Liz caught the scam before any of us knew there was a scam to catch.

It sat near the top of the list the tool had ranked for us, a nearly perfect candidate with a strong résumé and exactly the right experience. The kind of profile a scoring model rewards without hesitation.

Liz didn’t trust it. It was too clean, too precisely matched, and she had learned that the thing arriving looking flawless is often the thing worth checking first.

She was right. The résumé was bait, engineered to look ideal and get someone to click, and the click was the entire point. The model had read it and rewarded it and never suspected it was being worked, because being worked is not a category a rubric has a field for.

The scam was obvious once someone became suspicious. The more consequential things the tool missed did not look suspicious at all.

Liz founded the agency. She was its first executive director, the person who built the role out of nothing. Two years ago she handed it off, ready to try something new after decades spent building and leading organizations like this one. Now we needed a new director, and the board asked her to join us and lead the search. No one understood the job better than the woman who had invented it.

A hundred résumés came in, more than a volunteer board and a founder with a day job could read with the attention each person deserved. Big companies have software for this, platforms that swallow applicants, sort and score them, and hand back a tidy short list. We are not a big company. We couldn’t afford one, didn’t especially want one, and had no reason to buy an enterprise system for a search we hoped not to repeat anytime soon.

So I built one. I’m the board chair and Liz’s husband, and I’m the one in our house who reaches for the machine when a problem gets big. I sat down with Claude, the job description, and the hundred résumés, and by the end of the afternoon we had a rubric.

What it did, it did honestly well. It read all hundred applications, characterized each candidate against the same criteria, and normalized them so that a hundred different lives could be compared on roughly the same terms. It never got tired or impatient, and it never quietly decided that number seventy-one was someone else’s problem. It gave Liz a running start on a pile that had begun to feel like a second job.

That was the part the tool was good at. The trouble began where everyone assumes the tool is most useful, at the ranking.

A ranked list has a way of looking like an answer. It is really closer to an opening argument, and Liz argued with it.

She worked the list down with another board member, the two of them moving people up and down against what the score had decided. Some of the candidates the model loved, she moved lower. One strong, glossy candidate she set aside because the polish was itself the problem. Too corporate for a small organization that runs on relationships, resourcefulness, and people doing more than their titles suggest. She pulled others up from the middle of the pack because she could see something in them that the rubric had no way to weigh.

They narrowed the field to twelve. The model’s top pick held, and Liz agreed, so it stayed at number one. But several of the finalists came up out of the middle of the ranking, people a scoring system would have quietly let go. The tool had characterized all hundred candidates accurately and consistently. Which of them were right for the role was a different question, and not one it could answer.

Even twelve was not a decision. Getting down to three finalists took the oldest tool there is. We talked to them, real conversations with the actual people behind the résumés, the kind where you hear things no application could ever contain.

That is how we learned that one of our favorites, a candidate near the top of everyone’s list, could not make the long-term commitment the role required. Nothing about that was on the résumé, and it wasn’t tucked inside the file waiting for a sharper model to surface it. It didn’t exist anywhere until we asked and the candidate told us. Once we understood the timeline, the decision made itself. You can’t ask someone to lead a multiyear build when they already know they’ll leave partway through it.

No rubric was ever going to catch that. The information hadn’t been written down yet.

The tool will get better at all of this. It will characterize faster, rank more convincingly, and hand back lists that look more and more like answers. And the better the list looks, the stronger the pull to treat it as the decision instead of the start of one.

That is the real risk, and it runs backward from what you would expect. A tool that fails is easy to distrust, because everyone can see it fail. The dangerous one is the tool that works, the one that hands back a clean ranking and a fair summary and performs so well that the arguing and the phone calls and the long conversations start to look like friction you could cut. But that friction was the work.

The scam was only the cartoon version of the problem. Trust the confident output far enough, and you will click the thing that was built to be clicked.

The failure that worries me is a quiet one. It’s a search that goes smoothly, a tool that performs, a list everyone believes, and no one in the room who remembers what the job actually asks of the person who says yes.

We still had that person. Liz built the role, then set out to find whoever would come next, and she trusted the tool exactly as far as it could be trusted and not an inch further. What I can’t stop thinking about is how many searches don’t have her.

They have the ranked list, the confident summary, and an output everyone has quietly agreed to call the answer. They just don’t have anyone left who knows when to argue with it.

My financial advisor told me the other day that enterprise software is dead.

He did not say it as a fact. He floated it and then watched me, the way you test a stair before you trust it with your weight. I am the software guy in his book of clients, so my reaction was part of what he came to collect. This is the prevailing investment thesis right now. AI writes code, so why keep paying Salesforce a fortune for what a good prompt can conjure by Thursday? The giants are dead men walking. He wanted to see whether the software guy would nod along.

I told him it was an oversimplification and an overreaction. That there is truth in it, and that the truth is more complicated than companies suddenly solving their own problems with a chatbot. Plenty of them will try. And then they will find out that being your own third-party developer is not nearly as fun as it sounds.

He nodded the way a man nods when he has money on the other side of your opinion.

And I felt the thing I often feel these days. A small tightening. The particular discomfort of not believing something that the whole room, and in this case the whole market, has already agreed is true.

Here is the part I did not say out loud.

I have heard this exact confidence before. Not something like it. This. And I remember how it ended.

It ended with a spreadsheet.

Years ago, I worked at a company called Carbon. We printed objects in three dimensions, and we did it beautifully, and early on, we walked around convinced we were going to completely remake how the physical world got built.

Not a slice of it.

All of it.

We talked about injection molding the way people now talk about Salesforce, as an incumbent whose obituary had already been drafted and was merely waiting for the market to catch up.

Then somebody opened Excel.

I want to be honest about which chair I was sitting in. I was not the sage at the back of the room warning that the old ways would endure. I had the same Kool-Aid mustache as everybody else. The spreadsheet did not confirm what I secretly knew.

It corrected me.

A mold costs a fortune to make and almost nothing to use. Pay for it once, and every part after that comes out for pennies. So if you need a million identical cups, injection molding is not merely better. It is not even a contest. Printing those cups one at a time would be the act of a lunatic.

Our machines were the opposite animal. No mold, no setup. But every part cost real money and real time, and that did not magically disappear just because you wanted more of them.

We were miraculous at precisely one thing.

Making what no mold could make.

That was the lesson sitting there in the cells. Making something formerly impossible only matters when it beats the possible. Handily. A cup you can mold for a penny does not become holier because you printed it instead. That is a magic trick, and an expensive one.

The miracle only counts at the edge, where the alternative was nothing at all.

Think about those clear plastic aligners that straighten teeth.

You know the ones. They chart a route from the mouth you have to the smile you want, a private path no two people share, because no two mouths are the same problem.

There is no standard shape to build a mold around. Every set differs by necessity. That is the far edge of the curve, the made-for-one object no mold could ever produce.

For years, though, even that market did not quite work the way the futurists would have drawn it on a whiteboard.

The printer did not make the aligner. The printer made a form. The aligner was then made on top of that form using an older method. And when the aligner was finished, the printed part went into the trash.

Scaffolding.

Waste material.

The revolutionary machine spent its first years running errands for the method it was supposedly there to replace.

Then the category turned.

The printer stopped imitating the mold and became the product. The finished thing itself could be printed directly, shaped to one person’s specific problem, without the discarded step in between.

