Imagine an AI that could understand complex medical reports as easily as your doctor. That's the promise of large language models (LLMs) like GPT, which have shown impressive abilities in various fields. But how do they fare when faced with the intricate world of biomedical text? Researchers put GPT to the test, evaluating its performance on four challenging medical reading comprehension benchmarks. These benchmarks included various question types, from multiple-choice quizzes about biological processes to filling in missing information in medical abstracts. The team explored different prompting techniques, including a novel method called "Implicit Retrieval Augmented Generation" (Implicit RAG). This technique helps GPT focus on the most relevant parts of a text, similar to how a doctor might quickly scan a report for key information. The results were remarkable. GPT, even without prior training on these specific medical datasets (zero-shot setting), outperformed existing AI models on two of the benchmarks, setting new state-of-the-art records. On the other two, it achieved near-perfect scores. One key finding was that Implicit RAG proved particularly effective. By guiding GPT to pinpoint crucial information within long medical texts, this method boosted its comprehension and accuracy. While machine evaluation metrics showed GPT's prowess, the researchers also sought human expert feedback. They found that human experts generally agreed with GPT's generated answers, further validating its potential in real-world medical applications. This research suggests that LLMs like GPT could revolutionize how we interact with medical information. Imagine patients being able to quickly grasp the essence of their medical reports or doctors using AI to efficiently extract key details from lengthy documents. However, challenges remain. The cost of running large-scale experiments with GPT is a limiting factor, and the subjective nature of some medical answers makes evaluation complex. Despite these hurdles, the study highlights the exciting potential of AI to transform medical understanding, paving the way for more accessible and efficient healthcare in the future.