This doesn't conflict with my post. I doubt he made millions to "run off" with them.
More like the site exploded in popularity and that directly translated into higher upkeep costs both for code and storage. If projects like this do not have a monetization strategy from get go, to at least break even with upkeep, they just die out.
And for the sites which depend on semi-real-time importing it's even worse, since there is no such thing as "write once, run forever" code long-term. The externals and internals of the sites they depend on can and will constantly change and therefore require pretty humongous (or at least well organized) amount of constant work.