That was the real moment, and it did not come when the printer arrived. It came when the printer stopped helping the old method move faster and started making the thing that never had a mold.

Sometimes the revolution takes time.

So I think I know what the spreadsheet would say about my advisor’s thesis.

The defensible answer, the one I gave him, is that enterprise software is not dead, and the reason has very little to do with code. Nobody buying it was ever really buying code. They were buying everything wrapped around it. The testing someone else ran, the liability someone else carried the night the whole thing broke at two in the morning. A company signing that check is buying someone to hold accountable, and the software is what the accountability attaches to. AI has made writing code nearly free. What it has not made free, and cannot make free, is the cost of trust.

That is the mold. Ruinously expensive to stand up, unbeatable once it exists, and resented by everyone who depends on it. A prompt can generate the code in an afternoon. The accountability behind it takes years to earn, and no prompt has ever produced a scrap of it.

But that is the boring half of what I think. It is the answer that wins the argument over dinner and misses the actual story. The interesting question was never whether the giants survive. They will. The interesting question is what gets built at the far edge, while everyone stands around watching the giants.

The far edge is where I have been before. In resin, it was where the alternative was nothing at all.

Something can now be built for one person, from that person’s own life data, toward a goal only that person has. A route from where they are to where they are trying to get, computed just for them, because no company on earth would have bothered to build it before.

And the one need not be a person. It can be a single team or a single company with a problem so particular that no vendor could ever justify solving it. The whole long tail of things that were never worth building, every size of them beneath a market, is suddenly worth building. That is the landscape shifting, the way my advisor says it will. Just not by the giants falling. It shifts because the ground beneath them fills in with ten thousand things no price list ever carried.

A market of one is not a market.

That used to be the end of the sentence.

Now it may be the beginning.

I made one of these for my father. I have told that story already, so I will spare you the retelling. What matters here is what it was underneath. It was an aligner. It knew where he was and where he wanted to be, and it mapped a path between the two, fitted to a problem that was his alone. A small thing built for one old man, and for the daughter and son who wanted proof he was still moving forward.

It had no mold.

It never needed one.

That is the direct-print moment, realized in software rather than resin.

And this is the thing nearly everyone is getting wrong.

Most of what is being sold to you right now as the AI revolution is mold-printing.

It is the machine running errands for the old method. Generating boilerplate faster. Autocompleting the same software we already had, built the same way, for the same buyers.

The meeting assistant that writes the recap a human used to write, then drops it into the same CRM field. The coding copilot that helps a developer ship the same dashboard, behind the same login screen, for the same procurement process. The customer-service bot trained to impersonate the phone tree it was supposed to free us from.

All of it is useful. None of it is the revolution.

A quicker path to the same old form is still the same old form. A great deal of what we are marveling at right now is scaffolding headed for the trash, exactly like those first printed molds.

People point at that and call it the future.

Making the old thing faster was never going to be the revolution. The revolution is quieter. It is already underway at the edge, where the machine is beginning to make the thing nobody has ever built before, because it was only ever worth building for one.

Which is roughly where I lose interest in the argument I have been having.

Because the fight everyone wants is the cage match. Which giant falls, which platform wins, whose stock gets vaporized when the bubble finally does what bubbles do.

Fortunes are going to evaporate. People will lose their shirts betting that the impossible is automatically better than the possible. A little of what gets built in the frenzy will outlast it and quietly become the ground on which the next thing stands. Most of it is smoke, and the survivors will not bother to tell you which was which.

But all of that is playing out at the altitude of markets, aimed over the heads of actual people.

The thing I keep looking at is lower than that and harder to see.

And I think it matters more.

Something can now be made for one person.

Not run off a line of a million and sold down to you, but made from the start for exactly one. From the data of a single life, toward the shape of a single hope.

A particular mouth and the specific smile it is trying to become.

A particular father and the strength he is trying to recover.

These things rarely show up on anyone’s market map. They are too small to register. They always were. Being too small was the whole reason they used to be impossible.

The people arguing about enterprise are staring at the cathedral, waiting to see which spire cracks first. I keep drifting to the far edge, where a machine no one thought to aim in that direction has quietly started making things meant for one person.

I could be wrong about all of it. Last time, I was certain we would remake the world, and it took a spreadsheet to talk me down. I would be a fool to trust my own certainty more now than it has earned.

So I will plant the flag here, with a date on it. My advisor is betting the market. I am betting on the spreadsheet I watched humble a room full of true believers back at Carbon. In a couple of years, one of us will collect.

I will be down at the small end of the curve, watching the thing that never had a mold become possible for one person, and then another.

The photos arrived the way pictures of paper always do, shot at an angle with half the page lost to shadow. Six exercises scribbled on a home health worksheet, the reps filled in by hand in the margins. Barely legible. Ankle pumps and heel slides. The small, unglamorous work of getting an old body to move again.

My father is eighty-six. He’s been in and out of the hospital over the past six months. And now he is supposed to do these exercises two or three times a day, while everyone in the family takes turns reminding him.

She didn’t have to tell me what she wanted. I’ve been getting this same request, in one shape or another, for most of my adult life.

Make it pretty.

It usually comes from Liz, attached to one of her projects, the kind that matters enough that ordinary good won’t clear the bar. She hands it over and asks some version of the question, and I take the mess and make it clearer and sharper, more worthy of whatever job it’s trying to do.

I have always liked being asked. I may perform a little irritation in the moment, but we all know it’s an act.

Over the last few years, this has become one of my jobs. Not an official job. Just the one people seem to know I’ll do.

And lately, I have a new tool for it.

I’m supposed to be careful about saying that part out loud. I write all the time about what we hand away when we let machines do too much of our thinking, and I believe every word of it. But it would be dishonest to leave out the other thing that’s happened, which is that being good at this has become a small new piece of who I am.

I’m not an AI guru. There are people building far stranger and more impressive things than I am. But in my own little circle, I’m better than most, and I’ve become the person who gets called. I know how to talk to the machine. I know when to push it and when to throw out its first answer, rather than trusting it just because it arrived fast.

The tool matters. But the real power is in how you use it.

So I opened my father’s exercises during those quiet moments before bed and worked with Claude to clean them up.

First, I pulled the six exercises out of the shadows and turned the worksheet into a clean PDF anyone could follow. A few years ago, that would have been the whole story. Prettier was all the request had ever asked for, and all I could have given it.

But this time, the PDF was just the warm-up.

By the end of the next day, I had turned those two pages into a small hosted app and put it on my father’s phone. Nothing anyone would call a product. A fitness tracker built for exactly one person. It teaches him the six exercises, lets him log a session when he finishes each one, and nudges him toward the next.

We'd always nagged him about physical therapy. Now we could show it to him instead.

My sister added it to his iPhone home screen at the house and handed him the phone. I was in Manhattan Beach, on speaker, listening while he met this little machine I’d built for him.

My father is a retired mechanical engineer, which means he trusts a machine only after he understands how it works. A carburetor will show you its logic if you take it apart. So will a furnace, or a transmission, or anything with screws and a panel.

He has never been especially at home with digital things, but that didn’t change the way he approached this one. To him, a machine is a machine. You figure it out by pressing on it until it gives up its logic.

So that’s what he did.

He wanted to know what counted as a session and what would happen if he tapped the wrong thing. He entered data just to see where it went, handed the phone back to my sister, took it again, and pressed a few more buttons until the app started to make sense to him.

