I agree. Against this, I like Williamson's notion of 'knowledge first' - knowledge as foundational and in a separate zone from belief. But I've read Williamson, and even been to a little seminar run by him, and he's the nicest bloke - but every tiny possibility has to be explored by him too, footnote after footnote, and then there's the argument by Sproggins (2014) although Hackface (2015) would disagree...all that! I am a bit of a nit-picker by nature, I think that's why I enjoy the analytic approach mostly, but sometimes you've just got to see the bigger picture or you'll get awfully lost. — mcdoodle

Off topic:

I had greatly enjoyed Williamson's Knowledge and its Limits, OUP (2000). Though rather daunting in places, it doesn't suffer as much from the excessive narrowing down in focus that afflicts some later papers by him. Surely, the book format helps. But I am mostly indebted to John McDowell for the 'knowledge first' approach, expounded in a variety of papers including, most relevantly, Criteria, Defeasibility and Knowledge (usefully read in conjunction with John W. Cook, 'Human Beings', in Peter Winch ed., Studies in the Philosophy of Wittgenstein, Routledge & Kegan Paul, (1969), that McDowell is indebted to for the idea of indefeasible criteria). The best elucidation of the big picture afforded by McDowell's own account of the 'knowledge first' approach that I encountered is the second chapter -- 'Belief and the Second Person' -- in Sebastian Rödl, Self-Consciousness, HUP (2007). Lastly, I've bought recently Andrea Kern, Sources of Knowledge: On the Concept of a Rational Capacity for Knowledge, HUP (2017). I haven't read most of it yet, but it looks fantastic!