This is super-good – not just on ARGs, which aren't necessarily flavour of the month, but on designing difficult puzzles for a large number of people to solve, and how not to be surprised by how fast groups are at solving things when they have the network. Gating the experience with slow tasks – MD5 brute-forcing, for instance, is one nice idea; I also really like Adam's points about making sure players know precisely what is in-universe and what isn't, so there's never a question of whether something is right or not; just like a good cryptic crossword.