Yep. In that scenario I'm already assuming that trust depth would more-or-less still work the same way. What I was saying there is that DefaultTrust effectively puts 100 people on your trust list (quasi-directly, I mean; I'm not talking about what the final number would be after accounting for depth). I haven't pondered this deeply (maybe there's an off-by-one error I'm making about where in the recursion the 100 inclusions actually happen), but, my point was that accomplishing a similar level of "coverage" (with a completely-custom trust list) is going to be unrealistic for most users. (But, as I said later, it's also not obvious to me that that amount of coverage is necessary.)

I don't see that as a problem. Remember, its whole purpose is much simpler and more focused than DefaultTrust: it's just something that's meant to prevent new accounts from flying blind. The broader idea is to get everyone (new and old) to start making use of the trust system minus any default component. In fact, to really drive that point home, I'd make it so that you can either have a grayed-out trust list that contains StarterTrust (and nothing else) or have a non-grayed-out trust list that contains a set of manually-picked (excepting StarterTrust) inclusions and exclusions. I'd also make it so that you'd have to move on from StarterTrust at some point (as in, the reminders will eventually force you to curate your own list, even if that means leaving it completely empty). Normally, forcing anything rubs me the wrong way, but on this issue, the pros dominate the cons, I think.