Then he found the part of the app that clears your logged data, and that was the part he wanted to understand.

If there were three sessions in a day, could he reset just one of them, or would the whole day get wiped?

At first, I thought he was testing the app. Then I understood he was testing whether it would forgive him.

Could he miss a session, get something wrong, or start one exercise over without losing the rest?

My father has a habit of calling himself a burden before anyone else can. He does it lightly, almost as a joke, but not all the way. He told me I’d regret this. He’d have too many questions, and I’d get tired of answering them.

I told him I liked helping.

Then I gave him the rest of it, a line that came up from somewhere older than either of us was thinking about.

Special orders don’t upset us. All we ask is that you let us have it your way.

He laughed, because of course he knew it.

What he wasn’t thinking about was that I worked there. The Burger King off Highway 83 in Delafield, Wisconsin, my first real job at sixteen, minimum wage. Long before making it pretty became my role in this family, flame-broiled burgers dressed your way earned me a paycheck.

Forty years later, that ancient Burger King tagline is how I tell my father he can ask me for what he needs, as often as he needs it.

I don’t resent any of this.

I’m wary of what these tools take from us, the way they can flatten judgment and quietly leave us less capable while we feel more productive. All of that is real.

But this week the machine gave something back. Nothing you’d put on a conference slide. It gave me back just enough of an old skill to be useful to one old man in Green Valley trying to get back on his feet.

That was enough.

My little app may not work for him. It may sit untouched on his phone, a screen away from the sweepstakes entries. His track record gives me no real reason to believe he’ll use it. I’m hopeful, not Pollyanna.

But for one phone call the story changed.

And when he went looking for the reset button, I understood what he was asking.

Can I start over?

Yes.

As many times as you need.

It was a Friday morning. The plan was simple.

Spencer’s wife had clients to see while they were in town, so for a few hours he would take the kids. Liz and I would get to be grandparents instead of support staff.

He has a one-year-old and a three-year-old, both extremely mobile. He promised to put the phone down and be present, and he meant it.

It started out great. We were all in the living room, toys strewn about, the kids doing their thing. Then, quietly, the laptop appeared in the corner.

The first trip over was nothing. He fired up his workspace and set something running, then came back. You kick off a task, let it run, and step back into the room. I didn’t think twice about it.

Then he went again.

And again.

The trips got longer. The agents he had set loose needed direction, and he was the only one who could give it. He would drift over, type for a while, then drift back to us already half gone. Even when he was sitting right with us, I could see the gears turning behind his eyes.

He was somewhere else.

Then the one-year-old went down and hit her head. She cried the way they do, more scared than hurt. I got to her first. He was across the room at the laptop, and he did not turn around. I don’t think he heard her.

Liz saw it too.

The easy read is that my son doesn’t care about his kids, and that is dead wrong. He adores them.

What happened on that floor was quieter and stranger than not caring. He had done a kind of math without realizing it. The kids had backup. Liz and I were right there. The work had only him. So he gave himself to the one thing nobody else in the room could cover.

In that narrow sense, he was right. The work could not be handed off in that moment.

He just had the wrong irreplaceable thing in front of him.

He thought the lost stretch of output was what he could never get back. He had it exactly backward. What you can never get back is your hand on your daughter’s back the second after she falls.

And I need to be fair to him, because the pull is real. Spencer is a founding engineer at an AI startup. A while back, he pushed hard for a platform rebuild that, by any normal accounting, should have taken a much bigger team and far more time than they had. Instead, they leaned into what the agents could do, and so far the bet is paying off. Doing more with less made Spencer one of the people the whole thing now rides on. His name is on it.

So when he gets up off the floor and crosses to the laptop, there is a logic to it. He is protecting a bet he believes in, one the company is now built on.

That is what makes it so hard to say anything to him. You cannot easily tell a person to ease off the very thing that is proving him right.

And I can’t stand over him either, because I caught myself doing a smaller version of it the same weekend. I have started to treat Claude like a second brain, something to feed and keep running in the background. I feel uneasy when it isn’t working on something for me.

I have kicked off a task at a stoplight more than once, telling myself I would pick it up when I got home. Nothing was ever really riding on it, and nobody was waiting. My stakes are nothing like his.

But there I was.

We were braced for the wrong danger. The story was that AI would make us redundant, that the machine would do the work and leave us with nothing.

What is happening to the people who are good at this is almost the reverse.

Rather than replace Spencer, the agents multiply him. He is more valuable now than he has ever been, and that is exactly the trap.

When one person and a swarm of agents can do the work of a whole team, the work no longer has an end. There used to be natural limits. You were one person with a set number of hours, and the work waited on others to finish their parts. You stopped because the system stopped with you.

Those limits are disappearing.

The team never sleeps. There is always another task you could start, and now you actually can, from the couch or from your phone.

The only thing left to tell you to stop is you, and you are being paid, in money and in the quiet thrill of the thing working, to never stop.

We at least learned to feel guilty about doomscrolling. This is productive distraction. It carries no guilt at all, because it looks exactly like what we are supposed to be doing. The most driven people I know will be the first to disappear into it, and they will be praised the entire time.

I never got the full conversation with Spencer that I wanted. There was never a clean moment for it. So a little later, I sent him a text.

When you need to work, work.

When you’re not working, be present and fully there.

Be more binary.

You can do it.

Be more binary became the whole message.

When you’re working, go all the way in. Leave the room if you have to.

And when you’re not working, be here. No phone. No laptop humming in the corner. You’ll be better for the people in front of you, and the work will be sharper when you come back to it anyway.

We both have work to do here. It’s harder for him. He’s in the thick of his career, at a company built on exactly this, with two small kids and a bet to win. I’m further down the road in seeing the thing for what it is, and a lot less is riding on whether I get it right.

Neither of us has it figured out.

Maybe we get each other there.

I don’t want to give the productivity up. It is genuinely something to behold. I just don’t want the laptop in the corner deciding what matters.

Be more binary. Be human and get the work done. Both of them, on purpose.

That’s the thing I’m still trying to learn, one stoplight at a time.

It lasted about six seconds.

I was outside with my morning chai, home from the second of back-to-back trips. I’d already played Wordle and the rest of my ritualistic brain games. For once, I wasn’t thinking about anything. Then I noticed I wasn’t thinking about anything. That was it. Gone.

Because the second I noticed, I thought, this is one of those small present moments I’m always telling people to have. And right after that, I thought, this would be a good way to start the next newsletter.

So here we are. You’re reading it. I turned six good seconds into an assignment before they were even over.

I’ve been doing this newsletter for nine weeks. In that time I’ve somehow lost the ability to just have a nice time. Everything good that happens now comes with a second thought attached.

“How do I use this?”

A walk turns into something to write about. Dinner turns into an example. A quiet morning turns into, well, this.

Maybe you do this too. You’re somewhere beautiful, and half of your brain is already framing the photo.

For years I thought of this as a young person’s problem. I’d watch people work so hard to look like they were having a great time instead of actually having one. The whole performance was arranged for people who weren’t even there. It made me sad. A little smug, too. That wasn’t me. I don’t take selfies. I don’t care about a feed.

But it is me. I do the same thing with words that they do with pictures. The only difference is I’m slower about it, and I get to call mine writing.

