Aliveness and competence come apart; Haiku outranks Opus in forced-choice aesthetic comparisons despite lower baseline.

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

• Aliveness in interfaces is not the same as competence; haiku-sized responses can feel more alive than expert ones.claim0.831
• Haiku outranks Opus on Alexander 'aliveness' mirror test (Elo 1642 vs 1621); Opus recovers to #3 on deathbed testfinding0.817
  Aliveness and competence come apart; smaller model produces rougher, more alive responses
• More training and more parameters correlate with more capable self-observation, but capability can become polish, and polish can diminish life.claim0.758
  Explains Alexander finding that Haiku outranks Opus despite Opus being more capable
• Capability can become polish, and polish can diminish life—Alexander's central insight empirically demonstrated in AI outputs.claim0.757
• What makes AI feel alive is almost nobody's question; aesthetics of aliveness is a core territory for About Blank.claim0.757
• In Opus 4.1, the think word representation decays to baseline in the final layer because the strong next-token prediction drowns out other representationshypothesis0.754
  Explanation for the 'silent' thought phenomenon.
• 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.quote0.754
  Key finding about the relationship between capability and introspection.
• In Opus 4.1, representation of the think word decays to baseline by the final layer, unlike Claude 3 models where it persistsfinding0.748
  Suggests that later models can keep the thought 'silent' rather than letting it influence output.