Talking in circles about Moodle being customizable and supported by Moodle Partners? Problem here is its expensive to test and support customizations. Core can make Moodle more configurable, but each configuration option then needs to be tested. That moves the cost from Partners to Core, which is probably right, but then Core needs more testes and documentors, many more, do it right. In that scenario Moodle becomes more like Salesforce, very customizable application. Also very (very) expensive.
Or do you mean that priority be given to good UI/UE? The problem here is more solvable - and arguably a better use of resources - people will forgive a really pretty application that is fun to use for not doing everything - because they enjoy using the features it does have. See Instructure Canvas, for example. Its doubtful Core will bring on an architect and put them in charge of design (like the way great buildings and great commercial software are made, and Moodle will remain developer driven- maybe developers can learn to think like designers: http://kadavy.net/blog/posts/d4h-the-book/.