Accessible overview: https://plato.stanford.edu/entries/qm-relational/ In particular, this overview addresses the fundamental difference between Everett's theory and Rovelli's theory.
Original paper: https://arxiv.org/abs/quant-ph/9609002
A fluffy paper by the same author about QFT and its relational nature: https://arxiv.org/abs/hep-th/9910131
In short: The way you can evade the difficulties MWI addresses completely by changing one premise of the problem: that physical systems have states independent of their observers. In relational quantum mechanics, there is no universal wavefunction because there is no external observer of the universe. Call it the zero worlds interpretation. How parsimonious!
There is an element of Yudkowsky's sequence that deals with relative configuration spaces, but this is a subtly different concept and is really just a discussion from a clever guy unequipped with the right concepts about the difference between, say, affine spaces and vector spaces or torsors and groups. The idea that there is a configuration space independent of the observer is the assumption he misses.
In general I find Yudkowsky's extreme certainty in his own arguments to range from amusing to obnoxious. It's not hard to find places where uncertainty creeps in. It doesn't come from mistakes in his reasoning because what he considers he is usually meticulous about considering, but from what he doesn't consider.