Although I support some of Quantus' arguments, Segwit is actually a bad example - because it's a soft fork and there will be no chain claiming it's the "original" one.

A change to a more flexible supply model would always mean that a hard fork is involved - and that would mean that in the moment when the fork occurs, a chain would persist that continues with the 21 million limit. The people supporting this chain will always claim that they are running the original Bitcoin.

But actually the question "who's the original" is a social convention - because with every hard fork the same can occur. So in this aspect, Quantus is right.

Let's take the planned 2MB hard fork in November: I think if there is no agreement between Core and the Segwit2x folks, a minority chain will try to survive with the 1MB cap enabled and will claim that they are the original chain.

But the difference between the 1MB cap change and the 21 million bitcoins cap change is that the 21 million cap was "set in stone" at the beginning of Bitcoin's evolution, like cellard correctly points out, and is considered a fundamental principle. The 1MB cap, in contrast, has even described by Satoshi as a temporary measure (something Big blockers point out every day).

So the support for "21 million forever" is expected to be nearly unanimous (I think >95%) while the 1MB support may be a small majority or not even that. That's why a 2MB hard fork has much more chances to succeed in becoming "the Bitcoin" than a "+21 million" fork.