Maybe it's different in other fields, but from my background in physics it seems like if your hypothesis is wrong that is usually way more interesting than it being right. As long as it isn't just because of some contamination in the data.
Although, contrary to what I was taught in elementary school, most of the experiments in the physics department of my university didn't even really have a hypothesis. They were usually either of the form "we are going to do this thing, and see what happens", or "we're going to measure this thing more accurately than anyone before".
So I'm not a "scientist" really, I'm just a bit more of a theory-focused software engineer.
An example has been times where I really want to use a certain concurrency style, and I'm convinced that it should be faster than the way we're doing things before, so I will write a few non-trivial tests to make sure that's right, and I'll get inconclusive numbers.
That's still an interesting result. If you thought it would have a significant impact, and it doesn't, that means something is wrong with your mental model that lead you to think it would be faster. And if you can figure out why, that can help you find a solution that is faster.
Of course, that is still very frustrating if you are on a tight deadline and all you have is one thing you know doesn't work.
No doubt that it's interesting, but when there's deadlines that I need to meet "Throughput level X", and it's also infuriating when the theory isn't lining up with reality.
Although, contrary to what I was taught in elementary school, most of the experiments in the physics department of my university didn't even really have a hypothesis. They were usually either of the form "we are going to do this thing, and see what happens", or "we're going to measure this thing more accurately than anyone before".