Anyone can throw out a better-because-of-experience answer, and in fact I’ll do so now: I also have 15 years of experience in front-end dev and I disagree with your premise. It completely depends on what you’re trying to build, and SPAs are too heavily used at the light-interactivity end of the scale. We’re failing our users with brittle solutions that make us feel cleverer as developers.
Well, literally not anyone can "throw" such argument, only those who actually has an experience :) How many years of these 15 you are building SPApps?
It's ok to disagree. I'm in a camp "API + SPA" after my experience. I know as fact that you can greatly increase UI performance this way and build multiple apps (desktop, mobile, web) using single API.
So argument "I just think so" is totally not an argument for me.
I guess the other thing that you’re leaving out is the number of users your site will receive. For smaller numbers in a simpler market you can get away with more technically complex solutions. My last four years have been spent working on data driven platforms with a minimum of a billion individual visits a year, and for those typically prerendering the common parts and pushing them to a caching layer is much more computationally efficient and leads to a faster client.
If you’re dealing with a lot of custom data for each user they that plays less well into the prerender solutions. But again, it comes down to how best to serve your users’ needs.
As with many things, the answer is it depends on what you do. I inherited an app that serves about 1m requests a day, and server-side rendering of the HTML is one of the reasons why we're going to decommission this app, as the performance of the server-side rendering here is atrociously bad/unacceptable. We reimplemented many of the same features in a SPA that does a lot more overall and the performance difference is night and day between the two.
