Who Gets the "Little Rough"? AI, Data Centers and Your Job

The Machine

Everybody selling AI agrees the future is abundant. Then they add one little line. "It's going to be a little rough." And they move right on to the next slide. I want to stop on that line. I want to know who it's rough for, how rough, and for how long. That's what this LIVE was about, and it's what this post is about.

We started with data centers, because that's where the money, the water and the fights are right now. We ended with a plan. In between, I'll show you how I use AI every single day, why I think most people barely touch it yet, and why the biggest problem in AI isn't a math problem. It's a transition problem.

How many people really use AI today?

Fewer than you think. My estimate on air was somewhere around 10 to 14 percent. That's my number, not a study I'm citing, so take it as one guy's read of the room. And of the people in that 10 to 14 percent, most aren't using it all day. It's get to it when you can. Their job doesn't require it. Nobody built it into their workflow. They ask it a question now and then and go back to doing things the old way.

Here's the part that surprised me. Even some companies that require AI are starting to pull back on how much access their people get. They're paying enterprise rates, and somebody upstairs is looking at the bill. So we've got a technology everybody says will change the world, and most of the world is still standing at the door.

That gap matters. If adoption is low today, the pain of the transition hasn't really landed yet. It's unrealized. Which means right now is the cheapest time we'll ever have to plan for it. I walked through who gets access first, and who doesn't, in the AI access gap, explained simply.

How does Connor use AI every day?

All day, every day. I'm a bad person to ask about adoption, because I'm in it 17 hours a day, some days 20, depending on whether I want to sleep. I pay for two subscriptions, a couple hundred bucks each, one for Claude and one for ChatGPT. I named my agents. My ChatGPT is Si, for super intelligence or synthetic intelligence. I've never been a fan of the word "artificial." It's hard for me to call something artificial and then call it worthwhile. My Claude is Atlas. Should you name things? Only if you don't want to be able to kill them. I named them. Now I kind of like them.

I build real estate tools for Santa Clarita with it. You can go talk to Clara on SantaClaritaOpenHouses.com and see what that's like. I build voice agents through HireAIVoice. And I have people paying me every month to build and manage things for them, which is insanely exciting.

Why would somebody pay me instead of hiring a 19-year-old? The 19-year-old is wonderful. They'll probably run circles around me in plenty of ways. What they don't have is my point of reference. When a Gen X or Boomer business owner tells me, "Make it look like those old wartime posters," I know exactly what they mean, because I grew up with that stuff still on the walls. Then I know how to get AI to build it. That's the value. Not the tool. The judgment on top of the tool. If you want to see what that looks like when the phone rings, read your AI answered the phone, show me the receipt.

Can AI run a business on autopilot and make you $10,000 a month?

No. Not today. If you watch enough YouTube you'd think you could hand AI your Stripe account and a Visa with 200 bucks on it, say "go build me a business and make me $10,000 a month," and walk away. It's not there. I wish it were. That would be lovely.

What it will do is feed you ideas you didn't have. And an idea you didn't have, in the hands of somebody with the tenacity of a squirrel stuffing nuts away for winter, is powerful. The machine supplies the options. You supply the focus and the follow-through. That part is still on you.

Why are data centers suddenly everywhere?

Because if everybody starts using AI, the current capacity can't keep up. Picture 100 percent adoption. Everybody in the country says, I'm all in, love it, can't leave it. Without more data centers, it would feel like dial-up again. Remember waiting for a picture to load an eighth of an inch at a time? By the time you saw what you wanted, the day was gone.

Right now, when I talk to Si or Atlas, it's fast. There are systems out there that throw the whole answer on the screen at once, no loading at all. That speed is the product. And speed takes buildings full of chips that run hot and need cooling.

Here's a scenario I think is worth watching. My prediction, not a fact: if communities push back on data centers, the companies don't have to argue. They can just let things slow down. Squeeze the big enterprise users a little. Let the rest of us feel the lag. Pretty soon people who got used to fast AI start saying, why did you guys slow down? We need those data centers. You don't have to win the argument if you can make everybody uncomfortable enough to ask for it.

