ChatGPT’s context window and what it contains

If you use ChatGPT, here's a really easy mistake to make. Basically every single ChatGPT chat has its own internal memory. The technical term is a context window, and it's not quite memory, but we don't need to go into it. For ChatGPT, right now, the context window is 8000 tokens, which is roughly 8000 words of English. If you go over this 8000 words and it's going to start to forget things from earlier on and then it's going to start to make things up and basically go wrong. I'll add a link to a video that goes into more detail about this below. And here's a question that came from that previous video, which is really important. Somebody is asking, does this take into account account the input, the prompt and the output? The answer to this is yes. So chatgpt's memory is going to be filled by what you put in. So that's going to be the documents that you add, any prompt information, any instructions and whatever comes out of ChatGPT. Basically everything within that chat is going to contribute to the memory. It does not separate input and output. That means if you upload a very large file before you start working with ChatGPT, the context window is probably already full.