Lightning fast content discovery.
A blockchain is very different from an SQL or a non-SQL database. There are no built-in indexes. Blocks in the chain contain headers, transactions, and a hash of the previous block and that's it. That's why pulling old posts from the chain is so slow. Blockchains are designed for distributed consensus but not data storages for rapid retrieval or updating. A second-layer solution that puts everything in the chain to centralized database, which stays in sync with the chain and that can, if necessary, be verified against what's on the chain could perhaps partly solve the content discovery problem.
People should also learn to use tags properly to categorize their content.