in a notation based on predicate calculus, after which you can ask questions. You get out pretty much what you put in. Systems like that still exist, mostly as chatbots for answering the easy questions so customer support doesn't have to talk to you. A programmed FAQ, basically.

And those are hard sciences. Economics is hopeless. Psychology is worse.

(I made the unpopular comment at the time that expert systems would turn out to be more important than syntax-directed parsing, but less important than relational databases.)

As Fred Hoyle once said, "Science is prediction, not explanation". The parts of science that work lead to engineering that works and products that work. Belief is then unnecessary.