 Well, that's one of the things like we were talking to somebody that is using chat GPT to create syntax for flows and other stuff. So some of the complexity of Power Automate, which was great, but one of the limitations of chat GPT is that it's based on the date that it was trained on whether it has access to that data. So if you have with a lot of the PowerShell stuff, I don't know how long those would be out there whether it would be able to find it within its libraries. They even talked about that when we were in Redmond, right? They said that there are things that are being developed so fast that the models can't keep up. So they're like you can't expect anything near real time with AI unless you have a model that is completely open in terms of constantly consuming information. And to me, and from what I, you know, it's been all over the news, right? Everybody's talking about it. They went to the front of Congress and stuff like that about it too. But, you know, that is the scary part. Having an AI model that is actually listening in real time and going out and searching in real time. That's scary because that's when things can get messy, you know, with the responses and with, you know, the actions that people take off of those responses. I just, I see a lot of problems with that. So I did try. Good. I did try to use chat GPT to write that script for me to write a report. All I needed from the report was the SharePoint site, the title, and it's lock state. And I had the properties and I wanted to spit it out to Excel. So I asked chat GPT with that. I'm like, yeah, I did this whole script. I plug it in. It did work. Well, so there's PMPs versus SPOs and web versus site and all the stuff that I was taking charge. But you have to know how to create a prompt in chat GPT. Because if you don't prompt it correctly, it doesn't come back with the correct information. You know, when you started out, did you say, I need you to act like a developer? Okay. And I need you to, yeah, I need you to think like this. I need you, you know, and if you go out, I think on GitHub, there's a thing called awesome chat GPT prompts or something like that. And literally list hundreds of these prompts and they're full paragraphs. Just to ask one question is like a full paragraph because you're setting up the actual AI model with all this information. Like, okay, this is what I am. This is what I do. This is the information they want me to get, you know, all this other stuff. And then it comes back with more of a complete answer and a more accurate answer. But yeah, but there's work involved on the front end, you know, on your part to do that. But I think you can get a better, a more accurate result by, by, you know, prompting better. Yeah. Case in point, girl, don't code. Yeah. You were going to say something Christian. Yeah, sorry. I forgot now because I was thoroughly entertained by the answer. You were thoroughly bored by that answer. Checked out. Absolutely. What was, what was the, you said that for the, I've seen a couple sites and I'm more past the question now anyway. So this is the filler in between. But I found a couple sites and I've watched some videos of advice on how to structure your queries in chat GPT and things around there. What was the site that you mentioned? Yeah, prompts. Yeah. What was the site you mentioned? So if you go out to GitHub and you search for awesome chat GPT prompts and there's an entire repo. I found it. Perfect. I'll leave that in my notes there. Yeah. There's a, and he, that is literally updated on a constant basis where people are contributing. That's great prompts that they've actually used. One of the things I saw in there was I need, you know, I need you to act like a on stage presenter. I need you to create an abstract or this type of talk. And I was like, I probably know people that are using that. Yeah. There's some of that stuff. I would love to be able to leverage more, more of, I mean, I really am using it now for, um, for rewording stuff for, um, for like, outlining creation, summarizing like those other things, not for like flat out creation of content. You got to be careful. You got to be careful because I mean, an example is one of the, and this is, I don't know if this is even public yet, but one of the, uh, I want to say sites, um, that creates really quick technical content, um, in little tiny one, three, five minute lengths is being sued. Um, because they were using complete plagiarized information out of chat TPT. Yeah. Or, you know, their videos and for their, their scripts. And, you know, that's, that's a big deal. That's why I like tweets, creates tweets for me, which is great. And for SEO stuff. No, the, the, um, that's why I actually really like Jarvis as the tool, um, because it has with the paid subscription, the plagiarism checker built into it. Yeah. So it'll generate and, but it'll, uh, you know, identify the sources. And so you can actually cut and paste content in and it will identify if there's plagiarism there. Um, which is, I think any teacher now in this world that we're in would be smart to be able to digitally plug any student's paper in and right away identify plagiarism. Grammar. Grammarly now offers that as well. So Grammarly's jumping on that and they're offering smart colleges. That's great. Yeah. Well, somebody even mentioned that they're, they're talking about some kind of metadata around it to be able to identify when AI generated content. I'm like, I don't know how that would be possible to do that. There's no underlying code when you copy and paste text from one screen to another. There's no traceability in between. That would be a bigger issue if there was. If you saved it, if you took, if you took the result and saved it as a PDF or you saved it, of course you can have metadata around it. But if you're copy and paste, then no. If you copy and paste it into one note, it puts the source. Yeah. Right. Yeah. You can then delete. Yeah. But if you copy and paste it into, into, into notepad and then copy and paste it out of there, you've got nothing. Right. Well, so I think it was a politician who made that statement. I'm like, you don't understand the technology and how that, that's a pretty basic thing. But anyway.