I have been using LLMs to help me turn my journals into interconnected notes and sometimes it is so confusing to read the notes that it doesn't resemble any human would write. Its like the models are getting stronger while also losing its touch to write human sounding sentences on complex topics.