This is my first post here.
I suffer from a chronic medical condition that requires me to have blood tests every month and an MRI scan once a year. Naturally, I have been cataloguing diligently all the data since I was diagnosed with the condition about 15 years ago. I have equally designed a very rudimentary spreadsheet to keep an eye on blood results and make sure every single line (individually) is range-bound. And I have recently started uploading the anonymised blood test results to the Chatgpt and Deepseek in search of cross-correlations and causality (unknown unknowns).
All this has been fun until both agents spotted a rise in potassium level (called a mild hyperkalemia) last week and correlated this to an incidental adrenal gland lesion that was picked up on the last MRI scan back in October. This prompted me to investigate.
I care to understand the “stateless” limitation where agents maintain the conversation context for a single chat session only. Is there any easy way to overcome this, where someone like me can upload their data on a monthly basis and train an agent accordingly to keep everything in memory and run their analysis in the context of my full medical history?
Tony