It should be noted that the Bitcoin protocol is more difficult to change, compared to other crypto currencies, simply because it's the most widely used. The more users/implementations a protocol has, the more expensive change will be, simply because it affects more users (more software needs to be rewritten).

Similarly, changing the SQL protocol is orders of magnitude more difficult than changing, say, the GraphQL protocol, but does that mean GraphQL is likely to take over SQL? Or is it just a symptom of the fact that SQL is more popular than GraphQL in the first place?

A crypto currency that no one uses is really easy to change, because no one will be affected. Framing this as a significant competitive advantage, however, misses the point, since the nimbleness is a direct result of no one using it.