The fact that encrypted data is not a total mystery to Mega is the most troubling issue. On one hand, the reason behind implementing a block-based data deduplication scheme is obvious: storage is cheap, but it's not that cheap, and the distributed infrastructure providers supplying storage to Mega don't have to waste space storing non-unique data—instead of 10,000 copies of The Hobbit, the service would only store a single copy, freeing up terabytes of space (though the scale and scope of the deduplication isn't known yet, so this may be optimistic). On the other hand, even if the service doesn't know those blocks of data happen to be The Hobbit, the service does know which users own those deduplicated blocks, and if one user is implicated, there's proof against all the others, too.