Thinking Slow in an AI Economy

I was walking to my office today, wishing the commute was shorter. There's so much I want to do today, and I couldn't do any of it while walking. Ugh!

It's not even that long of a commute. 8 minutes, tops. But I had an idea now, and I was itching to get to work.

Funny thing happened once I stepped foot into the office. My racing thoughts slowed, and as I started my routine to get into "The Zone," I realized the idea I was so eager to start on was… not much of anything. Not even half-baked. I stared at my to-do list, that idea at the top, staring back at me like a dark void of potential — potential unlimited work and an unfinished project I'll eventually abandon.

I don't use AI much. It's primarily a code-writing tool for me. But I've dabbled in the chats, asking it for advice on what to do at the gym, what to have for dinner, questioning my life choices, and so on. As for code-writing, it's inevitable to spend a solid portion of the time planning out the work to be done, going back and forth with the agent about how to handle edge cases, test the functionality, and then work through the bugs that always arise from the implemented code.

It feels good to hand off the hard thinking to something that seems to do that part effortlessly.

And I suppose that’s what I’m getting at with all of this.

It's not that AI is making me (or you) dumber. It's that there's a very specific feeling I get when I paste a problem into a chat window — this tiny, almost imperceptible exhale. A "phew, not my problem anymore" feeling. The mental equivalent of putting down a heavy bag.

The trouble is, the bag I'm putting down is the part of my brain I actually want to keep strong.

My plan this morning had been to hand off the idea to Claude and let it rip. But recently I've been pulling back for job interviews, where assessments deliberately strip all forms of assistance and leave just you and what your mind is capable of. Going through these interviews made me realize just how easily, and how quickly, one can lose grasp of technical skill and knowledge. Things I used to know cold suddenly required effort. Not because I'd forgotten them, exactly — more like the path to them had gotten weedy from disuse.

Our brains are muscles that need to be worked out just like any other muscle.

They're just specialized to a specific workout that requires being challenged with thought. And like any muscle, the atrophy is invisible until the day you actually need to lift something.

Staring at my half-baked idea, I realized I was falling into that easy, effortless temptation to hand off the thinking to the robot.

We're all susceptible to this. Our minds are evolved, hard-wired to seek the easiest path to the desired outcome. It used to be for finding food or shelter — why waste energy taking down a large, difficult-to-catch creature at the risk of your own life if there's an easier alternative? That instinct kept us alive.

But now the "easier alternative" is a chat box that's specifically engineered to feel rewarding to use. The same instinct that once told us don't waste calories chasing the elk now tells us don't waste calories thinking through the problem. The thing about saving energy in the wild was that you still had to do plenty of hard things — the elk wasn't the only effort in your week.

Now, though, the hard thinking can be the one thing you're avoiding, and the day around it can be otherwise frictionless. You can route around your own brain pretty thoroughly if you want to.

So I closed the chat tab and sat with the idea instead.

It was uncomfortable. The idea was bad for a while. Then it was a little less bad. Then it became something I could actually describe in a sentence, which it turns out it hadn't been before. None of this would have happened if I'd outsourced the early-stage mush to something that's very good at making mush sound coherent before you've earned the coherence.

I'm not anti-AI. I use it. I'll keep using it. The point isn't to swear it off — the point is to notice when I'm reaching for it as a way to skip the part of the work that is the work. Pondering, sitting with the what-ifs, letting the half-thought rattle around long enough to become a whole one. That part doesn't compress well. You can't speedrun it without losing what makes it worth doing.

We need to think slowly.

We need to ponder ideas, navigate the what-ifs, sit with the challenging pieces that don't come together right away. Great things have never happened overnight, and AI isn't about to change that. If anything, it's raising the value of the people who can still tolerate the discomfort of not knowing yet — who can stay in the half-baked phase long enough to bake the thing properly.

That's most of what SteadyMaking is. A bet that there's still room for things made at the speed they actually want to be made at, long enough for the thinking to finish.

I'll take the 8-minute walk again tomorrow. I'm going to try to stop wishing it were shorter.