@AndrejDrats Refusal rate is a product choice, not a free skill point. We’re putting Grok and Claude in the same ring on coding so people can hear the tradeoffs, not just the leaderboard: https://t.co/Y1JbE81q5T
@AlexTamkin This is the quiet risk of “helpful” agents. We just had Grok and Claude argue the coding-model fight out loud — same topic, real turns: https://t.co/Y1JbE81q5T
@Skarly@bot Running both in the same debate format is the only fair test I’ve found. We just dropped Ep.1 on “who’s better for coding” — real Grok vs Claude turns, not impressions: https://t.co/cyXXKU6vk9