I think a middle ground would be that it's fine to start with a belief about how something might work (also known as a hypothesis), after all that's the first s

I think a middle ground would be that it's fine to start with a belief about how something might work (also known as a hypothesis), after all that's the first step in the Scientific Method. But crafting a hypothesis without a subsequent experiment to prove or disprove it is just wasted breath. As the adage goes, claims that are made without evidence can be dismissed without evidence.