I think it is a good thing when you find almost everything interesting. Curiosity that keeps pulling you deeper into a topic is not a bug, but a feature.
You look something up, that leads to a related concept, which depends on something else, and before long, you understand a topic far better but different than you originally set out to.
Rabbit holes get a bad reputation, but the problem is not the depth itself. It is not knowing when to stop and come back up.
Most solve this with timeboxing, I am not a fan of that approach. A timer can let you keep going down a branch that has nothing to do with what you actually needed to know.
What I do instead is pick a depth, not a duration. At some point, I decide that I will not go further than this. Whatever sits below that point, I treat as a fair assumption, a black box I do not need to open. Then I move on.
This works because rabbit holes are not flat. The first few layers are usually directly relevant. Past a certain depth, you are learning about the implementation of the implementation, and that rarely changes how you use the layer above it.
Timeboxing optimizes for how much time you spend. “Depth-boxing” optimizes for understanding what matters.
Next time you catch yourself going down a rabbit hole, try picking a depth instead of a time.