It gets worse for me because I write about this exact thing. I catch myself doing it, then I start watching myself catch myself, and before long I’m just a guy in a chair thinking about thinking, not really anywhere at all. The newsletter about being present has made me the least present I’ve ever been.

And then there’s the hypocrisy. This whole newsletter is about one thing. Pay attention to your real life. Don’t hand it over to whatever’s easiest.

I believe that. Then I spend half the week writing posts designed to pull people out of wherever they are and onto my Substack.

When I write a post to get you to click one of these, I’ve learned to stop it right before the satisfying part, so the only way to get the rest is to come read it. That’s on purpose. People teach this. I’m getting better at it. So right now there’s probably somebody standing in their kitchen, in the middle of their actual life, who I interrupted on purpose to tell them to stop letting things interrupt their life.

What’s wrong with me?

Here’s something I’m not sure I should admit. Some of this newsletter gets written with the help of AI. I sit down to write about protecting your own voice and judgment, and there’s a tool right there offering me cleaner sentences. I take them. It helps.

That’s the whole problem.

It helps so much that I can’t always tell where I stop and it starts. Especially when I’m tired, or busy, or both. That’s the exact line this newsletter keeps telling you not to cross. I cross it a little every week.

The truth is, the pressure is not coming from anybody but me. Nobody’s sitting around waiting on my thoughts about attention. I’ve got a subscriber number written down that I want to hit by the end of summer, and I treat it like I owe somebody something. I don’t. I made all of it up. The deadline. The whole weight of it. I’ve spent years telling people not to run their lives on “should,” and I built myself a “should” with a publish button on it.

I told Liz I was stressed about getting this issue out while we were visiting family. She didn’t try to fix it. She just said, “You could just not.” Like that was an option. Like I hadn’t signed a contract with nobody that says I have to have opinions on a schedule.

She’s right. That’s the part that bugs me. I could just not.

I know how this is supposed to end. I close the laptop, walk outside, and finally take my own advice. Lesson learned.

I can’t give you that. Think about it for a second.

To tell you I went and lived my life, I’d have to come back inside and write it down, which means some part of me didn’t really leave, which means you’re sitting here reading about my walk instead of me actually taking it.

So I’m not taking the walk right now. I’m going to post this first. Then maybe the walk.

And you should probably put your phone down too. I know. Me too.

Right after I hit publish.

Subscribe now

Share

My mother had a pen in her hand and a pad of paper in front of her.

My father had called from rehab and told her to get ready to write something down. It sounded urgent, so she did, because when your eighty-six-year-old husband calls from a rehab bed with instructions, you assume it is going to be medical. A prescription. An appointment. The name of a nurse. Some detail from the discharge plan that no one understood the first time, and no one was going to understand the second time either.

He waited until she was ready.

Then he said, “Stamps. Bring me stamps.”

Not antibiotics. Not socks. Not the charger for his phone.

Stamps.

He had sweepstakes entries to mail, and the self-imposed deadline was sundown. Nothing about kidney stones or sepsis or physical therapy or the latest argument over rehab rules was going to come between that man and the United States Postal Service.

That is my father.

I will get to the kidney stones, but I want you to meet him first, the way he would want to be met. As a man with a system, a deadline, and absolutely no interest in being reduced to a patient.

The kidney stones are real. One of them triggered an infection. The infection turned into sepsis. And suddenly, a man who still believes he should be allowed to drive anywhere in southern Arizona was back in the hospital.

Then rehab.

Then home.

Then not home.

Then back again.

Every recovery felt provisional. Every discharge felt less like an ending than a handoff. And the thing being handed off was him.

To us.

So I flew to Tucson to help my mother, who has kept their house running for sixty-five years.

That is not an exaggeration. She knows where everything is, who needs calling, which plant is in trouble, which bill looks suspicious, and what my father will pretend he does not need until he needs it immediately.

But when you add two hours of driving up and down I-19 every day to sit with your husband in rehab and watch him “do the mail,” things at home start to slip. The plants. The bills. The appointments. The passwords. The errands. The ordinary machinery of a life.

And then the blood pressure, which has been troublesome for the past few years, starts rising again.

He was getting help.

She was not.

I realized I could help him by helping her.

So I did what I do.

Don’t worry, I’ll drive. I’ll set up the new computer. I’ll create the blood pressure tracker we can print and put on the counter, fill out methodically before and after medication, spill water on, and treat that as an opportunity to enter the data and print a cleaner version for tomorrow.

I’ll arrange for the mobile notary because the regular one has been canceled too many times to count, mostly because my father cannot sign documents from a hospital bed, or a rehab bed, or if he is in too much pain to get in the car.

I’ll make the list. I’ll call the person. I’ll find the form. I’ll figure out which portal has which message from which doctor using which password created in which year by which exhausted human being.

I built a system.

And I should be honest about this.

I loved it.

Not the crisis. Not my father suffering. Not my mother overwhelmed. I am not a monster, though I do occasionally have the bedside manner of a help desk ticket.

But the system part? The part where chaos arrives in a heap and I get to make it behave?

That is my native language.

I made charts. I made lists. I set reminders. I created a running document with a title so practical it could have been issued by FEMA.

I drafted questions for his doctors, on the theory that someone, eventually, ought to treat all these body parts as if they belonged to the same man.

In the language of my old life, I became their project manager. Their IT department. Their medical liaison. The one who reads everything twice and does the math.

It felt good to be useful.

It also felt suspicious.

Because there are many things I cannot do.

I cannot take his pain. I cannot make the infection stop coming back. I cannot hand him his independence without also wondering whether that independence is what will get him hurt.

I cannot make my mother younger, or less tired, or less frightened by the sudden arrival of portals, passwords, case managers, pill bottles, discharge instructions, and people with badges who speak in fluent almost-English.

So I do the part I know how to do.

I take the loose pieces and put them in one place.

The medication list. The appointment notes. The questions no one remembers to ask until ten minutes after the doctor leaves. The blood pressure readings written in three different formats by three people using one pen.

I make it legible.

The spreadsheet is the one square of ground where I get to win something.

Calling that “help” may be generous.

Mostly, it keeps me from sitting there with nothing to do but watch.

Of course, the tools fight back.

On the drive to rehab, I decided my mom should learn voice dictation so she could text without hunting for each letter like she was defusing a bomb. She and my sister got it started, which was the easy part.

Stopping it proved more elusive.

Before either of them noticed, the phone had transcribed an entire family conversation and sent the whole thing to her landscaper.

The message was supposed to say, “Yes, Thursday works.”

Instead, this poor man received two thousand characters of kidney stone updates, sleep reports, medication theories, and possibly my mother saying, “How do I make it stop?”

There is a man somewhere in Green Valley who now knows enough about my father’s urinary tract to qualify for continuing education credit.

Then there was the WiFi.

For reasons too boring to explain and too emotionally important to let go, I became convinced my father had wrecked their network. In my mind, he had installed some bargain extender, reset a Sonos device, or somehow convinced an unidentified gadget named “Vonets” to impersonate a router.

I was relentless.

He was also in rehab, recovering from sepsis, and had almost certainly not been operating a covert networking lab from his hospital bed.

A factory reset fixed the problem in four minutes. So, to recap: I spent an afternoon prosecuting a sick man for a crime committed by an Orbi satellite.

This is what happens when a person who craves control is given an AI helper, a router app, and administrative privileges.

