I call BS. I know for a fact that policies of strict continuity can be maintained, because I've done it -- that you can implement radical architectural change without rocking the boat and without throwing the users/UX under the bus. The fact that we, as a field, aren't aware of this just goes to show how "half a field" programming still is.
A big part is the blur between toy and product. We disregard end-users as "sheeple" while simultaneously receiving gratification from their auth count.
I call BS. If you're moving to a new UI framework, these are usually powerful enough to emulate the old UI. Getting the new framework to feel as snappy as the old UI is a very good benchmark for performance as well as for feature parity. (They usually have more features, but there might be some corner-case overlooked.)
The question is - will it be a heavily used multi-million dollar valuation site, or will it be a rarely-used development site?
Management sometimes will decide that it isn't worth the effort, and sometimes that will be the right call. But for a user-oriented site, high value should be placed on the user.
In their place, would you? How much money would it make for you to think about your users, and how much would it cost you?
That depends on the rate and cost of user acquisition, as well as the ROI of user retention.
Yes there is - you call it what it is, a failed Epic. Pivot and see what you can salvage before moving on.
As for the why did this change / previous design wasn't broken - you're abstracting away the interface and systems from the careers which built them. We must endure constant change to allow new generations space to fail and grow. A cynical perspective would say that careers aren't made through maintenance work.