The Case Against Algorithmic Feeds for Long-Form Writing

Imagine if your bookshelf reorganized itself every morning based on what you read most last week. By Friday you would have nothing on it but the book you keep going back to because you cannot get past chapter four.

This is what algorithmic feeds do for writing. They do not optimize for the article that changed your mind; they optimize for the one you bounced off ten times because the headline made you angry.

What I want from a writing platform is the opposite: a slight randomness, a chronological discipline, a chance for the long thing nobody else read to find me. The algorithm cannot give me that, because nobody else read it. The algorithm needs prior signal. The good essay does not have prior signal yet.

So Froodl will have the smallest, most embarrassed feed algorithm we can ship. Sixty per cent chronological from the people you follow, thirty per cent recent activity in topics you follow, ten per cent trending across the platform. No personalization beyond that. No mysterious "for you" page that you cannot turn off.

If you want to discover writers, follow more topics. If you want surprise, hit Explore. If you want a stream, that is what Following is for. The platform should be legible, not magic.