Honestly, I'd feel much better if people were standardizing QUIC and we simply run HTTP/1.1 over it.
Instead we now have transport layers that are application specific and 3 completely different web protocols with none of them being considered legacy, 2 of them being complex enough that people aren't very willing to move.
That does not look like a good foundation for anything.
Part of the process while moving QUIC through IETF was separating it from the HTTP layer, which is why now HTTP/3 and QUIC are different things - although given history and involved players, the HTTP use case was a priority, but other companies are looking to use it underneath other protocols. From my understanding, compared to HTTP/2, the stream concept has moved out of the HTTP spec in the underlying level, where it hopefully can benefit other use cases too.
Indeed, and the encryption is split too. Overall it is starting to look nice to use, although I need to read the spec in more detail to understand a bunch of the details.
Instead we now have transport layers that are application specific and 3 completely different web protocols with none of them being considered legacy, 2 of them being complex enough that people aren't very willing to move.
That does not look like a good foundation for anything.