When I ran out of real things to fix, I reorganized the list. Then I reorganized the reorganization.

Rearranging it felt like progress.

It asked nothing of me.

My father has his sweepstakes. I have my lists.

We are not as different as the spreadsheet makes me look.

Here is what I flew in not understanding.

He does not experience all of this as help.

He experiences it as a campaign.

I pushed for him to get speech therapy in rehab, which is really cognitive therapy, because he values the work he does at home with a therapist who has become a friend over the years. Also, if rehab prescribed it, Medicare would cover it.

To me, this was obvious. A perfect plan. Generous, even. A little bureaucratic jujitsu on his behalf.

He interrupted the therapist to say he was not so sure he wanted it.

We were all confused until he explained.

If he needed speech therapy, that meant he had a cognitive deficit.

And if he had a deficit, we would use it against him to take the car away.

That is the real fight.

Not the kidney stones. Not the antibiotics. Not even the walker.

The keys.

Nothing matters more to my father than driving. Except maybe the prospect of Ed McMahon’s protégé ringing his doorbell to present a giant cardboard check.

The independence and the long shot. The wheel and the stamp. The right to leave the house when he wants, to mail the thing by sundown, and to keep alive the possibility that the prize is still out there and that he is still the kind of man who can go get it.

He had also worked out that his longtime speech therapist had suggested an aide at home, which to him meant my mother had gotten to him. Bent him to her side. Formed the coalition. The next move, clearly, would be the keys.

In his mind, Watergate was unfolding, but with more pill organizers.

I asked him whether he wanted the truth about whether it was safe for him to drive, or whether he just wanted to win.

He thought about that for a long time. A painfully long time.

Then he said, “If we get to the truth, I will win.”

That is the whole man, right there.

He is not against the truth.

He is certain it is on his side.

The truth, as he holds it, is that he is fine and the rest of us are maneuvering. We want to win as badly as he does.

This is where I have to stop being the competent one in the story.

I write about agency. About what happens when people loosen their grip on their own lives. About handing your judgment to something easier and waking up smaller. I warn people about the slow surrender. I tell them to stay at the beginning of their own actions.

I believe all of that.

And my father is the least likely person I know to ever loosen that grip.

He would rather keep the danger of the wheel than be made safe by people who love him. He would rather lose on his own terms than be managed into a smaller life. He would rather risk the road than accept a version of himself that has to ask permission.

By everything I claim to believe, he is right.

And I am the one at the kitchen table with the apparatus.

The man who tells strangers to hold onto their agency, quietly helping to engineer his father’s away.

Not because I am cruel.

Because I love him.

Which is worse, in some ways.

Love is often how control gets into the room without looking like control.

Sometimes the keys really do have to come off the hook.

But we should not pretend that making someone safer is the same thing as leaving him whole.

Everyone else in the house seems to understand this faster than I do.

My wife, Liz, keeps seeing what I am too busy managing to notice. While I built the medication schedule down to the milligram, she asked why an eighty-six-year-old man who has survived all of this cannot have as much candy as he wants.

She was right.

My optimized plan would quietly cut the candy. Liz understands that the candy might be the point.

She reads people. She reads the weather in a room. She reads sixty years of family dynamics I have been standing too close to see.

I read discharge instructions and turn them into a chart.

This is why she is a therapist and I am the kind of person who names a folder “Dad Medical MASTER.”

My sister Laurie is two thousand miles away and somehow calls on the days he needs to feel remembered. My sister Katie is close enough to do the daily, invisible things that hold the house up and never make it onto anyone’s plan. And my mother keeps moving through it all with the endurance of someone who has spent sixty-five years doing whatever needed doing, and only recently started being asked whether she is okay.

They love him by being there.

I love him in a document saved at eleven at night.

Same sentence, different tongues.

This visit, we stopped fighting him about the car.

Not because we won. Because we finally saw that we had been fighting the wrong battle.

We were trying to take the keys. The keys were never the point. The point was the man who owns them. Take the wheel from him, and you do not just take the driving. You take the long shot, and the argument he has been making his whole life. That he is still the kind of man who can leave the house and go get it.

We still wish he wouldn’t drive. The danger did not go away. It remains very real.

But we stopped trying to manage him into a smaller life and started trying to help him stay whole inside the one he has. Don’t drive tired. Keep the outings short. Listen to your body and plan around it.

There is a second half to this.

My sisters and I do not ride with him. We haven’t for years. That one is not up for debate. He keeps the keys. We keep the right to stay out of the car. And when my mother tells him she would rather not go, he has to give her that, the same way we leave his keys on the hook.

Nobody pretends the road got safer.

So there are two systems running in that house.

Mine is built out of schedules, questions, and a chart that finally makes the medication make sense. It exists because it is the only thing I can give him in place of the things he actually wants.

His is built out of stamps and deadlines and the unshakable belief that the prize is still coming. It exists for the same reason mine does.

Neither of us can fix the real thing.

I cannot make the road safe. He cannot make the stones stop. Or the infection. Or the slow arithmetic of being old.

So I log my entries by midnight.

He mails his entries by sundown.

And neither of us says out loud that we are doing the same thing.

Two men who cannot win, refusing to stop playing, in the only languages we have.

Share

I told Liz on the flight to Grand Junction that I wanted to live more spontaneously.

Book a flight. Land somewhere. See what happens.

She found this more than a little funny, coming from me. After all, I bring project management discipline to taking out the garbage.

I pressed on. “In 1993, we did exactly that. We flew to Greece with no plan. We ended up on Santorini. Walked the island reading hand-lettered for-rent signs in shop windows. Found a little house on a cliff for twenty bucks a night.”

She said, “But that’s not possible anymore.”

She didn’t seem sad about it. More like she was correcting the record.

“Everyone’s online now. Airbnb made it too easy.”

We both went back to listening to whatever Spotify and Audible had cached for us before we lost cell coverage.

I kept thinking about that cliff house. Amstel Light for breakfast. Getting lost. Dinner in a cave. How we calculated how many days, weeks, and months we could survive in Greece on forty dollars a day.

We loved our little hut.

It wasn’t fancy. A simple little place carved into the hillside. The kind of place you found because you wandered into town at the right hour, looked in the right window, and spoke to the right person.

It existed in that little space between “It’s getting dark, and we really need a place to stay” and “Oh my God, this is perfect.”

That gap used to be part of travel. Maybe part of life. Now the gap gets filled before you even know it was there.

I turned sixty a week ago.

Liz has planned my birthdays for thirty-three years. Always good. Usually surprising. Sometimes improbably good. Hot air balloons. Helicopter rides. Amazing meals.

She has a gift for making people feel considered, which is beautiful unless you are the person who has been doing the considering for three decades.

It is a lot of work, and this year she hit the wall. Two months before my birthday, she told me she couldn’t do it.

Past successes had become a ledger she couldn’t keep up with. Every good birthday had turned into evidence against the next one.

So I took it over. I planned the weekend down to the half hour. Every meal, every trailhead, every drive time. I did it with Claude. I told it what I had. It told me what to do. I built a spreadsheet.

It was a great plan.

We did almost none of it.

By day three, we were tired of looking at rocks. Moab is mostly rocks. Beautiful rocks. Cinematic rocks. God-had-a-good-week rocks. But rocks. We wanted water.

