Eignex Blames Feeds for AI Slop
Rasmus Ros argues that AI slop is really a recommender-system problem: feeds optimized for engagement amplify synthetic content, while the model itself is just the supply.
Hot take: this is a solid systems-level critique with the right villain. The post is strongest when it reframes “AI slop” as a ranking and incentive problem, because that explains why the same content can be harmless in one context and toxic in another.
- –It makes a clear distinction between tools that remove friction for a task and feeds that exist to retain attention.
- –The argument is more persuasive than a generic “AI is making us dumb” take because it locates the failure in objective design.
- –The strongest practical implication is that alternative ranking goals, like informational diversity or delayed user endorsement, matter more than content provenance alone.
- –It is opinionated, not product-centric, so the value here is the framing and analysis rather than a launch.
DISCOVERED
45d ago
2026-05-03
PUBLISHED
45d ago
2026-05-03
RELEVANCE
AUTHOR
AWildMonomAppears