There's now a memory feature in Chat GPT that lets you use info from past chats for more personalized replies. Now, you can finally stop repeating yourself.
However, there should be an easier way to select parts of chat, maybe tags. Because sometimes you ask queries that are not for you. That could mess up the context. As AI tools are used for so many different tasks, it is actually detrimental. The dream would be to have a sort of drop-down option to categorize memory into different types.
Advertisement
Work From Home as a Live Chat Assistant
This is a great feature in Chat GPT, but there are dozens of 'in-memory' MCP servers. While not likely as seamless or slick, I'd much rather have interoperability and more control. The next improvement that will be really useful will be memory zones. Different memory zones for different projects.
Chat GPT's Memory Feature:
I have many questions. Does it work across models? Meaning, if I selected one model, asked a question, and then went back to another model. Does that model know about the other conversation? Will this finally affect my CustomGPTs? It's a great thing, but without it finally working for my CustomGPTs, the joy is only half as great.
A lot of us make so many random queries about completely unrelated topics that it might just end up polluting the responses we get with noise.
It would be nice if you could have a work setting that refers to and remembers work-related stuff, and you could have a personal setting or individual user settings. So, if other family members are using it, they’re not changing their behavior.
Comments