On September 28th, around 8:37 at night, I told Muse, “Blog idea tonight I was running 3 different AI. Muse, claude and grok bot. Got to be post in there somewhere.”

I was at the dining room table. I turned it into my desk. My laptop sits on a stand connected to a curved monitor that I had split screen, and my computer cluster runs to my right with another monitor hooked to it that shows my daily schedule, what I need to focus on, the weather, and what tasks I need to complete today. I was wearing shorts and a T-shirt that says Don’t Panic on it, from The Hitchhiker’s Guide to the Galaxy, and I had a fruit punch or something like that. I was working on a bunch of different things and I just noticed I had three separate, completely different AIs going at the same time. Actually it was four, counting Jev. I’ve been doing this for maybe three months, and that kind of blew my mind.

Muse is an AI from Meta. I use it as my personal assistant, it’s where I keep track of the things that are going on.

Claude Code is where I do all my projects, so that’s where the coding happens. That night I was in it working with Brittany, one of my AI agents that runs on Hermes, on getting Jev to categorize and label everything in my Obsidian vault so Brittany would be able to find things easier. Jev is an AI from TypeSafe, and I’ve written about building with it before.

The Grok bot is helping me organize some things for a business I’m starting, which is going to be the grand reveal I guess. One of my Grok bots is an inbox bot I named Mae, and that night she was getting set up to go through all my emails and clean them up, with Jev doing the categorizing for her too.

I get my emails on my phone and I just kind of ignore them. A friend of mine emailed me and I guess it got lost in all the other stuff, because I didn’t catch it. Mae told me, hey, this guy, we haven’t responded yet. That was good to know.

The way it works is I have three email accounts, two Gmail and one Zoho, and Jev goes through all of them and sorts everything into a bunch of categories like promo and noise and newsletters and personal and finance, and it does the flagging, and then Mae goes in and deletes only the categories I designated, the newsletters and the promos and the noise, without having to ask me anything. Jev is so cheap that having it do the categorizing saves a lot of tokens. So far nothing has been deleted that shouldn’t have been.

None of them know about each other. If I wanted to know what Claude was doing I had to ask Claude, and if I wanted to know what the Grok bot was doing I had to ask the Grok bot. It’s like the right hand not knowing what the left hand is doing. I wasn’t copying and pasting between them, though. They were on separate tasks, each one running its own thing in its own little universe, and with the screens open I could just read them. It never really got messy.

I guess you can say I’m the boss and they work for me. I tell them what to do. Just like normal employees they do some things autonomously, but I’m the one who set those up. Some people want it to ask permission on everything and some people don’t want to think about it at all. I think I’m kind of in between.

I just went to a wedding with friends I’ve had since high school, and most of them are not real happy about AI right now. I told them where I stood and what I was using it for, that I’m not doing it to replace me, I’m doing it to enhance me. They said OK, what I’m using it for seems pretty good.

Three or four months ago, if you asked me who runs four AIs at once in the evening, I would have pictured somebody in a dark room with six monitors up, wearing a hoodie, hacking stuff, not some guy in shorts and a Don’t Panic shirt with a fruit punch who’s trying to start a business, with his personal assistant working on stuff and the coding going and Jev categorizing for the other two. I kind of wish I was the dark room guy with the hoodie and six monitors. I just have a laptop and I don’t think it would support six monitors. I don’t fit the mold of a hacker. I’m a vibe coder.

Right now, doing this, my setup is about the same. Claude is on the left part of my curved monitor. Muse is on the right side, looking through my past sessions to help jog my memory. The Grok bot is on my laptop screen, and I asked it what I was doing on September 28th too. When I sat down to answer questions for this post I had to stop and go figure out what Claude was even working on that night. I think I’m wearing the same shirt.