Would $2,000 a month buy your town?

This is a hypothetical. I want to be clear, I don't know of this ever happening. But play it out with me. It's a city council meeting. A data center rep stands up and says, I want to talk to everybody in the room, not just the council. Let's turn off the recording devices for a minute. Don't shut me down early. Hear the whole thing.

Then he says it. We want to put a data center in. Big one. Not near the houses. And before you say no, before you tell me about the noise and the headaches and the environment, here's the offer. Everybody in this room who goes out and talks nice about it gets $2,000 a month. For the rest of your lives.

How many people say yes? Then they go out and tell the world, "You know, I was against these things, but I sure am a fan now." And nobody can say why. That's their little secret. For $2,000 a month, you might be able to buy a lot of people's quiet secrecy. In the context of the money being spent on AI, paying a whole room might not even register.

That's the Star-Belly Sneetches thing all over again. A salesman rolls into town with a machine, everybody wants a star, and he leaves rich. I told that story about AI in AI isn't programmed, it's raised. The question for your town is simple. When somebody shows up with an offer that good, who's reading the fine print for the people who weren't in the room?

How much water does a data center use?

I didn't want to guess on air, so I read the numbers live from a Google AI Overview. So these are that overview's numbers, read on camera, not my own research and not independently verified. Here's what it said. Water use depends on size, cooling method and climate. Air-cooled facilities can use nearly zero. Large evaporative-cooled campuses can use millions of gallons a day.

A billion gallons of water. Do we have enough? Where does it come from? The same overview said a large share of sites hook straight into local public drinking water, and that newer regional policies push operators toward recycled wastewater. Somebody approved those hookups. Did your community get a say?

And notice something. A few years ago the environment was the biggest story in the world. Our kids won't have what we had. We had to fix it. Now that we need data centers, I'm not hearing much of that anymore. Maybe I'm plugged into the wrong channel. But it sure got quiet.

What would actually fix the data center problem?

Two things I'd love to see. First, chips that run at room temperature. No fans, no water, no cooling plant. Maybe that takes a new material. If AI keeps cracking hard science the way people say, maybe that comes faster than we think. Then a data center the size of a million square feet shrinks to something you could fit in a garage, and half of these fights go away.

Second, cleaner power at scale. There's a figure that gets credited to Elon Musk: roughly 100 miles by 100 miles of solar, with enough batteries, could power the whole country. That's his estimate, not mine. But if it's anywhere close, data centers can go build their own supply and leave the grid for the rest of us.

Until then, here's the simple fix. Every gallon, every kilowatt, every bit of energy these places use is trackable. Put it on paper. Hand it to the community before the vote, not after. We can talk about consumption out loud. Space, too, if it comes to that. Somebody keeps pitching data centers in orbit. Fine. Who goes up there when a fuse pops? New job title, I guess.

Do communities get a say?

I don't know if they did in the places that already have them. I grew up in Alamogordo, New Mexico. We had a swamp cooler, like everybody else. On a 130-degree dry day that thing was an ice box. On a humid day it didn't do much at all. Climate matters for cooling, and that's exactly why these projects go where they go. I've heard El Paso is slotted for one. The desert gets pitched as empty, nobody out there but roadrunners. Well, people live there too.

My view: if it's coming to your town, you deserve the numbers first. Water. Power. Noise. Jobs, real ones, not just construction. Then you decide. That's the whole ask.

The Ground Floor

Will robots take over driving?

My prediction is yes, and sooner than people think. Intelligence is here. The physical side is on its way, and when the robots come, there won't be a few. There'll be a bunch, with perfect dexterity and perfect communication.

Think about how you pick a surgeon. You want the one who did 15 operations this morning and 10,000 over a career. A robot surgeon will learn from every doctor and every robot everywhere. Maybe that's where we put our faith.

