OpenAI is now rolling out ChatGPT's memory improvements to free users as well. On Tuesday, the company announced that all users will now get personalized answers based on their previous conversations, although free users will only get a light version of it. Also, OpenAI's coding AI agent Codex will now be able to fully connect to the Internet, which can be allowed in the setup phase.

ChatGPT's memory feature is now for everyone.

OpenAI informed about the expansion of this new feature on its official X (formerly Twitter) account. This memory feature was first introduced in April for ChatGPT Plus, Team, and Pro subscribers. Now it is being brought to free users as well. Earlier ChatGPT was only able to save limited information, which could be manually deleted or changed from the Settings > Memory section. Now the new update allows the chatbot to take context from the user's previous conversation.

This means that now the user will not have to repeat the old things every time while starting a new chat. You can directly continue the conversation and ChatGPT will recognize your previous conversation. Example: “Based on our discussion on last quarter’s report, tell us what the goals should be for the next quarter.” ChatGPT will now be able to understand such instructions better, although free users will get a light version of it. OpenAI did not clarify what the difference would be between this and the paid version.

Codex gets internet access.
OpenAI CEO Sam Altman informed on X that Codex is now being given access to the entire internet, although this feature will be off by default and will have to be turned on at the time of setup. OpenAI has warned that turning on internet access can lead to some security risks such as misleading instructions to AI, leaking of sensitive code, and creation of code containing malware.

Disclaimer: This content has been sourced and edited from Amar Ujala. While we have made modifications for clarity and presentation, the original content belongs to its respective authors and website. We do not claim ownership of the content.