After reading it fully, I've another, deeper problem with this (still great) article : Bayes' theorem totally disappears at the end. Hypothesis that exactly match the observation bitwise have a fixed probability (2^-n where n is their length), and those which are off even by one bit is discarded. There is no updating of probabilities, because hypothesis are always right or wrong. There is no concept left of an hypothesis that'll predict the position of an electron using a probability distribution, and there is no room for an hypothesis like "the coin will fall heads 75% of times, tails 25% of time".

That's both a problem for the article itself (why even speak of Bayes' theorem, if at the end it doesn't matter ?) and to find the "truth" about the universe in a quantum world (even if you accept MWI, what you'll actually observe in the world you end up being in will still be random).

I understand that going into the full details on how to handle fuzzy hypothesis (like, algorithms who don't just output one result, but different results and the probability of each) would make the article even longer, but it would still be a good thing to address those issues somewhere, IMHO.