Two Big Announcements from OpenAI and Nvidia
Today, there are two pretty big announcements from OpenAI and Nvidia. Let’s quickly cover the OpenAI announcement. They’re saying, “We’re testing chat GPT’s ability to remember things you discuss to make future chats more helpful.” This feature is being rolled out to a small portion of free and plus users and it’s easy to turn on or off. I, unfortunately, am not one of those small portion of users, but let’s take a look at the announcements.
OpenAI Announcement
So, the big news from OpenAI is about memory and new controls for Chad GPT. They’re saying, “We’re testing memory with Chad GPT remembering the things you discuss across all chats. It saves you from having to repeat information and makes future conversations more helpful.” You’re in control of Chad GPT’s memory. You can explicitly tell it to remember something, ask it what it remembers, and tell it to forget conversationally or through settings. You can also turn it off entirely. A few people are already getting it, and they have plans to start rolling it out to everyone pretty soon.
How it works is that as you are chatting back and forth with Chad GPT, you can tell it to remember something specific or let it pick up the details itself. Chad GPT’s memory will get better the more you use it, and you’re going to start to notice improvements over time. For example, if you’ve explained that you prefer to have meeting notes with headlines, bullets, and action items summarized at the bottom, Chad GPT will remember that and recap meetings in this way moving forward. Let’s say you’ve told Chad GPT you own a neighborhood coffee shop, when brainstorming messaging for a social post celebrating a new location, Chad GPT will know where to start. If you mentioned that you have a toddler and she loves jellyfish, when you ask Chad GPT to help create her birthday card, it suggests a jellyfish wearing a party hat. If you’re a teacher, it’ll remember information about your students when making lesson plans, etc. And of course, your control. You can switch it off anytime in settings.
Personalization Memory
If you have access to this, you’ll have this new tab called “personalization.” So, if you log into Chad GPT, click on your name in the bottom-left settings, and see if you have that personalization tab. If you wanted to forget something, you can just tell it. Interestingly, you can view and delete specific memories in that same personalization and manage memory. As you can see here, it looks like Chad GPT is remembering some of the things that you’ve told it, whether that’s through things it selects by itself to remember or things you’ve explicitly told it to remember, and you’re able to delete them one at a time.
They’re saying your memories, like the models, will be trained on those to improve the models for everyone. You can turn this off in data controls and, as always, they don’t train on content from Chad GPT teams and Enterprise customers. They also add a temporary chat similar to the incognito window for Chrome browsers, so if you don’t want it using your memories, just use that. Temporary chats won’t appear in history, won’t use memory, and won’t be used to train our models.
Custom instructions can still be used to provide direct guidance. So, for explicit information instructions, you can add it to your custom instructions. These memories seem to be a more fluid ongoing thing that you can change quickly, that will be updated automatically. They’re saying certain sensitive information like your health details won’t be remembered unless you explicitly ask it to. They also explain how team and Enterprise customers will benefit from this feature.
Nvidia Announcement
We also got some big news from Nvidia and Jensen Huang. That’s coming up in just a little bit. Stay tuned! My name is Wes rth, and thank you for watching.
