> Bringing up user experience when talking about branch predictors is meaningless.
If working on branch predictors does not impact the user experience, then arguably that work is meaningless in the context of building real products -- it may still be interesting for academic purposes.