We of course should not shy away from evolving the Donut incentive model. There is plenty we have learned about what has worked vs not worked and I believe there are some changes we could make to make the model more clear, consistent, and effective.

The overarching aim is to reward contribution. A key challenge therefore is how to identify that contribution. At the moment we rely heavily on Reddit to give us karma metrics which we use to bias weighting. Reddit does not allow any discrimination based on who is voting on content and this, in my opinion, is a major issue. The signal from established members of a community should have a greater weight to identify what is a contribution.

The following suggestions seek to replace Reddit's aggregation, remove failed mechanisms (tip signaling), and extend successful ones (approved users, pay-to-post).

• Remove incentives to signal. This seems to just promote tip farming
• Replace tip signaling with comment-to-vote. For purpose of donut allocation posts would be weighted by the number of comments from approved users (gov weight > 20k). Commenting is easy and accessible on all platforms.
• Only comments above a certain length (100 chars?) would be eligible to earn Donuts.
• Like pay-to-post, to combat farming and spam there is a fee of 10 donuts (deducted from comment earned donuts) for Donut eligible comments
• Approved users (gov weight > 20k) can give more weight to a comment with a reply that includes !glaze
• Current tip based signaling (I believe) accounts for only 10% of the distribution. The new distributions would be entirely based on comment-to-vote and replace the Reddit karma aggregation. Eligible comments and posts could either be treated with equal weight, or changed to something like 80/20 posts/comments. IMO, eligibility from different flairs (ex. COMEDY at 10%) could be removed.
• Signaling for both posts and comments would be analyzed, with the potential for cheaters to lose all their CONTRIB.