Fractal Feed

Karmyq's feed is personalized to each member's evolved trust calibration. The same pool of requests appears differently to different users — ordered by a blend of skill match, trust distance, community relevance, and urgency, where the trust distance component is calibrated to each user's personal cross-community openness.

How it works

Each user's trust model contains three parameters that evolve based on their interaction patterns:

• Depth weight: how much repeated relationships count in your trust score
• Breadth weight: how much diversity of connections counts
• Cross-community prior: your baseline trust toward people from other communities

When your cross-community prior is high (earned through positive cross-community interactions), you'll naturally see more requests from people outside your immediate community. When it's low, your feed skews toward your established network.

Your evolved parameters

Visit your Trust Score page to see your current calibration. The parameters are read-only — they evolve automatically from your experiences, not from manual configuration. You can pause evolution globally if you prefer your model to stay fixed.

The arc

The fractal feed is the third phase of a three-sprint arc (ADR-046):

1. Individual trust evolution — your personal params learn from your interactions
2. Community evolution — your community's model drifts from aggregate member patterns
3. Fractal feed — the feed uses both your individual calibration and your community's evolved model to rank what you see