Large language models (LLMs), like GPT4o, are being explored to enhance dermatological image interpretation by combining human-understandable descriptions with clinical context, mirroring human semantic reasoning. Unlike traditional computer vision that relies on pixel pattern matching, LLMs can dissect an image description and consider relevant medical history to boost diagnostic accuracy. GPT4o achieved over 80% accuracy on a complex dermatological dataset when context was provided, outperforming models like Google Lens which do not employ context. However, these LLMs may be swayed by incorrect information, highlighting the need for advocates like MDandMe to carefully curate and contextualize data. While LLMs can aid in diagnosing minor conditions, they lack the definitive judgment of a dermatologist and are seen as supplementary rather than replacements in medical settings.