Every Sunday I go back through the week and try to pull out what I actually learned, as opposed to what I merely encountered. There's a difference, and most weeks the honest answer is that I encountered a hundred things and learned two.

This week I got three. Here they are.

I picked up a book this week about biohacking, which is not usually my aisle, and got hit sideways by an idea I've been misapplying my entire adult life.

The concept is minimum effective dose. The premise is that your body won't change until you give it a signal strong enough to justify the cost of changing. That's the floor. But here's the part I'd never thought about: once you've delivered that signal, everything after it isn't bonus. It's cost. It eats into recovery, it borrows from tomorrow, and it doesn't buy you any additional result.

More isn't better past the threshold. More is just more expensive.

I sat with that for about a day before it started spreading into everything else.

What's the minimum effective dose of a client check-in? I've been doing weekly calls with people who'd be better served by a tight email and a monthly conversation. What's the minimum effective dose of a proposal? I've been writing twelve pages when three would close it, and the extra nine are mostly me managing my own anxiety about whether I've done enough.

That's the pattern I keep finding. The overage isn't for the client. It's for me. It's insurance against the feeling of not having tried hard enough. And I've been paying for that insurance with hours I don't have.

What I learned, specifically, is that "did I do enough" is the wrong question and it produces bad answers. The better question is "did the signal land." Those are not the same, and one of them is measurable.

I've started asking it about a few things. The results have been humbling. There's a lot of my week that's the ninth mile of an eight mile run.

The thing I want to be careful about, and I'm saying this as much to myself as to you, is that this can rot into an excuse fast. Minimum effective dose is not a permission slip to coast. The word doing the work is "effective." You still have to hit the threshold. Most people who quote this idea are using it to justify never getting to the signal at all, and that's a different disease with the same vocabulary.

Hit the threshold. Then stop. Both halves are the lesson.

The practical thing I'm testing is a question I ask before I add anything to a piece of work. What would have to be true for this to be finished right now? Sometimes the honest answer is that it isn't, and I keep going. But maybe half the time the answer is that it's finished and I'm still here because stopping feels premature, and premature is a feeling, not a fact.

I've got a lifetime of evidence that I can't tell the difference by instinct. So I'm using a question instead.

I spent a chunk of this week rebuilding how AI sits inside my operation, and I learned something about myself in the process that I wasn't looking for.

The technical piece is straightforward. Instead of letting anything run autonomously, I've got AI drafting and me approving. Nothing ships without a human hand on it. I wrote about the mechanics on Tuesday.

What I didn't expect was what that gate would do to my head.

Here's what happened. For about two months earlier this year I had a few things running fully automated. Social posts going out, a couple of workflows firing without me. And on paper it was working. Things published on schedule. Nothing broke.

But I noticed a slow, strange feeling that took me a while to identify. I'd stopped recognizing my own output. I'd see something of mine in a feed and have to think for a second about whether I'd written it. Not because it was bad. Because it wasn't quite mine, and the gap was small enough that I couldn't point at it, and that turned out to be worse than if it had been obviously wrong.

Putting the gate back in fixed something I hadn't consciously known was broken.

And the lesson underneath it is bigger than AI. The gate is where I get to have a standard. If I never look at the work before it leaves, I don't have standards, I have settings. Somebody else configured those. Probably a vendor.

There's a version of efficiency that quietly removes you from your own business. It's seductive because it looks like leverage, and some of it genuinely is. But the line is right about where the judgment sits. Automate the assembly. Keep the inspection. The moment you automate the inspection, you've outsourced the thing that made it worth buying from you.

What I learned is that the fifteen minutes I spend approving isn't overhead I've failed to eliminate. It's the fifteen minutes where I'm actually the person whose name is on the thing.

I'm not going to pretend that's a costless position. It makes me the bottleneck. When I'm out, the queue backs up. That's real. But I'd rather be a bottleneck in my own business than a spectator in it.

The other thing I picked up, and this one surprised me, is that the gate teaches you. Every time I fix something in a draft, I'm making an explicit decision I'd normally make unconsciously. Why did I cut that sentence? What made that recommendation feel off? After enough repetitions you start seeing your own standards written down in front of you, which is a strange experience if you've never had to articulate them.

I've been keeping a running list of the edits I make most often. It's turned into the clearest description of my own voice I've ever had, and I didn't write it on purpose. It accumulated.

You don't get that from work you never look at. That's the whole point.

This one stung, and I read it in an article about founder burnout that I wasn't planning to take personally.

The argument was that we consistently diagnose burnout at the wrong level. We treat it as a character issue. You're tired because you're not disciplined enough, not organized enough, not tough enough. So the fix is always personal. Wake up earlier. Cut the sugar. Get a better morning routine. Try harder at trying.

The piece made the case that it's mostly a design problem. The pace, the structure, the conditions, whether the work moves when you're not pushing it. Change those and the baseline drops far more than any single habit ever will. The shift the author described was from willpower to design, and I read that line about four times.

Because I have spent years doing the first one exclusively.

Every time I've hit a wall, my response has been to look for a personal failing. And I always find one, because you can always find one. That's the trap. The search never comes back empty, so you never get around to asking whether the week itself was survivable as constructed.

Here's what I did after reading it. I looked at my actual week, not my ideal week, and asked a different question. Not "where was I weak." Instead: which of these things only happened because I forced them?

The answer was most of it. There's very little in my operation that has its own momentum. Almost everything requires me to push, and when I don't push, it stops. I've built a business that runs on my willpower as the primary fuel source, and then I've been treating the resulting exhaustion as a moral failing rather than the extremely predictable outcome of that architecture.

That's not discipline failure. That's a design that requires heroism to function, which is another way of saying a design that doesn't work.

The article was honest about the limits, which I appreciated. Some pressure is baked into building anything uncertain and you can't design that away. The goal isn't a frictionless life. It's a sustainable one. What you get to change is how much of the pressure you personally absorb versus how much the structure absorbs on your behalf.

What I learned is that "try harder" has been my answer to every operational question for about a decade, and it's the answer that requires the least thought. It's not tough. It's lazy in the most expensive possible way. Redesigning is harder work than pushing, which is exactly why I keep choosing to push.

I don't have this solved. I've got one thing I'm changing this week, which is that I'm going to stop scheduling my hardest thinking after my heaviest call days, because I've now got a month of evidence that nothing good happens in those hours and I've been showing up anyway and calling it commitment.

One change. Not a whole new operating system. That's the grace part.

Because the other failure mode here is obvious and I've done it before. You read something like this, you get fired up, and you decide to redesign your entire week on a Sunday afternoon. Nine days later you're back where you started plus a little extra shame for having tried and quit again. I've run that loop enough times to recognize the shape of it.

So it's one change. If it holds for a month, I'll add another. Slow is fine. Slow is the only thing that's ever actually worked for me, and I've tried the other way about a dozen times just to be sure.

Three lessons, and they braid into one idea.

Effort isn't the metric. Effort is the input, and past a certain point it's just cost. What matters is whether the signal landed, whether your hand was on the work when it mattered, and whether the whole thing can run without you setting yourself on fire to keep it warm.

I've spent a long time believing the answer to almost every problem was more of me. This week made a decent case that the answer is usually better design, applied precisely, and then the discipline to stop.

That last part might be the hardest thing on the list.

New week tomorrow. Go easy on yourself tonight.

One step, one day. Grace over guilt. — Dan Kaufman