Decide on how we want to model filtering issues and repos

By @chshersh:
Though, it's an open question, how do we want to model filtering. For example, when a user enters labels difficulty:easy and difficulty:medium does he want to have issues that have both labels easy and medium or any of the labels? In this case, probably the first. But what if here enters difficulty:easy and refactoring? Maybe now he wants to have these two labels. We can't read the user's mind so we either will stick to one default or support both strategies. Not now, just something to think about.