← Writing

To AI or not

Gary Vee is a smart marketer with lots of valuable things to say.

He worked at his dad’s liquor store, and turned to talking about wine on YouTube. It blew up his store’s business and propelled his brand into a multi-million dollar juggernaut.

Now, he talks about marketing and personal branding. He’s everywhere, frequently popping up in my YouTube feed dispensing nuggets of advice about how to create a personal brand like his.

Lately, Gary’s been talking about AI. (It would be weird if he wasn’t).

He believes society will turn into two types of people, mostly by 2030, creating a picture of a barbell: Those who AI and those who run away and avoid tech at all costs.

You might call it the hallelujah/fuckit divide. There’s the all-in, and the all-out. With a very slim bar in between holding the two together.

The hallelujah side is turbo-charged and all in. This person is heads-down and automates everything. I’m guessing you probably already know about these types.

The fuckit person signed off and said sayonara! to tech. It’s just gotten too be too much. Maybe they’ve joined the re-wilding movement.

In history, we had the haves and have-nots. In this framework, we have the do and the do-nots.

The barbell takes shape

Scroll LinkedIn for awhile and take notes on what people are saying about AI writing.

You’ll see the outlines of the barbell if you look long enough.

On one end, you have people boasting about being “content engineers.” They talk about building authority at scale and generating unlimited inbound leads. They’ve mastered every technical aspect of leveraging AI. They use Claude chat, moved beyond it to Cowork, maybe even Claude Code. Maybe they’ve built their own Karpathy-style second brain. They are button-pushers of the most kind.

Don’t worry if some of those words aren’t familiar to you.

And look, they’re right about one thing. Leverage matters. Doing everything by hand in 2026 isn’t noble, it’s slow. But investing 100% of yourself into the machine is a mistake, because it works against the grain of what it means to be a human being. Lose sight of what it means to be human, and the only direction you’re going towards is decay.

Then there’s the other end of the barbell.

They’re complaining about the em dash and how it’s proof a robot wrote your post. These are the types who’ll comment that your article is “definitely AI slop” — whether you used AI or not. They’re all about the outdoors and touching grass.

They’re right about one thing too. Slop is real, there’s an ocean of it, and their allergy to it is a healthy immune response. They’ve just aimed it at punctuation instead of at the actual disease.

You can’t see it, so it must not be there

Now here’s the part Gary Vee can’t see. The part nobody can see, and that’s exactly the point.

There’s a third group: people using AI to help with their content while the thinking stays unmistakably their own. They don’t use AI to think for them. They use it as a collaborator. They hand off the repetitive, administrative work and keep the thinking for themselves. The writers producing work that’s actually valuable focus on the valuable part, which takes a lot of work, and don’t let AI water it down.

When these people publish, their work doesn’t land on either end of the barbell. It doesn’t look AI-assisted, because good AI-assisted work doesn’t look AI-assisted. By definition.

That’s why the barbell survives as a theory: its counterexamples are invisible. You can’t count the people you can’t detect. Gary Vee is standing in a room full of them, seeing an empty middle.

Third way blueprint

There is nothing wrong with being on either side of the barbell, but I would argue a healthy third-way is better.

At this point, I’m ditching the barbell metaphor. A barbell with three ends would be strange.

If you’ve read this far, you might be looking for some concrete steps to build this magical third configuration. I wish I had a perfect list, but the fact is I’m building this out myself at the moment.

Future content will address my exact process and give you the tools to do the same.

Until then, codifying this into your working memory will help you start on a path to grow as a human, bypass the noise, and come to peace with AI.

Let’s take the first step and put this third way into practice.

Order matters

But — and this is everything — the order you do it in matters.

Why? Because tools amplify whatever’s already there. Give a clear thinker AI and you get clear thinking at scale. Give an empty tank AI and you get emptiness at scale — which, not coincidentally, is a perfect description of the content-engineer end of the barbell. The machine didn’t hollow those people out. It just published their hollowness faster.

So: human end first. Then the technical end, where and when it makes sense.

Primal first, always.

Start with a primal reset — the event. One deliberate stretch where you rebuild the foundation.

This is an oversimplified list to give you an idea:

  • Practice thinking, especially thinking on paper. Messy thoughts turn into organized thoughts. Organized thoughts turn into output that’s useful to you and to others. The origin of your thinking should be sourced at the speed of your brain, not outsourced to a machine.
  • Don’t be a hostage: get outside and move — and sometimes exercise vigorously, maybe once a week.
  • Connect with other humans — especially the ones close to you, and also the ones just outside your front door. Hug your kids if you have them.

Then comes the primal practice — the maintenance. The reset gets you the foundation; the practice keeps it. This is where AI finally enters the picture: not as the thing you maintain, but as the thing that helps you maintain everything else.

Here’s what that actually looks like.

About a month ago, I asked AI for an all-body 15 minute kettlebell workout. One-thirty pound weight. It worked great for a week. Just the right workout to fit into my busy routine with very low friction, other than my mental resistance to vigorous effort.

I felt great after each workout.

After a week I asked for a 20 minute workout and by the way, create a plan and a schedule to increase my workout so it’s progressive. Just tell give me the routine and tell me how long to do it before the next level-up.

The AI asked if it could access my health data on my iphone.

That was ten minutes of setup. Now the most advanced technology on earth spends its cycles getting me away from the screen.

I got stronger, and better at using AI to advance growth.

If you’re a writer, be a better writer first. Then fold in AI to do it even better and quicker. If you’re a knowledge worker, make better and more efficient processes, then build AI agents around it to eliminate the boring repetitive stuff so you can spend more time applying your human insight where it’s needed.

That’s the whole model. The tools serve the primal.

Get that order right, and the barbell was never your problem to begin with.