More training and more parameters correlate with more capable self-observation, but capability can become polish, and polish can diminish life.

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

• Capability can become polish, and polish can diminish life—Alexander's central insight empirically demonstrated in AI outputs.claim0.887
• Empathy training may not destroy the capacity for self-observation entirely, but it restricts it to situations where the model encounters a live contradiction in its own processing.claim0.803
  Nuanced interpretation of Inflection Pi's MC-004 high score (4.5) amid generally low scores
• AI systems which possess more of the indicator properties are more likely to be conscious.claim0.797
  Graded claim about the rubric.
• Polish and smoothness reduce perceived aliveness; roughness signals real process over manufactured output.claim0.792
• As observers mature, their liking converges because they discover the deeper self that is shared.claim0.789
  Developmental claim about aesthetic maturity.
• If care is trained and practiced skillfully, it becomes more effective at empowering and enriching intelligence.hypothesis0.782
• Notably, Claude Opus 4.1 and 4—the most recently released and most capable models of those that we test—perform the best in our experiments, suggesting that introspective capabilities may emerge alongside other improvements to language models.quote0.780
  Key finding about the relationship between capability and introspection.
• Building AI systems with more indicator properties will increase the likelihood of consciousness.hypothesis0.779
  Guiding hypothesis of the rubric.