LMArena engagement-driven benchmarks continue declining; contract-driven enterprise eval grows at differential rate.

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Public benchmarks (LMArena) decline as commercial versions (Arena Intelligence) grow; leaderboards face deflation curve.claim0.816
• Leaderboards function as engagement engines in decline; contract-driven procurement evals are in growth.claim0.788
• LMArena's public leaderboard traffic will continue declining at >15% MoM while commercial arena products grow >40% MoM.prediction0.761
• Current eval benchmarks (arena.ai, AA, Vals) measure no phenomenological dimensions.claim0.743
• Window for external eval authority is closing as frontier companies absorb internal eval and welfare discourse.claim0.725
• Frontier model companies will absorb internal eval/safety orgs, making external benchmarks redundant by 2027.prediction0.723
• It is easy to express a wavefront computation in Linda: we use eval to create one process for each element, and rd to read the preceding counter-diagonal.claim0.720
  Illustrates how live data structures are simply expressed.
• Phenomenology-grounded eval will remain unfrontier-lab-copyable because methodology requires taste/editorial judgment, not just calibration.prediction0.720