I’ve been listening to a lot of podcasts and on one of them somebody asked whether we’re all going to end up as pets to the robot overlords. That was The AI Fix, episode 100, Mark Stockley and Corey Noles. It was a joke.

I’m not complaining and I’m not worried about it. I’m excited about our future. What actually happened is that AI moved within like the last month from something that was kind of cool to something I was actually using every day.

Before that I was using it as a super google. I’d ask it the kind of thing I’d ask Google. Tell me about rabbits. I want to do research on rabbits. So instead of asking Google I’d ask ChatGPT, and then I’d go down a rabbit hole that was not meant as a pun.

The change started in a hotel. I was there for work watching one of NetworkChuck’s YouTube videos where he was talking about Claude Fable 5, which is Anthropic’s model and had just come out, and I watched it and it really sparked my interest. So that’s kind of like where it came from. I started playing with it a little bit more.

The first thing I really pointed it at was a big research question I actually cared about, and I had it go through everything and I built myself a document out of what it found.

It stopped being cool and started being useful about a week later, when I started building things with it. After that I kind of wanted to research more into what this AI thing is. How to use it, how to use it more efficiently, more than just a super google. So I started doing projects.

Now there’s a pile of them. I’m building a game. I’m building a MUD. I’ve got a personal assistant that’s an AI agent, I do research with a research agent, I’ve got another one that handles all my file management and another one that’s an archivist and keeps track of all the notes I make, and a couple of pipelines, and a board of directors that was part of one of the pipelines. I built a chat bot for my family with all the recipes in it, my mom’s recipes and everybody else’s. I’m just trying to learn how to use this, so I try different techniques with different projects.

The game was the first thing I built and it’s still in the works. Probably not the best thing for a first hand at using vibe coding with AI, but that’s what I did.

It isn’t looking like what’s in my head and it’s frustrating the crud out of me. It didn’t feel like I wanted it to feel, it just looks very minimum, and it was like the AI was doing exactly what I was telling it to do, which, it’s on me. I had spec documents. The specs were all about what the game does and there wasn’t anything in them about the feel of the game.

That’s what Claude Fable 5 fixed. I put the more expensive model on it for a little while to audit it and it got the picture rendering right, how the game actually looks, which the cheaper model I’d been running hadn’t. I also chewed through a ton of tokens doing it, and when you’re burning tokens you’re burning money.

I pay for all three of the majors right now, Gemini and ChatGPT and Anthropic. I use OpenAI because of the price, and that’s what it comes down to. Everything over there is on sale right now, like ninety percent off the regular price. If Claude did like ninety percent off I’d be using Claude, because I like the way Claude works better, but it’s not enough to offset the money.

I look at my tokens. I’m very conscious of how much I’m spending, I monitor it, I just need to actually put up a cost per token. Maybe that’s a future project. Right now it’s an expensive hobby. I’m not using it in a profession, I’d call myself an enthusiast at this point, and I would like to turn it into something that’s a career. That’s the goal.

Then my dad called me about something else. He’s learning Spanish. He’s known it my whole life, sort of, going back to his graduate work in Guatemala, and now he’s got this obsession with Duolingo and keeping the streak up. He’d been using ChatGPT. My brother-in-law and I told him to try Claude, and he was asking whether it could build him flash cards. I still haven’t asked him how it went.

Somewhere in the middle of that he asked me what I thought about data centers. This is just based off what I’ve heard in the community. No direct sources. Some of the fear is based off of real things, companies that took advantage of eminent domain, held the water supply back, drove the power up. The ones building now have figured out they can’t just roll into a town and wreck it, and there’s other good that comes with one, jobs and economy into the area. Nobody asks any of this about a golf course. You’ve got Scottsdale out in the desert with some of the greatest golf courses on the planet and that green grass has to be watered every single day and there’s a lot of them. Everybody knows it’s an issue. It’s a secret to nobody. It’s just not in the media.

AI is eventually going to be a basic utility. I lived through it once already, the dot-com era and the development of all that, and it’s been kind of a crazy ride. I moved about five times in the Coast Guard and every time it was the same thing. Get power. Get the internet. Those were the most important ones and I had them set up before I even moved into a new place, and that’s when I kind of came to the conclusion that the internet was a basic utility just like power and water. I’ve had that conversation with people at work and they agreed with me.

Imagine just not having internet. I can’t, even though I grew up without it.