I agree with you, but I wasn't commenting on that aspect of it.
I was pointing out that popular protocols are much harder to change than unpopular ones. But claiming that this fact will cause the unpopular protocols to take over the popular ones doesn't make much sense. By that logic the least popular protocol is the most likely to win -- completely backwards, as far as I can see.
A more salient example is the TCP/IP protocol! You cant juat update it, since it's required for everyone to agree on it!