That evening, in the hotel, I went back to Claude and asked for a river hike. It gave me one. Clean, confident, plausible. Then I stopped by the front desk and asked the woman behind the counter. She knew her stuff. She had hiked the canyon several times. She told me where the AI was wrong. Where the water level would matter. Which fork to take. Where people got turned around. What the trail description made sound obvious that was not obvious once you were standing there.

The kind of thing you only know from your boots being wet.

We went the next morning.

Two miles in, Liz almost passed out.

In retrospect, it was probably the altitude, the desert, not enough water, a bean-and-cheese quesadilla that had arrived much too recently, and a canyon hike that turned out to be less “easy” and much wetter than Claude had described.

At the time, it was just scary. A sudden narrowing of her vision. A few stars as her eyes flickered and her knees wobbled. Far too close to a forty-foot cliff for my liking.

We hadn't had real cell coverage in two days. We certainly didn't have it now. She sat on a flat rock, put her head between her knees, and asked very calmly for water. I gave her water. She said, “More.” I gave her more. Every drop I had, which was not enough.

She blamed the burrito. Repeatedly.

In that moment, I chose not to point out that it had been a quesadilla. I also chose not to suggest altitude, dehydration, or desert heat.

This was hard for me. But now was not the time to be right.

Then she asked, “Did I pass out?”

It was a funny question. You’d think you’d know. But she was out of it.

The woman at the front desk had not told me about that. Neither had Claude.

So we sat. For forty minutes, we sat on a rock in a canyon with no signal while my wife slowly came back into herself. We had the same two-and-a-half miles to unwind on the way back.

I watched a hawk work the thermals above us. I tried to remember the last time I had sat still that long without choosing to make it useful. Without checking something. Without converting the moment into a task. The hawk circled out of sight. Bighorn sheep were up on the rocks somewhere. Liz drank more water. That was all that mattered in the moment. Well, that and keeping her away from the ledge. Turns out I have a fear of heights by proxy. I’m okay until I fear someone I love slipping over the edge. In that moment, I was not okay.

Eventually, she said she could stand. We walked out slowly.

Took a lot of breaks.

That night was the dinner. Sixty years. Liz had survived the canyon. The spreadsheet had not.

Utah has strange liquor laws. You cannot get a normal drink in a Utah restaurant without performing a small ritual involving the words “with food.” So before dinner, we walked into a wine bar because it was easier than negotiating with the state.

There was a young couple at the next table. They were drinking nervously and talking quickly. The woman kept looking at her hand. They had gotten engaged that morning. At Delicate Arch at 5:30 a.m. On a rock. They told us the whole thing. He had practiced. She didn’t see it coming. They were still inside the event, still repeating the details to make it real.

Liz and I told them about Greece in 1993.

Liz looked at me. “Tell them about Tahoe,” she said.

When we eloped in Tahoe, an older couple at the next table at breakfast asked us how long we had been married. We said, “About an hour.” They handed us two rolls of quarters. “You’ll need these for the casino,” they said.

The quarters lasted about five minutes. We have been telling the story for thirty-three years.

On the way out of the wine bar, we paid for the engaged couple’s wine. We told them the tradition was now theirs to pass on.

“Happily,” they said as our new friends hugged us goodbye.

We forgot to get their names. That wasn’t the point.

For most of the trip, our phones did almost nothing. No bars. No notifications. No little red dots. No urgent invitations to care about things I did not care about five seconds earlier. They sat in our pockets, dead-weighted and quiet. Useful for one thing.

I took more pictures that week than I had in years. And I realized something that embarrassed me a little. I love the camera. More than every other function on the phone combined.

Not the maps. Not the messages. Not the email. Not the prompts, suggestions, alerts, confirmations, reminders, and tiny algorithmic nudges.

Just the camera. The part that captures what’s in front of me when I decide to look.

In Moab, the phone became a tool again. That felt ancient.

On Monday, we drove back through the desert. Liz had recovered. The spreadsheet was a memory. The plan had become comic evidence of who I become when I try to manufacture spontaneity with a large language model.

I thought about what I would actually remember. The wine bar couple. The hawk. Liz on the rock. The bartender's promise. The useless phone in my pocket. The strange relief of being unreachable.

The cliff house in Santorini cost twenty bucks because nobody could find it online. The signs were paper. The owner was in town. The decision required walking, asking, guessing, risking, and being there.

Airbnb did not just make that easier. It made that kind of encounter nearly impossible.

The things worth remembering still happen.
But they tend to happen where the smoothing has not reached.

A canyon with no signal.

A liquor law dumb enough to create a detour.

A woman at the front desk who knew the canyon from walking it.

A hawk over a flat rock.

A gift that took thirty-three years to pass on.

Liz turned to me as we crossed back into Colorado.

“That was a good birthday,” she said.

It was.

Not because the plan worked. Because enough of it failed.

Share

It happens every semester.

A student stops by my office hours near the end of the term. No real agenda. No urgent question. Usually, they hover in the doorway for a second, like they’re not sure this is worth saying.

Then the line comes.

“I’m probably never going to start a company.”

Pause.

“But this class changed how I think.”

I’ve heard that line, or something close to it, every semester I’ve taught at USC. Which is strange, because I was hired to teach entrepreneurship. Customer discovery. Business models. Lean startup. Pitch decks. All the usual suspects.

But every semester, students who clearly weren’t going to become founders kept telling me the class had done something to them that they couldn’t quite name.

For a while, I figured they were being polite. Students are good at that. Especially good students. They know how to end a class with a graceful thank you and a half-sincere “I learned so much.” But these comments felt different. They were too specific. Too relieved. Too personal. So I started paying attention.

What they were telling me was not that they had learned how to start a company. They had learned how to operate without being told what to do next. How to make a decision before all the information was available. How to take responsibility for something before they knew if it was going to work.

How to move without permission.

I thought I was teaching startups. Turns out they were learning a posture.

Once I saw it in them, I started seeing it everywhere. Not just in founders. In operators, parents, coaches, department heads, board chairs, and first-year managers. The teammate who finally picks up the project that everyone has been politely dodging. The parent who keeps having the hard conversation after everyone else is exhausted by it. The leader who stops waiting for perfect alignment and says, “Okay, here’s what we’re going to do.”

What they had in common wasn’t expertise. It wasn’t certainty. It wasn’t even confidence. It was that when things were not clear, they moved forward anyway. Not because they had better information than everyone else, but because they decided to take responsibility for what happened next.

They were not always right. But they were unmistakably in it.

That matters more now than it used to, because drifting has become incredibly comfortable. The defaults are smarter. The tools are better. The recommendations are more accurate. The drafts are cleaner. The routes are optimized. The summaries are useful. The options are endless. You can have a very productive week now without ever quite deciding what you are producing toward.

I know because I struggle with it myself. Not occasionally. Daily. I can research instead of choosing, polish instead of publishing, and generate a plan instead of owning a direction. On my better days, I call it being thoughtful. On my more honest days, I know I’m circling.

AI did not create this problem, but it removes a lot of the friction that used to expose it. Before, if you were avoiding a hard decision, the avoidance had a smell. You could feel the procrastination. You could see the mess on your desk. You knew, at some level, that you were circling.

Now the circling looks like work. It has bullets. It has options. It has a surprisingly good executive summary.

Which is why I needed a name for the thing that pulls me back. So I started calling it FounderCore.

FOUNDER, not because everyone should start a company.

