It is how giant publishing houses self-destruct in the gaming space. They fail to realize how difficult it is to build up talented devteams. Everything becomes about maximizing profits in the end. Between the shitty monetization tactics and the terrible working environment they've created, they end up destroying their ability to make good games. I fully expect more mediocrity from Xbox/Activision-Blizzard, if not declining quality.
Open world games need two types of fast travel. The first is your standard type, which is pretty much a teleportation ability. That should be greatly limited. At most, just for cases where you need to travel across the entire map, and should be hidden behind some kind of in-game explanation like "you're taking a boat/plane/subway" or whatever.
The other one should be some way of moving really fast across the map so previously explored areas aren't a chore to move across. Literally fast travel, and not teleportation. And no, conventional solutions like horses or cars is still not fast enough. It's still minutes of mindlessly moving from point A to B in most cases. It needs to be truly fast. Spiderman 2 actually did explore this concept pretty well, with ideas like catapulting yourself or using a wingsuit to glide long distances. Other games need to come up with someway of allow players to cross huge distances in in a few seconds.
Without delving into the question over how good the game is, this sounds like a company that simply has the wrong processes in place. A case of "working hard" instead of "working smart." As a result, they waste a lot of time and resources on things that ultimately don't matter. I'm sure the person in question worked really really hard on the game, but it's mostly pointless and ineffectively effort.
Developer feedback is usually about answering questions that the players have, or finding bugs that were missed in QA. What Bethesda is doing is quite a bit more ridiculous.
Given how massive their game is, I'm doubtful. So much of what they did in the first game will have to be rebuilt. Compared to just reusing most of the original assets and code, this sounds like a lot more work.
Going from their existing RED engine to Unreal is basically the same idea. Almost nothing from the original Cyberpunk game is going to be easily translated to the new platform. I think CDPR just set their development timeline back by at least 3 years.
Starfield heavily leans on procedural generation. It would be many times bigger if it didn't. BG3 has something like 170 hours of recorded dialogue. Cutting down means getting of features.
So no, there is no room for optimization here. These games are just going to be that big, period. People just need to accept that they will have to get giant SSDs in the future.