Sunday, October 25th, 2026

This week had a theme I didn't plan. Everything I read, heard, and stumbled into kept circling the same question. How long does it actually take to become someone new, and what are you allowed to bring with you?

Turns out the answers are longer than I wanted, and better than I feared.

Three lessons. None of them complicated. All three are already changing how I'm running next week, which is still the only test I care about for this letter.

I read a breakdown this week of where the famous sixty six days to form a habit number comes from. It's a 2010 study out of University College London. People picked one new daily behavior and tracked how automatic it felt for twelve weeks.

The median was sixty six days. But the range ran from eighteen days to a projected two hundred fifty four. And exercise habits trended slower than the rest, with a median of around ninety one days.

The twenty one day rule everybody quotes? It didn't come from a habit study at all. It came from self help writing in the 1960s.

When I read that, I did the math on my own life.

I started a new workout routine a few weeks ago. By the twenty one day rule, it should be about to click any day now. It doesn't feel close. There's still a little negotiation in my head most mornings. And I'll be honest, I'd started to read that negotiation as a sign. Like maybe this routine isn't the right one. Maybe I'm not built for it. Maybe I should tweak it, swap it, start a different plan on Monday.

Here's how that goes sideways. You use the wrong timeline, so you misread normal as failure. Then you quit something that was working, start something new, and reset the clock to zero. People call that inconsistency. I think a lot of it is bad math.

By the actual research, I'm not behind. I'm about a quarter of the way there. Maybe less.

The second finding from the study matters even more. Missing a single day didn't materially affect habit formation. One miss doesn't reset you.

I can't tell you how many habits I've abandoned in my life because I missed a Tuesday, decided the streak was broken, and let the whole thing go. As if the habit were a glass vase and one miss shattered it. It's not a vase. It's more like a trail through tall grass. Skip a day and the trail's still there. Skip a month and the grass grows back.

So here's what I'm changing.

I'm giving every new habit ninety days before I'm allowed to judge it. Not twenty one. Ninety. If it's exercise, I'm not even asking the question until after New Year's.

I'm adopting a simple rule I've heard for years and never actually lived by. Never miss twice. One miss is life. Two in a row is the start of a new habit, and it's the wrong one.

And I'm tying each habit to a fixed cue, the same way the study participants did. Not "sometime in the morning." After the first cup of coffee. After Sheldon's walk. After I close the laptop for lunch. The cue does the remembering so willpower doesn't have to.

This applies just as much to the business. The daily outreach. The weekly numbers review. The content you swore you'd publish on a schedule. Most owners judge a new business habit after three weeks, decide it isn't working, and go chase a new tactic. Give it ninety days. You'll be amazed how many "failed" habits were just interrupted ones.

The article that got under my skin this week was from Scientific American, about AI and skill loss.

There's a study in there I keep coming back to. Endoscopy specialists in Poland, every one of them with at least two thousand colonoscopies under their belt, were given an AI tool that flags precancerous growths. Some days it was available. Some days it wasn't. After a few months of using it, their detection rate on the days they worked without it dropped from 28.4 percent to 22.4 percent.

These were experts. Two thousand procedures each. And they got measurably worse at the core skill of their job once a tool started doing part of it for them.

There was a second study, on software engineers. The ones who used an AI assistant on a task scored 50 percent on a quiz about what they'd learned. The ones who didn't scored 67. The AI group was especially weak at diagnosing errors. They got the work done without understanding it.

I use these tools every single day. I've built a big part of my work around them. So I read that piece with a knot in my stomach, and then I asked myself a question I didn't love.

When was the last time I wrote the first draft of an ad, a headline, or a hard email with nothing open but a blank page?

The honest answer is longer ago than I'd like.

Now, I've talked here before about my rule that I create and AI critiques. I still believe it. But reading this, I realized a rule about process isn't enough. I also need a list of which skills I'm not willing to let slide no matter how convenient the tool gets.

So I made two lists.

The rent list. Skills I'm happy to hand to a tool and not lose sleep over. Formatting. Summarizing a long document. Pulling research together. Turning my messy notes into a clean outline. Cleaning up a spreadsheet. If I lose a step on those, nobody gets hurt.

