The only major Google Maps redesign was from the original boxy low-fi design to the modern pan & zoom gestural design with overlays. The redesigns after that were mostly incremental and rarely affected the general look and feel of the site. There have been a number of design changes recently to unify it with the mobile app and some would argue those weren't always for the better (much like YouTube where it sometimes feels like the UX is getting worse every time).
Google.com always was fairly minimalist to begin with and that never changed much except for additions like the cards added to the mobile site to bring it in line with Chrome for Android. Again, whether these additions are good or not is heavily disputed.
GMail is yet another case of a site that never underwent a significant redesign but merely iterations on the same look and feel.
I can't say anything on CNN or Uber because I don't frequent them sufficiently to have observed any redesign other than Uber's confused rebranding earlier last year.
The way the new design for Reddit is being described reminds me of the Digg redesign. If you don't remember Digg: Digg was essentially pretty much like Reddit, the biggest differentiator was the community. Digg completely redesigned the site because it was failing hard. Digg today is irrelevant and what little community there was has left before the redesign largely been displaced by the redesign (and refocussing).
Changing aspects of a design that are so strongly part of its look and feel (or it's "DNA" if you insist on being poetical) is extremely risky and extremely easy to do badly. It's especially dangerous if you apparently only do it because you want something new/prettier.
HN and Craigslist don't look ancient because of a lack of effort, they look that way because the designs work and there's not much to be gained from changing it.
Any large, sudden redesign is heavily disputed - that much is obvious, and for good reason, since it changes habits. Maps itself goes through continuous redesign, but each modification is "incremental" as you call it. The end result is fairly drastically changed if you look many years back, but well.... I guess it only had incremental redesign, all along the way.
I guess what I'm saying is that if reddit launches one big-bang "completely new reddit" - yeah, I agree, it's very likely to fail. But if they incrementally change it to something else, and backtrack on the bad decisions - few people will notice the radical change, they'll just remember the many small ones.