There are a few qualifications here:
1) There are problems that can be practically fixed and there are problems that can't be practically fixed. Sometimes, a problem that isn't practical to fix can become practical depending on the scope of a project. Aurora's client/server architecture and animation systems could certainly be re-written, but I don't think that would fall in the scope of an expansion (for example). The Witcher shows just how much Aurora can be modified given enough time.
2) It's important to remember focus when looking at an engine. While I honestly don't think Aurora's animation/state updating is good for anything (sorry to be blunt, but I can't think of a good reason for it to be engineered as it was), certain other aspects lend themselves better to NWN's intended purpose than the IE's intended purpose. For example, IE game spells were all created in .spl files through an editor called Spell-O-Matic (yes, really). In NWN, most of a spell's functionality came from a script. Creating and debugging spell scripts can honestly be a pain in the *** when compared to the .spl data-driven format, but it allows a lot more flexibility. Also, a 3D engine of any sort makes developing new character/creature content (models, animations, etc.) much, much easier for hobbyists.