I was truly amazed. Usually, retrospective medical studies are a real torture. Constantly making inferences from patient files and stories, not being able to assure their accuracy, and having to spend so much time for so much data. Today, I wrote a code using OpenAI's API and ChatGPT that finds the features I wanted from a made-up story that I generated. It's really interesting because it selected all the information correctly from the entire text. I asked also prompted it to do a risk calculation too, but since I didn't specify which risk to calculate, it came out in a silly manner, but even that was admirable. The formula was related to height, weight, and blood pressure, very amazing intuition. I don't actually know if this type of formula really exists, but next time I'll try to create a child patient and assess it with my own expertise. Despite everything, even with this formula, it still turned out to be a great result. I made a small intervention on the string it returned due to this silly formula, and when I ran the code after removing the '/n' characters and replacing them with paragraph spaces, it automatically converted to a pandas data frame. This is amazing. Big kudos to OpenAI.