Vitalik wanted changes that were outside of consensus to make his ideas work; thus he started his own chain. A better discussion would be about changing consensus rules to make mining fees lower by restricting mining fees to only very small bitcoin transactions with minimal flexibility (to make encoding arbitrary data harder). Removing P2SH and requiring disclosure of public key to go with address before including a transaction in the chain would remove a lot of surface area for embedding data. And enforcing a maximum number of outputs of at most 2 per transaction (only change and payer; remember fee is just whatever is left over, not an output) would restrict data embedding even further. But this would very much harm bitcoin utility (no more time decaying multisig) for example!