I agree that webs of trust are the way forward.

Power corrupts, and any set of curators will eventually fall to corruption, but the nice thing about having software that lets you explicitly trust people (instead of implicitly trusting the platform) is that the switching costs are low. You don't have to wait for somebody who is both tech savvy enough to build a competing platform and also trustworthy enough to curate the content, you just need the trustworthy part. You just twiddle your trust settings and you've got a fresh start.

I also think that the who-trusts-whom-in-what-category dataset would be incredibly useful. For instance, you can follow the edges until you get a cycle, and that cycle is going to be at the heart of a community of experts.

Or if you're having trouble seeing eye to eye with someone you can consult the trust graph and find a mediator, or if they're not to be found then you'll find thought leaders who exemplify the disagreement.

You'll have to be hygenic about who you trust though, and that's not the lazy path forward. So culturally I think any project of this sort is going to face an uphill battle.