I think it's likely to be a product of internal incentives and how these create a culture. I would be prepared to wager than it's much easier to get a promotion by working on a new hot thing than on maintaining or improving an existing thing.

Scroll forward a couple of years and you have a reputation for just abandoning existing things rather than improving them, because people individually maximise their own career advancement by migrating off difficult maintenance and improvement projects on to "new hotness" projects only to abandon them again later for an "even newer hotness".

Small organizations have an immune system that rejects bullshit like that but as orgs grow they institutionalize and it becomes harder and harder for them not to succumb.

I think your analysis of DevOps and SRE is spot on.

Most developers make terrible operators. Devops shops also sometimes shift release management and QA tasks onto developers.

Most developers are terrible at those roles too (and most are uninterested in learning them). I’m good (for a developer; mediocre-to-poor vs specialists) at some of the non-developer roles I mentioned, but I hate doing them, and am better at developing software anyway.

Continuous deployment “saves” the system by just throwing something over the wall at customers. This means you always meet the development metrics that get passed up to upper management.

SRE’s optimize for minimizing the number of pages generated by the monitoring system.

All of the positions that acted as customer advocates or considered long term development roadmap (beyond a few sprints) are eliminated in such shops, so even successful products end up being a flash in the pan. At best, they get too popular to shut down, and immediately stagnate.

I’ve seen this play out in multiple organizations from multiple different angles over the years.