This still does not solve the problem of "portability" where you would like to use multiple models from multiple providers however there is no way to port your memory from ChatGPT to Claude. I wonder if Manually copy pasting this is too cumbersome - or simply because most do not care about memory.
Which is it for you?