“For those asking for details — here’s more context and data 👇” A new benchmark study called WorldTest shows humans outperform every major AI model in real-world reasoning. GPT-4, Claude 3, Gemini 2… all lost to human participants. For y