'We've Hit an Intelligence Overhang': Claire Vo Argues Model Personality, Not Raw Capability, Is Now the Deciding Variable

ChatPRD·medium signal

In the companion write-up to her Opus 5 benchmark, Vo makes the broader claim that raw intelligence has become table stakes and 'the actual experience of working with a model — its personality, its directness, its usability — is the next major frontier.' She contrasts Opus 5's apologetic hedging with GPT-5.6 Sol's blunt 'don't trust me automatically, just use me when I prove valuable,' and reports Opus 5 telling her not to evangelize AI because it might hurt people's feelings. The builder takeaway: model selection is drifting from leaderboard position toward tuning fit for the interaction mode you actually use, and verbosity is now a measurable cost in agent loops, not a style quibble.