AI Finds Heart Disease Cure? - I Made LLMs Analyze Medical Studies - Reasoning Test

This video explores the analytical capabilities of large language models (LLMs) in medical research by testing whether they can produce novel insights about heart disease. Watch as Micah Johns challenges four advanced AI models (3.7 Sonnet, Gemini 2.0 Pro Experimental, DeepSeek R1, and ChatGPT o1) to analyze multiple medical studies on heart disease and generate concise, significant realizations that could lead to better treatments or prevention strategies. Each model produces a unique two-sentence response synthesizing information about statins, inflammation, neurohormonal mechanisms, and multi-targeted approaches to cardiovascular disease management. The experiment tests the assumption that language models cannot yet produce meaningful new analysis or findings on complex medical problems.