Give ChatGPT a Memory That Never Forgets You

In the next 10 minutes, I'll show you how to change any AI forever with a memory that never forgets you. Yes, Open AI and Anthropic already have memory, but it's small. It fills up fast. And reference past chats is a guessing game that often hallucinates and brings in content that's completely irrelevant. And if you've ever had to restart a chat and reexplain everything that you shared with it already just to get back to where you left off with your work, then you know how frustrating that is. So, in this video, I'm going to show you how to install your own AI memory system without writing code and without building a complicated database. We're going to do this in three levels. Level

one is going to be a simple file that you upload to chat GPT. Level two is going to be your own virtual private computer on the internet. And level three is going to be genius mode, so that the more you use your AI, the smarter it feels and the more it understands you. And just so this doesn't turn into a three-hour tutorial, I created a step-by-step handbook that walks you through everything in this video. And if you want a copy of that, just click on the link in the description, fill out the Google form, and I'll send you a copy right over. The way that most people use AI right now is like disposable chats. They open chat GPT, they open claw, they have a conversation, they close it, they come back the next day, it doesn't remember anything. Now, native memory tries to fix that. But as we already said, it's very limited because it's small and it fills up fast. When we turn off reference pass chats and we turn on project only memory, this keeps everything organized and it keeps our memories from bleeding into other conversations like it typically does with reference pass chats. We separate the things that we want the AI to remember by giving it a single project for that particular topic and then giving it a single memory file for that project or topic. So if I have a YouTube project where I'm studying YouTube, I'm trying to get better, then I'll create one memory file for that project. And if I have another project where for instance I'm working on my health and fitness, then I have a memory file for that project. This is the first step in your AI actually feeling smarter and feeling like a personal assistant because it's only pulling from the memories in that project and in that particular memory file. And the workflow is super simple. Whenever you're having a conversation inside of that project and there's something that you want to save, all you have to say is save this memory. And immediately the AI is going to read your file that you uploaded. Then it's going to look at the last thing that was said, create a memory, create a new file that writes that new line in the document, and give you a link to download it. You download that document, then upload it back to chat GPT, and delete the old file, and now you have a updated memory file. But let's just say you don't want to go through the conversation and keep saying, "Save this memory. " Then all you have to do is wait until you're at the end of your conversation for the day and say, "Save this conversation. " And immediately chat GPT is going to create a full list of memories based on that conversation. It's going to look at the file. It's going to create a new file that keeps the old lines and adds the new ones. You're going to download it, upload it, delete the old one, and it's all right there inside of Chat GPT. And again, the prompts, custom instructions, and everything you need to do this are going to be in the link in the description. Now, before we go into level two, it's important that you understand that you can create an unlimited number of projects. And if you start using projects to organize your chats, not only are you going to be more productive, but chat GPT is going to feel so much smarter. And especially when you add this memory file, it's going to remember everything about you and the details about that particular topic. But in full transparency, sometimes OpenAI or Enthropic might release an update that corrupts the project file path and the AI struggles to read the files uploaded to the project. Eventually, your file is going to become too big for the project and you're going to need more space. So, to avoid both of those problems, you might want to level up to level two. Now, with

