Design-wise, storing the entire content in the blockchain is not that difficult. They could keep in the blockchain just the references for the actual posts and a bit of meta-data (revision number and delta-modifications, sort of like a versioning system in which you store just the differences between various versions) and the actual files in a distributed filesystem. Even beyond the current design, platforms like NXT/Ardor could allow spin-offs in child blockchains. I don't see the technical implementation as a bottleneck.

The actual bottleneck might be initial vision, where founders envisioned something reddit-like. I've read the white paper and I don't remember the word "clone", but there were many references and probably reddit was a big inspiration. But as the community grows and the direction of Steemit becomes clearer, they may realize that they can't build a Reddit clone, because, you know, Reddit is already there :).

The tradeoff is between of "just another Reddit-clone", and the "first ever Steemit". That's what I wanted to say by "opportunity cost". They do have a big opportunity with this project.