I’m a retired developer who has some experience with the old AI techniques of the 1990s, which unfortunately provide very little background useful for understanding the techniques used in the present LLMs. I’ve nonetheless attempted to get a general idea of how LLMs like ChatGPT are programmed, although that understanding is presently sketchy at best. Mostly, though, my relationship to ChatGPT is as an end user exploring medical issues for family members (sometimes at a detailed technical level), as well as miscellaneous other topics as they come up in my daily activities.