This looks really cool! There's a lot here that I agree with. I strongly recommend avoiding the urge to prematurely automate processes involving trust (or gran

This looks really cool! There's a lot here that I agree with. I strongly recommend avoiding the urge to prematurely automate processes involving trust (or granting people additional powers). It would allow you to grow faster, but you don't need to grow fast. Trust signals are useful when they are costly (at least, costly to fake). And if your site can be driven by an API, that changes what is considered costly. Getting a clique of accounts that upvotes the same way you do is not so costly if you're just driving an API. It may be counterintuitive, but... imagine you have some metric that you were considering using as a criterion for granting trust to a user. Let's say you've got the "TrollPercent" metric that shows the percentage of someone's content has been tagged as trolling, or the "KeynesianBeautyVoter" metric that shows how well someone's votes correlate with everyone else's. Then rather than automatically changing that account's status on the basis of these metrics, their account is merely flagged for a human to review. A human moderator can then say "oh, that person's content really is a cut above" or "wait, this person has only ever accessed the site via the API and they've somehow upvoted hundreds of posts in the last 10 minutes..." Anyway, none of this concern is meant to detract from the essential goodness of the idea, and I hope this succeeds.