xperiment is complete. Good, good. Now there might be a difference on the browser version versus the app. I don't know about that yet.

Well, let me tell you about my AI experience yesterday. If you were like me, and I like to think everybody is, you probably said to yourself sometime in the last year or so, I should spend some time working with AI just so I don't get left behind. Anybody say that to themselves? And a lot of you are probably already using it for work. But I said to myself I'm going to wait. And so what I thought I would do was build a little app, which in the AI world, if you use OpenAI ChatGPT version, would be called a GPT. So it lets you talk your way into creating an app. And so I tried that, and I said to myself, what can a clever person in his senior years who doesn't work with technology all the time anymore, like how far can I get?

Here's the problem. The interfaces for using AI are weirdly way behind the AI. So no matter how powerful the AI is, getting you to use it as a user is weirdly, it's almost ridiculously non-obvious. For example, I probably spent an hour yesterday not knowing that there was a switch where I could intentionally change it from version 3.5, which wouldn't do anything for me because it didn't have the features I needed for what I was doing, to 4.0. I spent the whole time trying to debug why it wasn't doing what I wanted, and I wasn't even in the right version. Now why does it not default to the current version, or why doesn't it flash and tell you I'm definitely on this version so you can do these things? Right now, I mean it's just, that's just a user interface, you know, total botch. That's all it was.

So I spent hours just trying to figure that out. Do you know how hard it was even to find the official ChatGPT app? You go to the App Store and there's so many crazy little apps that have the same name, and they try to make a logo that looks sort of like the official one but different. It's 90% garbage apps or apps that are not the ones you want. So I spent maybe an hour, maybe an hour even trying to look for the app. Have you ever spent an hour looking for the app when you knew the name of the company? I mean that's so weird. Why was it so hard?

But then when I had the actual experience of building it, what's new about version 4 is it can access an external file. So you can feed it a file and then ask questions about it and build a little app around its special knowledge of maybe that one file or what you taught it. And that little special knowledge would be your GPT. So it was actually a freaky experience building an app by talking to it. But that kind of worked, meaning I just had a conversation with it. And the thing that's really mind-blowing is that it makes suggestions after it fails. So you'll say, can you do this? It will try and it will say I can't do that with the file you've given me, but I'm going to try this instead. And I'm like, what? Seriously? You're actually keeping me informed what you're trying? A whole new look. And then it tries another one. But it was freaky. It talks to