The own list. Skills that are the reason anybody pays me. Diagnosing why an ad isn't working. Writing a hook that stops somebody mid scroll. Reading the room on a sales call. Knowing which client problem is the real one and which one is the symptom. Making the call on price. If I lose a step on those, I lose the business.

Here's the practice I'm putting in place for the own list. Once a week, I do a rep with no tools. Write ten headlines from scratch on paper. Diagnose an underperforming campaign with nothing but my own eyes and the numbers before I ask any model what it thinks. Draft the hard email myself start to finish. Then, and only then, I can use AI to check my work.

The researcher in the article had a great way of putting the danger. People can perform at a high level because they're borrowing the skill from the AI, without ever building the skill themselves. That's fine for the rent list. It's a slow disaster for the own list.

If you run a team, this is double important. Your junior people are building their skills right now, in a world where the tool can do the first draft of almost everything. If they never do the reps, they never become senior. Somebody has to protect their practice time, and that somebody is you.

Make your two lists this week. It takes ten minutes. Then schedule one no tools rep on your calendar for something on the own list. The point isn't to use AI less. The point is to stay the kind of person who'd know if the AI got it wrong.

The third one came from an article about changing careers in your forties, and from a short book I picked up this week, and they ended up landing in the same place.

The article's big argument is that a career change in your forties is a pivot, not a restart. The skills that made you good, reading people, building trust, structuring an argument, handling a difficult personality, don't disappear when you change direction. They travel. What changes is the container.

I needed that reminder more than I'd like to admit.

When I rebuilt this year, there were plenty of days where it felt like I was starting from zero. New city. A business rebuilt on a different foundation. A lot of people who'd never heard of me. It's easy to look at all that and feel like a rookie in his late forties.

But that's not what was happening. I was starting from twenty five years of experience in a new container. The phone calls, the copy, the strategy, the instincts for what makes people buy. All of it came with me in the moving boxes. I just wasn't giving myself credit for unpacking it.

The book, How to Be Good at Life by Blake Mair, pushed on the same idea from a different side. Mair argues that we measure a good life by too narrow a scoreboard, mostly money and status, and he lays out a wider one. Character. Relationships. Growth. Resilience. Financial responsibility. Service. When I looked at my own rebuild through that wider lens, I found a lot more already in the moving boxes than I'd been counting.

But the part of the article that actually stopped me was about guilt.

The writer, a coach who spent years on the hiring side, says the thing nobody names during a big change is guilt. Guilt about disrupting your family's income. Guilt about the years people spent supporting the old version of you. Guilt about walking away from something that made other people proud.

And her answer wasn't a pep talk about following your dreams. It was this. You answer guilt with a plan. With evidence and sequencing, so the people who depend on you aren't being asked to just trust a leap.

My whole brand is built around a phrase with the word guilt in it, and I'd never thought about it that way.

I've always framed grace over guilt as permission. Give yourself grace. Let go of the guilt. And that's still true. But this week I realized there's a second half I'd been skipping. Some guilt is just noise and you should let it go. Some guilt is information. It's telling you there are real people affected by your choices and you owe them a plan, not an announcement.

Grace doesn't mean ignoring that kind of guilt. It means listening to it without letting it drive.

So here's the exercise I did, and I'd recommend it if you're in the middle of any kind of change.

Write down what transfers. Every skill, relationship, and hard won lesson from the old chapter that still applies in the new one. Be generous. The list will be longer than you think.

Then write down the guilt. Specifically. Not "I feel bad." Who, exactly, are you worried about, and what are you afraid it'll cost them?

Then, next to each line of guilt, write one concrete step that answers it. A timeline. A safety net. A conversation you need to have. A small test before the big move.

When I did it, the guilt didn't disappear. But it went from a vague cloud over everything to a short list of things I could actually do something about. That's the difference between guilt that paralyzes and guilt that points.

Three lessons. Give your habits ninety days and never miss twice. Decide which skills you're renting and which ones you have to own, and practice the ones you own without help. And remember you're not starting from zero, and the guilt that shows up during a change deserves a plan, not a lecture.

Here's the small version for this week. Pick the one new habit you're most tempted to quit. Write today's date on a sticky note and add ninety days to it. Put it somewhere you'll see it every morning. That's the first day you're allowed to decide whether it's working.

Until then, you just keep showing up. Even on the awkward days. Especially on those.

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