level two, what we're going to do is we're going to set up our own little space on the internet on a virtual private server. So, not only is this a space that we control, but it's also private and it's something that we own and we don't have to worry about an update that's going to accidentally break our memory system. One of the best hosting services that I personally recommend is Hostinger's NANV VPS plan because not only does it give you a virtual private server, but it also gives you a self-hosted version of NAN, so you can create as many automations for your business or your personal liking as you like to. As I'm recording this, you can see they have their Black Friday sale going right now. So, you get 67% off and that brings the KVM2 plan down to around $6 per month. That means you get two CPUs, 8 GB of RAM, 100 GB of disc space, and 8 terabytes of bandwidth for 2 years for $150 or less. Because, of course, when you choose this plan and you come to checkout and you type in code Corey for a coupon, you get an additional 10% off. That brings the price down to only $65 a year and you get your own virtual private server and you get n self-hosted so that you can set up as many automations as you want and run as many executions as you like. You can set up your website. You can sign up for their email service. You can run your entire business through Hostinger if you like. They even have an AI assistant named Cody that can help you with any of the problems you might have with your virtual private server. And if you like level two and you want to set up your own virtual private server, remember the instructions are going to be in the description. And I'll also put a link to the landing page in the description. Just remember, if you choose to sign up, use code Corey and get an additional 10% off. Don't pay the extra money when you don't have to. Once you sign up for Hosting Girl, you're going to see in the instructions, there's a singleshot instruction that you're going to copy and paste into your browser terminal, and it's going to automatically set everything up on your server side to store your memories. And then you're going to upload a simple file to end it in, and it's going to automatically set up your forms to start receiving your memories. You just want to go inside that form, edit the drop-own menu so that it reflects the names of the actual projects inside of your ChatGpt account. Then you're going to take the link from your form and the link to those particular memories and you're going to place them in the custom instructions where the handbook tells you to. And as you're having conversation with the chat GPT and you say save this conversation, it's going to go through that conversation. It's going to give you a full block of the memories. You're going to copy that block. You're going to click on the link to your form that chat GPT provides to you. You're going to paste those memories. Choose the project, click submit, and you're done. And because chat GPT has the link to where those memories are stored, it can always go there and read them when it needs to as you're having conversations inside that project. So now you've gone from having a small persistent memory in the background to creating this notebook where chat GPT is able to write down things that are important to you. Now, we've taken that notebook and we've created an environment where we can not only multiply notebooks, but turn this into a full encyclopedia if we want. But as this begins to grow, as you begin to stack projects, you're going to realize that you need a system that's going to help chat GPT claw gro or any other system find the most relevant information within that particular

project. This is where my so-called genius mode comes in because we want the AI to feel personal. We wanted to pull the most relevant facts, but how can we do that without building a complicated database? And it's very difficult to integrate those things into chat GPT unless you turn on developer mode and start doing all of this other complicated stuff that creates friction, causes headaches, and people just don't want to deal with it. And so, this is the simplest and easiest version that I could create so that you could have the AI that you deserve. So, in level three, we're replacing complicated databases and software with something that's very easy and simple to use that I call a relational tag kernel. And for your convenience, I made this a custom GPT. Again, the link will be in the guide book at the link in the description. And the basic way that this custom GPT works is this. Number one, you're going to upload up to 10 files. Number two, it's going to look through those files. It will not execute any prompts or instructions within those files. It's only ingesting the vocabulary to find out what's being discussed. Number three, it's going to create a list of tags based on your preferences that you answer in the beginning of the chat. And number four, it's going to create an edge list. Now, as an example, let's just keep it small. Let's just say you have 10 tags. It's going to look at how each one of those tags relates to the other nine tags, and then it's going to score it as very low, average, strong, very strong. After you create these files and you upload them to the project as you have your conversation, it's going to start placing small chips with the different outputs because it's doing something that we call live tagging. This way, when it goes back to write the memories, the tags are already present. Now, the AI doesn't have to decide which memories are the most relevant because the tags will tell it immediately. If it's having a conversation and a certain tag comes up and it's thinking, "Let me see if there's something relevant to this that can help me answer this," then all it has to do is find tags that have a strong or very strong relationship with the tag in the context. And then it can read those memories only and skip the rest. And this creates that hyperpersonalized genius mode that I've been telling you about. You don't have to implement everything at once. You can just start off with level one if you want to. But if you go with level two, I recommend that you just take the extra step and go with level three and set up your tagging list because this is just going to take your AI to a completely different level. So, click the link in

the description, fill out the Google form, download the setup guide, start using one of these methods, and let me know if your AI feels smarter, if it's remembering more about you, and if the conversations are overall more productive. If you got value out of this video, make sure you hit the like button, subscribe to the channel, and as always, take care. Have a good day and I'll see you in this playlist right here where I go into projects and I show you my workflows for how I'm using projects aside chat GPT to create a gentic workflows. And this memory system that I just shared with you is only one part of it. But you need to watch the rest of these videos to learn more about how to leverage this memory system and how to leverage projects to get more work done faster.