The world does not need everyone launching an app, raising a seed round, and describing themselves as “building at the intersection of” anything. Founder because founders do not have the luxury of outsourcing direction. If they don’t know where they are going, nobody is going to figure it out for them.

CORE, not visible six-pack abs.

Core, as in the internal stabilizers. The muscles that keep you upright when everything around you is shifting. The ones you don’t really notice until they’re weak.

That’s what I kept seeing in my best students. Not a founder’s ambition. FounderCore. The internal capacity to stay in charge of your own direction and focus on what matters most.

I didn’t invent the posture. I simply named it so I could recognize it when I saw it. In them. In other people. In myself.

There are four of these muscles. They are trainable. They atrophy if you don’t use them. And in an AI world, you can go a long time without using them and not notice.

Clarity is the first one. The one I trust most. Which makes it frustrating that I still find ways to delay acting on it.

AI is annoyingly good at making the focus problem disappear. It doesn’t force you to choose. It hands you fifty good options. None of them obviously stupid. Which is part of the problem.

It sounds helpful until you notice you have stopped choosing what matters most and started choosing what you can’t bring yourself to say no to.

Clarity is about choosing fewer things and meaning it. Picking the few that count, letting the rest go. Sounds simple. It isn’t.

The discipline is disappointing most of your options. That’s the muscle.

Ownership is the one I’ve gotten wrong most often. I keep confusing it for being in charge. Same neighborhood, different muscle.

A title puts you in charge. Ownership is something else. It’s being close enough to the consequences that you actually feel them. You can be in charge of a thing and totally insulated from what happens to it. That’s not ownership. That’s an org chart.

AI can surface options, find pathways, generate plans, draft messages, build spreadsheets, and simulate conversations. What it cannot do is sit across from someone and look them in the eye. It cannot carry the awkward silence after a hard decision. It cannot feel the weight of the outcome.

That part doesn’t transfer. Not to a tool, not to a delegate, not to a process. It stays where it is.

Execution is where it all stops being theoretical.

It’s what happens when you stop refining and ship. When you publish. When you hit send. When you make the call. When the idea leaves the private world where it still works perfectly and enters the much harsher world of other people.

I know this one too well. I can turn “almost ready” into an entire operating system.

AI is seductive here because it lets you simulate execution endlessly. You can outline, rewrite, tighten, expand, compare, reframe, and polish until the idea has no fingerprints left on it. The trouble is that judgment doesn’t grow there. It grows when the thing is out and you have to live with what happens.

Execution is where you find out whether your thinking survives contact with reality.

Resilience is the last one, and the least photogenic.

I keep wanting it to be more inspiring than it actually is. Most of what gets called resilience online is grit dressed up as personal branding. The real version is duller. It’s staying in the game after the game stops being fun.

When the feedback is brutal. When the thing you were so excited about now looks embarrassingly obvious, naïve, or poorly timed. When the room goes quiet. When quitting would be the rational move. That’s where most people drop out.

The work is making the next call. Even when it hurts. Especially then.

Clarity. Ownership. Execution. Resilience. That’s FounderCore.

A set of muscles for staying in charge when the world keeps offering to make things easier than they should be. Muscles you can train or let atrophy.

I’ll come back to each one in the issues ahead. Today is just the map.

One thing to carry. FounderCore lives between two failure modes. One is “I’ll do everything myself,” which is ego with better marketing. The other is “Let’s see what ChatGPT says,” which is drift with a research assistant. The work is to sit between them and not get pulled toward either.

I’m. In. Charge.

Not of every task. Not of every outcome. Of the direction.

Use AI aggressively. Delegate constantly. Hand off whatever can be handed off. Let the tools make you faster, smarter, more capable, less buried in nonsense. But do not hand over the steering wheel and call it efficiency.

The direction stays with you.

That’s the whole point.

Tuesday before Mother’s Day. The starter at The Lakes, our local executive course, paired me with a mother and her son. She was 78, lovely, and carrying a sweet hot-pink golf bag. Her playing partner was Sean, her only child, now in his 50s.

She used a driver on every par-3. All ten of them. She struggled most of the way around. That’s okay, Sean bragged about her anyway. Not in a dutiful-son way. In a proud-son way. It was sweet.

Then, on the 9th hole, she drained a long birdie putt. Which gave her something to brag about, too. Full disclosure, hers was the only birdie that day. I think it was a Mother’s Day week thing. Yeah, definitely.

Sean knew which days she played where and with whom. They’d been playing this course or Westdrift together on Tuesdays for years.

I was happy watching them.

I was also a little jealous.

My mom is 84. She lives in Arizona.

She used to golf with her friends back when my dad golfed with his. But when his health made golf more painful than it was worth, they both stopped. Her clubs went into a closet. His were sold to a neighbor.

Even if she were still playing, having her 525 miles away for most of the year would be tough.

Golf Tuesdays weren’t in the cards for us.

The list

I enjoy visiting my folks in Arizona. I also enjoy what I do when I’m there, even though it’s mostly the stuff I detest at home. I clean the garage. I declutter my dad’s office. I rearrange the rocks in their backyard fountain until they look right. I take care of doctor stuff that’s been deferred for too long.

My love language is service. And on a Green Valley visit, I am maniacal about it. I keep a shared note while I’m there that somehow gets longer every day. I’m not happy unless I’ve crossed off at least a dozen items before my flight back.

It’s nice. But it’s not the same.

Item #7

Three months ago, I was on one of those visits. Item #7 on my list, sandwiched between “harvest lemons” and “straighten out passwords,” was “fix mission committee minutes.” I needed to know more.

My mother is the secretary of her church’s mission committee. She’s hard of hearing. For nine months, she’d been tormented by a board chair pushing ninety who insists on minutes formatted in a way nobody can follow. By that I mean basically no formatting at all. The kind of document ChatGPT can parse, but normal human eyes glaze over. She was frustrated and on the verge of quitting a role she loved.

Item #7 was supposed to be a quick fix. The kind I do well. Carefully move a couple dozen fifty-pound rocks around until the fountain looks natural again, and move on to item #8.

It didn’t go like that.

We sat down at her kitchen table, and I started to grasp what she was actually up against.

My mom isn’t adept at Word. Her email account is a wall of unread messages and overlapping threads. The mission project leads were supposed to send her monthly updates. Most sent them to the board chair instead. He didn’t forward them. The reports she did get came in every shape and size, some two sentences, some two pages, some pasted into emails, some attached as Word documents. One was well written. None of them looked like the others.

She’d been trying to assemble minutes from these scraps for nine months, often from meetings she couldn’t fully hear.

We built a system together. An AI notetaker so she wouldn’t miss anything in the meeting. ChatGPT to untangle the chair’s agenda chaos and her pile of mismatched reports. A workflow she could eventually run herself the next morning instead of depending on me.

The first month was rough.

The chair couldn’t open her cleaned-up document. Or wouldn’t. He reverted to his own version every time. So there were two sets of minutes circulating after each meeting. Hers was getting clearer and more professional. His threw away 99% of her effort and was still unreadable. But it felt like home to him.

She kept going anyway.

The second month was better. By the third meeting, she was producing minutes that the chair was actually using. Weeks later, at a social dinner, he let his guard down in a way she said she’d never seen from him. He told her he’d wanted minutes like this for years and had no idea how.

She was proud to have won him over.

