Noticing some of the people on Twitter with big audiences who onboarded early to Farcaster and churned spending more time here. 

Still not casting a ton, but lurking, liking, etc.

(Not crypto people, either.)

Targeting is my growth problem on every platform.

I am not one content vertical. I’m one-hundred. Random segments of these will surface on any feed-based platform, each a non-sequitur to the last. 

This creates a revolving door where alienation and attraction reach equilibrium and growth stops.

The range of interests following me becomes so wide that I am held hostage and only post the most broadly-appealing or likely-viral content.

It’s not impossible for today’s machine-learning social media recommendation engines to facilitate more cosmopolitan followings, but it’s too resource intensive. These engines want to maximize user behavior KPIs with minimum CPU usage, which forcing followings into single content verticals facilitates.

Humanity is the dog being wagged by the resource-minimizing-machine-learning-recommendation-engine tail. I’m just not that type of dog.

Chocolate rainmaker. Singer. Thinker. Voice

guess on an open protocol like FC, the choice of different algo led content recommendation can be a solution to this. Some algos might require additional spends in warps, $DEGEN etc as only a few people might want to use them ( & they may be resource intensive like you say). While other basic algo promoting virality can be free...