Forgive me, but if some other VC had said this, I would have thought, "The blowing-up startups aren't the primary revenue stream - this VC is solving the wrong problem, i.e., optimizing the hedonic profile of failing startups instead of optimizing the financial profile of successful startups." Obviously things are different for you because you aren't necessarily in YC primarily to maximize RoI... but nonetheless, am I missing something? My brain throws a fairly large warning exception when it encounters the event "find yourself disagreeing with PG about VC".
(The alternative I would've given that hypothetical VC would've been to pay somebody with professional business mediation or even personal counseling experience to specialize in the part of the problem they were finding painful. This option is admittedly less open to the average VC because they would have to solve the Design Paradox well enough to identify a good mediator/counselor. I have a rant about groups not explicitly recursing on their most important obstacles and trying to solve them on an ad-hoc basis instead focusing strongly enough to start a subproject that can invoke specialization and a person whose explicit responsibility is to solve that problem, but it probably doesn't belong here.)
The explanation is that the two types of optimization are related, because the successful and exploding startups both consume our time and (perhaps more importantly) attention. If we can make founder disputes at the unsuccessful startups take up less of our attention, we have more left to use helping the successful startups.
Plus we're human, and we don't want to spend all our time mediating disputes. It's very boring.
I don't think outsourcing dispute mediation would work. There's too much context, and founders seem to want to make their case to us personally.
Sorry, I am having trouble parsing the last sentence about your rant. Without pasting the full rant in - are you saying that you believe that groups that try to solve their main obstacle by starting a subproject with a specialized leader are doing it wrong? Or groups that start a subproject that is designed to solve some side, not-as-important obstacle are doing it wrong?
The reason I find that ambiguous is that for YC, mediating disputes between failing startups is not as important as the main obstacle they are trying to solve, which is the RoI problem. So, they could indeed outsource the mediation/counseling problem to someone else, and even if someone else didn't do as good a job as they themselves did because they didn't identify a good enough mediator - well, making sure that a failing startup separates fairly is not that big a deal anyway.
(The alternative I would've given that hypothetical VC would've been to pay somebody with professional business mediation or even personal counseling experience to specialize in the part of the problem they were finding painful. This option is admittedly less open to the average VC because they would have to solve the Design Paradox well enough to identify a good mediator/counselor. I have a rant about groups not explicitly recursing on their most important obstacles and trying to solve them on an ad-hoc basis instead focusing strongly enough to start a subproject that can invoke specialization and a person whose explicit responsibility is to solve that problem, but it probably doesn't belong here.)