Source: The Next Web
Marc Andreessen's assertion that ChatGPT outperforms 99% of doctors reflects a familiar pattern among tech investors: extrapolating from narrow benchmarks to sweeping capability claims without accounting for the messiness of actual clinical practice. The gap between passing standardized exams and handling real-world diagnostic complexity—where patient history, physical examination, and contextual judgment matter—remains wide, and promoting otherwise risks undermining trust in both AI and human medicine when integration, not replacement, is the only realistic near-term path.