Read news on artificial reasoning with our app.
Read more in the app
New study reveals top AI models (GPT-4o, Claude 3.5, Gemini 2.5) completely fail the classic "Stroop" psychological attention test, exposing a fundamental limitation in artificial reasoning.