Hello. My name is Patrick.

I've been using AI and AI tools to help me write software more efficiently. And one of the newer use cases I've found for AI tools has actually helped me with a lot of the other applications in my life, like note-taking and desktop publishing, et cetera.

And so the software I use for note-taking is called Obsidian. Basically, it'll take any of your journaling, food recipes, other notes, and save them into your Dropbox. So they live on your computer, on your hard drive.

And so when you're thinking of the same tools we use for AI code gen, one of my favorite tools is Aider, which basically lets you add and remove files from your source code repository, and it'll synthesize them and understand your entire software project, such that you can begin to ask it to develop with you, develop things for you. So if you wanted to add a contact form for your website, you can load all of your source code files into memory and just ask Aider, "add a contact form to my website."

And so what I've started also using now is outside of like the programming sphere, using Aider with Obsidian.

So one of the things, as a demo, I wanted to add, create a document describing my computer backup strategy and schedule, just to make sense of it. So I have a few notes in my Dropbox folder that describe how I back everything up. I take a backup every morning at 5 a.m. And then once a month, I also send it off to somewhere on the network or into the cloud encrypted, what have you.

And so the cool thing with Aider is it can do all this, right? It can create one document describing everything very formally and technically. So it'll have, if you use this "slash add" command, you see it'll do this fuzzy search and can start finding the different notes that describe my computer backups. And I'm able to ask it to prepare a document for me.

So: create a document, backups.markdown. Please describe backup strategy and schedule and storage locations in a Mermaid diagram. So Mermaid is a syntax for describing diagrams and flowcharts in text, which is very useful because of course, ChatGPT and Claude and other large language models, you know, they speak in text. But, we'll be able to render it as a graphic.

I'm gonna add some more. So please, please also add—you always gotta say please. Please also add key facts as a bulleted list at the bottom. And finally describe the 3-2-1 backup methodology at the top in its own section. So of course, three, two, one means you always need three copies of a backup on at least two different devices or storage mediums. And most critically, you need one backup offsite. It doesn't do you any good if you have all three copies in your apartment and then your apartment goes up in flames. So you always need one backup on like a USB drive at your parents' house or saved encrypted to the cloud, wherever.

So this is enough to be able to prepare this document. So it'll take like 15 seconds, I think. And you see the large language model thinking. This is the markdown file it's preparing for us. And it's actually gone through and read like a few thousand words of documents in my own like note-taking folder on Dropbox.

It's asked me to create this file. Yes. And that's all it takes. So now if I use the GitHub command-line tools, it's very easy for me to create a Gist. A Gist is just like a single webpage hosted on github.com that lets me share this document with whoever. And so I'll go ahead and do that.

And as soon as we open this, we expect to see this. So it did exactly what we prompted. The 3-2-1 backup methodology is described at the top. It prepared this nice little flow chart of how my PC backups at 5 a.m. every day. They're retained for 28 days then on a network share. I keep three months of them on a secondary network share. And then finally also I keep the last two full backups encrypted in the cloud. So I can go back like six months or whatever the case may be.

And all this was just from the large language model, you know, reading my own notes, synthesizing, understanding, and then outputting this document for me.

So it's pretty cool to think that all the AI tools that are being built right now for developers also have use cases for all the other software that we work with. As long as it's— generally it's gonna be amenable to things that are already working with text files, what have you.

And I just wanted to share this use case as kind of a jumping off point for how we can actually build apps that are more tailored to do all this for a lot of the rest of the world who may not be tech savvy enough, of course, to install Aider or let alone like a Python runtime on their Windows PC.

So anyway, have a good rest of your day. Ciao.