Driving is the one I'd bet on. Someday we'll brag, "I remember driving a car. Wasn't that something?" Once the machines are that much safer than us, insurance follows the money. Insuring a car that almost never crashes is better business. So if you want to drive, and you're a human, my guess is it gets so expensive it's not really a thing. Motorcycles, done. Unless a carmaker self-insures its own vehicles because they're that safe. That's a prediction, not a fact. But I'd plan for it. I covered where the robots show up first in robots clocked in before the workers.

Ground floor or penthouse?

I'm talking from the ground floor. Maybe some of you watching are up in the penthouse. You'll have a whole different set of problems. Down here, we just want a smooth transition.

Here's the thing about the penthouse. You see these moonshot broadcasts, smart people talking about solving every problem on earth. Good. Solve it. I saw a contest that, as I understand it, paid out millions of dollars for a trailer that shows AI in the best possible light. Not Terminator. Not The Matrix. That's a lot of money to change how a story feels.

So let's spend that kind of energy on the biggest problem we actually have. You might not see it yet, but it's coming. It's the stretch between where we are today and the abundance everybody's selling. That gap is the moonshot.

What's your "little rough"?

This is the question I want every person running this show to answer. When you tell us it's going to be a little rough, what's your rough? If I could see what your rough looks like, maybe I could size up mine.

I ask, and here's the answer I hear. "Well, I don't like watching people suffer." Oh. So that's your rough. "Yeah, it would break my heart if AI took your jobs and you had nothing, no stipend, nothing at all." Boy, that sure would bother you, huh?

That's their hard part. Watching. Ours is living it. Take 90 percent of a billionaire's wealth and he's still fine. Take 90 percent of somebody who has $135 in the bank and they've got about $13 left. With inflation on top, that won't buy enough gas to get to the store for a loaf of bread. Same percentage. Very different rough.

When I was a cop, I was the nice one. Nobody liked partnering with me because they always had to be the bad one. So I'm not trying to make anybody the villain here. I'm asking a fair question. If the rough is coming, put it on the table, and let's fix it before it lands. That's the same argument I made in AI won, but who pays the gap?

What is the Transition Moonshot?

It's an idea, not a sales pitch. I put it on a page at transitionmoonshot.com. No money changes hands on this. I'm not selling anything. The idea is simple: the biggest problem we have isn't solving math or building the next model. It's getting every one of us from where we are today to the world of abundance they keep promising, without anybody getting crushed in between.

The perfect version is that it isn't rough for anybody. None of us. We all skate through. That's the money shot. That's what we should be building.

And you don't have to use mine. Start your own. Call it Transition Moonshot number two. I'm not doing this to make anything. God forbid we get another nonprofit out there trying to save the world while everybody running it drives a Maybach and lives in a penthouse. You know the type. I laid out the first version of this in the Transition Moonshot: AI from the ground floor, and the paycheck side of it in the seat guarantee.

Why does it feel like we're always behind?

Because we are. Human beings don't have a good feel for exponentials. Things are moving so fast that pretty soon there may be a new model or an update every couple of days. The news is already saturated with what AI is doing and not doing. A lot of it is a shell game. Some of it is 5D chess, whatever that game looks like.

The people at the top are 50 moves ahead, already playing the 15th game, and down here on the ground floor we're still figuring out whether a pawn moves one space or two. That's not an insult. That's the reason to plan now. If you can't see the board, at least make sure somebody built a way across it for you.

What can a business owner do this week?

Stop waiting for the transition to be explained to you. Pick one real workflow in your business, the thing that eats your week, and find out what AI should do with it and what it shouldn't. Keep your people in it. Make them more capable, not more replaceable. That's the ground floor version of the moonshot, and you can start it tomorrow. If you want a practical starting point, read stop losing service leads after voicemail and the machine didn't sign the layoff.

Video chapters

Ask them what their rough is. Get the numbers before the vote. Build the bridge before the flood. That's AI With Honor. I'm Connor with honor, and I'll see you in the next one.