Often the older people get, the more they relish in using their skills to great effect.
I think they found that their skills and experience were no longer relevant given the constraints they'd set for themselves (budget and design goals). Coding an "evolved" late 1980s blobber (basically what Ultima Underworld was) turns out be different from trying to twist the Unity Engine into the Ultimate Action RPG with Dark Messiah combat, Thief stealth and Arx Fatalis magic (or whatever).
Do you really believe this? Their professional development didn't end with UU, they went on to found LGS and made SS1/2, T1/2 etc. Do you think concepts such as engineering proofs of concepts, functional tests, scoping and prototyping were unknown to them by the time LGS folded? That they didn't have to manage projects and negotiated deliverables with publishers? I really have trouble believing that.
For similar reasons, I have trouble swallowing the narrative you suggested
Doctor Sbaitso. Sure, UU may have started with geeks at MIT, but they went on to found a company and publish games into the year 2000. They surely had to do the above before making concrete statements about what will or will not be in a game, and even if
those changed, no game released by LGS (that I'm aware of - but I'm certainly not a LGS aficionado) was in as remotely as bad a state as UA.
This is why I expect there were people involved in the production of games at LGS that were critical, but did not work with Otherside. As with a lot of collective creative efforts, names get attached to projects often unfairly, because it simplifies a media narrative. And while people love reading and digesting nice little stories about visionary X or Y propelling something to greatness, the true picture is often a lot more complex and relies on a lot of people that go relatively unnoticed.