Memory access changes after hitting the cap on 4o/4

[sysmai]

Since the release of the memory feature, I have used it. Anyone that has used the memory feature in prompting, knows that when you have a decent amount of information about a single topic in memory; that both GPT-4 and 4o can get confused. This is no problem, I just tell it to list everything that it has stored in memory about that topic in the current conversation. After that, it is able to reference the conversation instead of memory and that corrects the issues. Recently, I hit the cap on GPT-4, which normally sends me to the 4o model and that is fine. However, this time it grayed out both 4 and 4o sending me to 4o-mini. This model would not list out my stored memory regardless of what I prompted. It would summarize and tell me it can’t list the memory like 4 and 4o does. It says something about security and memory management. Regardless, once I hit my cap and was forced onto the 4o-mini model. I could not use the GPT as I needed because it kept referencing memory incorrectly and I couldn’t get it to correct itself. Basically, due to the cap, I was blocked from being able to use the features I needed to use until my timer ended. This is a problem for people that make custom GPTs. At the very least, the mini model needs to be able to list out memory so people can make adjustments to the memory itself and not just have the option to delete it through settings.