ChatGPT Expands Its Memory: OpenAI Opens It Up to Free Users

ChatGPT Enhances Its Memory for More Personalized Interactions

OpenAI recently announced a major advancement in the functioning of ChatGPT, aimed at improving the memory of the virtual assistant. Thanks to a new architecture called Dreaming V3, the tool is now capable of linking past exchanges to the current needs of users, including those using the free version.

The goal of this update is to make ChatGPT more coherent from one conversation to the next. By remembering details such as the user's first name, profession, or even preferences, the assistant avoids unnecessary repetitions and enhances context understanding. This helps reduce those awkward moments when the AI, which is supposed to be personalized, asks for information that has already been provided.

This update relies on an architecture called Dreaming V3, which allows ChatGPT to draw more finely on memories from past conversations. The tool no longer waits for an explicit instruction to remember but can connect elements already shared to tailor its responses.

A Technology Accessible to All

Until now, ChatGPT's memory primarily relied on recorded memories that users could request to keep, as well as the history of discussions. With Dreaming V3, OpenAI has managed to reduce the computational power required to operate this system by about five times. This means that this once-costly technology is becoming accessible to free users.

The rollout of this update is gradual. Plus and Pro subscribers in the United States will benefit first, while Free and Go users, as well as those in other countries, will follow in the coming weeks. Thus, not everyone will see the difference immediately, but access to this feature will soon be expanded.

In practical terms, ChatGPT could better retain a recurring project, a preferred writing style, or a previously mentioned professional constraint. For a heavy user, the gain can be significant: less context to repeat, fewer generic responses.

Personalization Under Control

OpenAI emphasizes giving users control over the data remembered by ChatGPT. It is possible to review, modify, or delete the information retained by the assistant. This flexibility allows users to choose between a more personalized AI or reverting to a more traditional operation.

Although this memory enhancement may make the tool more useful, it also raises questions about the sensitivity of personal data. Preferences, habits, and personal constraints can enrich the AI's responses but require careful monitoring to avoid any intrusion.

The memory of AI assistants is becoming a major issue. Those who succeed in personalizing without being intrusive will have a distinct competitive advantage.