ChatGPT's memory feature now available to free users: What it means for users

OpenAI has officially launched a significant upgrade to its memory feature for free users of ChatGPT, a capability that was previously reserved for Pro plan subscribers. This enhancement empowers the AI chatbot to deliver more personalized interactions by recalling user preferences and past discussions. Initially announced in April, the rollout is now reaching a wider audience, allowing ChatGPT to better assist with tasks such as writing, learning, and offering general advice by referencing recent chats. OpenAI shared the news on X, mentioning, 'We’re starting to roll out a lightweight version of memory improvements to Free users. In addition to existing saved memories, ChatGPT now references your recent conversations to provide more personalized responses.' Users can easily assess how much the chatbot remembers by asking it to summarize their interactions with the prompt, 'Describe me based on our chats.' Although the model may not retain information from several months back, it can remember details from more recent exchanges. For those who wish to opt-out of the memory feature, OpenAI provides options to disable it. Users can navigate to Settings > Personalization > Memory to fully turn off memory or manage previously stored data. In conjunction with this upgrade, OpenAI has also broadened access to its Codex coding assistant. Once limited to Enterprise, Team, or Pro subscribers, it is now available to ChatGPT Plus users. Codex aids developers in writing, debugging, and testing code, and a noteworthy addition in this update is its ability to access the internet while coding. This feature allows the assistant to install packages, interact with staging environments, and gather necessary resources for specific tests. However, internet access is turned off by default and must be activated manually. Users retain control over which domains Codex can access and the HTTP methods it may employ. OpenAI is closely monitoring the system for potential misuse, including 'prompt injection attacks,' to ensure the coding assistant functions securely and effectively.