How far I want to push it and how far I should push quality are often very different things.
That said, it will often be easier to strive for higher quality by inspiring the team and getting involved. I'd see this more as taking on the lead developer role (which is certainly not mutually exclusive to the architect role) - but someone definitely ought to be doing this.
The situation I described in my "first experiences" entry refers to an offshore development team - a team which hurls code over the wall for review close to a deadline. Ideal? Not at all! Getting involved would no doubt remedy a lot of this but would naturally raise the cost of development considerably and would probably be contrary to a rather fine balancing act between various external companies.
I'd agree with Richard; inspiration is often one of the most effective means of mentoring. Every so often, though, you meet a team member or, as I've found, an entire team that is just not inspired or motivated by the same things you are. No doubt there are then other more effective approaches to try...
