Were all invented long ago and were all first of their kind. Creating a new standard when there are no wide spread alternatives is easy, doing the same where everyone is already invested is hard. E.g. that's why payment systems suck universally.
As someone who does a fair bit of work in networking fields, I think that's a pessimistic view. Most people don't see when the underlying infrastructure develops, in no small part because the standards and compatibility issues are so carefully considered, but that doesn't mean there aren't newer standards and protocols being developed all the time. We're now on something like the fifth mainstream WiFi standard, for example, and while someone buying a new laptop of getting a new box from their ISP might not know what all the 802.whatever markings mean, they still experience much faster speed and higher reliability compared to the earlier technology. An example from much further up the stack is that we're starting to see wider support for Web serving using HTTP/2, which is a big change from its predecessor.
Even with payment systems, we've seen multiple contactless payment technologies become established very rapidly in recent years, and developments like Chip-and-PIN cards a few years before that. Of course online payment processing is also a much more developed and competitive industry today than it was even five years ago, which again is partly because both the technical and the regulatory frameworks have opened up in recent years. SEPA in Europe is a good example here.
I like optimistic people (no sarcasm). But I think you're overestimating the cooperative abilities of Homo Sapiens. Every bank, store and pet garage invents their mobile payments nowadays. This is very bad and leads to segmentation of efforts which shouldn't exist in the first place.
Also, I'd argue HTTP/2 is not such a huge improvement as many make it out to be, but I can't deny it's some improvement compared to 1.0/1.1 -- that's a fact.
I don't know about HTTP, but the first three were definitely not first of their kind, particularly when you take into account how they operate now compared to their original incarnations.
GSSAPI and Kerberos, for example, both predate SSL (Kerberos by nearly a decade if I have my dates right). SMTP was originally intended only to transit mail protocols across networks, clearly evidencing that there were internal (incompatible) mail protocols before then. UUCP and FTP were commonly used before SMTP to transfer messages; it took over a decade after its invention for SMTP to finally see off UUCP, and a few more years for X.400 (invented more or less concurrently) to fade away as a potential competitor.
I think we can easily replace "were the first of their kind" with "were invented in times when there was a desperate need of a good protocol to do X, Y and Z" and you and the parent poster would be in a full agreement, don't you think?
I think his general point was -- even if there were some ugly corners of the technology, people were like "you know what, this is the N-th try and we really REALLY need this tech, let's move forward with it and fix the problem later". I think we all know how that usually ends, don't we?
It ends with a lot of legacy baggage and huge economic pressure to not change anything. So we come to our present dilemmas (outlined in the original post).
Were all invented long ago and were all first of their kind. Creating a new standard when there are no wide spread alternatives is easy, doing the same where everyone is already invested is hard. E.g. that's why payment systems suck universally.