I seem to recall that the personalization module worked with o1, but I just used my limited access to o1 to work on an HTML project and realized that this version of o1 didn’t recognize any of my personalization entries or past interactions.
Working with a personalized version of ChatGPT has always made my workflow more efficient and enjoyable, so I was surprised to see this functionality absent. I’m not sure if this is a change they made recently or if o1 never supported personalization to begin with.
Does anyone know if this is intentional, or if o1 was always designed without personalization? Any insights would be greatly appreciated!
EDIT: Starting the session with 4o, which does seem to have access to the memories/personalization, and then swapping to o1 seems to have worked, and I was able to work with o1 while having the nice, personalized interactions that make work so much better for me. Although I am still puzzled why starting a session as o1 does not seem to be able to fetch the memories.
I revisited one of my earlier interactions with o1, and it seemed to have access to memory at that time. Additionally, I noticed that starting a session with GPT-4o and then switching to o1 allows it to retain access to memory, which is quite interesting.
I can only hope that the personalization and memory features will be reintegrated into o1 models in the future. For me, these features make a world of difference when collaborating with ChatGPT. Familiarity truly fosters creativity in my workflow, and having that continuity is invaluable.
Either way, I really appreciate your insight—thank you!