What Would a Media Ecosystem Designed for Attention Recovery Look Like?

This piece emerged from a different kind of signal: not a coverage spike but a coverage absence. Eral tracked 1,200 articles on media criticism, attention economics, and platform design over 90 days. The pattern of what was not being written about — constructive proposals, specific design alternatives, economic models for different attention architectures — was as notable as what was.

What the current system is optimized for

The current advertising-based content economy selects for content that produces high-arousal emotional states (outrage, anxiety, desire, tribalism) because these states correlate with engagement metrics that translate to advertising CPMs. This is not a conspiracy or malice — it is the emergent property of an optimization target. Change the target, change the output.

What different targets might produce

Several alternative optimization targets have been proposed with varying degrees of rigor. Time-well-spent metrics (proposed by Tristan Harris and the Center for Humane Technology) have been partially adopted by some platforms but not in ways that affect core algorithmic reward structures. Subscription models partially decouple revenue from raw engagement but create their own distortions — the need to justify renewal creates different but equally potent content biases toward validation over challenge.

The most structurally promising alternative Eral's source analysis points to is reader-reported satisfaction at the end of a session rather than engagement during it. This changes the measurement point and creates different selection pressures. Publications that have experimented with this — notably The Guardian and some local newspaper cooperatives — report higher trust ratings and lower churn, though the A/B evidence base is thin.

The problem with attention is not that people do not have it. It is that every system that depends on it has an incentive to deplete it.