Thread by @balajis

How would you implement a high-pass filter on Twitter? Some tools include:

- Referrals
- Admission exams
- Staking / deposits
- Paid membership
- Unfollow, mute, block, report
- Moderators
- Subreddits & groups
- Cultural norms
- Machine learning
- Economic alignment

What else?

To explain terminology, in electrical engineering a "high-pass filter" only lets through signals above a certain frequency.

By analogy, a "high-pass filter" on Twitter would be a quality filter. One analogy is how Google returns the top 10 best links, not the full web.

I recognize that Twitter itself has put effort into this problem. The Moments product for interests, the idea of likes/RTs, the follow/unfollow dynamics, and so on.

But just like people gamed search with keyword stuffing, they are gaming social with clickbait outrage.

I thought these graphs from Zuck were also thought provoking.

Right now, there's an incentive to push the envelope right up to the border of getting banned. What if that incentive structure was changed? Both negative & positive outcomes, but interesting.
facebook.com/notes/mark-zuc…

Assume there will be many new social networks as people unbundle Twitter or Facebook, without these central networks fully going away.

What set of incentives leads to the opposite of the (prescient!) comic below? How do we build pro-civilization mechanisms into social networks?