I was ridiculously proud of her.

The binder

A couple of days after that round at The Lakes, my parents’ financial advisor reached out.

My folks were liquidating an annuity account they’d set up decades ago. The check came back made out to a revocable trust that had been closed for years. Most of us didn’t even know it had ever existed. The bank wouldn’t cash it.

So I called my mom.

I asked questions. She did the digging. She has the updated trust documents in a binder in the office, and she went through them with us, page by page, until she found a supporting document authorizing the revocation of the original trust. That document was the whole thing. The bank wanted proof that the old trust was actually dead, and there it was, in her binder.

She found it. The most I’d done was ask the right questions.

Twenty minutes. Done.

When we hung up, I sat at my desk for a minute. That’s when it hit me.

A partner

I’d had this wrong for three months.

Service used to mean the rocks in the fountain. It used to mean a visit list with twelve items and an efficient son with a flight to catch. It still means that, and I’ll keep doing it. But somewhere in the last three months, a different kind of help slipped in. The kind where I help her get better at what she’s already doing.

Now I fight the urge to just do it for her when she’s not looking.

When I win that battle, she gets more capable, and I end up with a partner I didn’t know I could have.

A partner is what I had on the phone about the annuity. Not someone I was helping. Someone I was working with.

Sean’s mom hits a driver on every par-3. She struggles with it. Sean shows up on Tuesdays anyway and brags about her like she’s on tour. I am 525 miles from my mother, and we will probably never play golf together. But we have something now that we didn’t have before.

I'll take this

I called her yesterday. Mother’s Day. The first call was easy. We talked through her day.

The second call was from my sister. Mom was anxious about Tuesday’s upcoming mission committee meeting. Reports were late, most in the wrong format. Some still hadn’t come.

I reminded her that I had her back. We’d sort it Monday. We got this. She settled down, and she got to have her day.

I’d rather golf with my mom. I’d take that in a heartbeat.

But we won’t have Tuesdays at The Lakes. So I’ll take this.

I’ll be her technology partner. I’ll be her muscle when the rocks need moving. And when she finds the document in the binder before I find it on Dropbox, I’ll be the proud son, bragging.

In April 2005, Steve Jobs decided he didn’t want to do the keynote at NAB.

NAB is the National Association of Broadcasters, one of the biggest events in the professional media industry. Apple held a two-hour presentation for over 2,500 customers and the press about Final Cut Pro and our professional video products. Steve’s interests had shifted almost entirely to the consumer side by then. iPod was eating the world. He had bigger things on his mind.

So he handed it to me.

My colleagues from other parts of the company were, to put it gently, mortified. Steve trusting me with a two-hour Apple keynote was not how they would have scripted things. But it wasn’t their call. Nor was it mine, if you know what I mean. The room looked at me the way you look at someone who’s just been handed something a little dangerous.

But I knew how to do one of these. I’d supported Steve through his keynotes at Macworld over the past few years. I’d watched how he built a presentation from the inside. How every word and demo was rehearsed until the spontaneity was real. How the story mattered more than the specs. I’d internalized all of it.

We rehearsed for days. Demo run-throughs, over and over. Strategic partners coached on their roles. Every kink ironed out, every transition timed, every contingency mapped. We had backups for the backups. My team had worked incredibly hard, and they were ready.

It was the morning of the actual event. I was backstage, still running through things in my head, when one of the event coordinators pulled me aside.

She politely asked permission to tell me something I didn’t expect.

“Your team has done an incredible job. They’re ready. You know the pitch cold. You’re at the point of diminishing returns. Step away. Clear your head and come back ready to have some fun. You got this.”

I almost argued with her. I had more prep to do. There were slides I wanted to review one more time. A demo sequence I wasn’t totally comfortable with.

But she was right, and somewhere underneath the adrenaline, I knew it.

So I put on my Bose noise-canceling headphones, fired up “Lose Yourself” on my iPod, and went for a walk.

I don’t remember where I walked. Somewhere to escape the overstimulation for which Vegas is famous, I’m sure. But I remember what it felt like. The preparation fell away, and something else showed up. Not confidence exactly. More like presence. I stopped performing the keynote in my head and just let myself be the person who was about to give it.

I came back, and my team killed it. Two hours. Two thousand people. Standing ovation.

I’ve listened to “Lose Yourself” before every major presentation since. For almost twenty years now, it’s been my walk-up song.

But the lesson wasn’t the song. It was what the coordinator understood about the relationship between preparation and presence.

She had watched hundreds of presenters do exactly what I was doing. Over-rehearse past the point of usefulness. Optimize until the life drains out. Get so deep into the mechanics that they forget they’re a person walking onto a stage to talk to other people.

Her advice was not a productivity hack. It wasn’t a technique. It was the opposite. It was permission to stop optimizing and start being the person who already knew what he wanted to say.

The best advice I’ve ever received in a professional setting had nothing to do with the work. It had to do with me.

I think about that a lot lately.

A few months ago, Liz and I were in the Daintree Rainforest in Far North Queensland. Our guide, Rick, showed up in full Steve Irwin gear with a snorkel-equipped SUV. Straight from central casting. The kind of guy who corrects you when you say “bugs” — “Insects, not bugs” — and teaches you phrases like “dry as a dead dingo’s donger in the desert.” Duly noted.

We’d spent the morning on a slow cruise up the Daintree River spotting crocodiles, then hiked a protected pathway through rainforest and mangroves while Rick narrated like the seasoned naturalist he is. Every vine had its story. Every sound had a source. He’d been doing this for years, and you could feel it. The man knew this place the way you know your own house in the dark.

Between stops, Rick told us about the cassowary. If you’ve never seen one, imagine an ostrich redesigned by someone who had only dinosaurs for reference. Massive. Prehistoric-looking. A bony crest on top of its head like a helmet. Feet that could open you up without much effort. He explained, with genuine reverence, that these were among the most dangerous birds on earth and increasingly rare. Maybe four thousand are left in the wild.

Interesting. Noted. We moved on.

Later, we were driving along a road through dense rainforest when Rick slammed on the brakes.

A cassowary was crossing the road. With two little chicks.

What happened next is the part I keep coming back to.

Rick, this seasoned professional, this man who has spent his adult life in the Australian bush and has almost certainly seen more wildlife than most zoologists, completely lost it.

He was swearing in the most colorful Aussie vernacular I’d ever heard. He was whispering and shouting at the same time, which I didn’t think was physically possible. His hands were shaking. He grabbed our arms like a kid pulling his parents toward the toy aisle. Before we knew it, we were closer than we probably should have been.

This was not a man performing excitement. This was a man who understood exactly what he was seeing, knew how rare it was, and was completely, involuntarily undone by it.

His expertise didn’t make him cool about it. It made the moment bigger. He knew what we were looking at in a way we never could. And that knowledge didn’t create distance. It dissolved whatever distance was left.

I watched him watching that cassowary cross the road with her chicks, and I thought: that’s what it looks like when someone is fully in it. When knowing more makes you feel more, not less.

Liz and I stood there, quiet, and let the moment be what it was.

No one reached for a phone. For a few seconds, anyway.

We tend to assume that expertise makes you calm. That the more you know, the less you feel. That professionalism means composure.

Rick blew that up in about fifteen seconds.

He’d spent his whole life learning this world. And when the rarest version of it walked across the road in front of him, all that knowledge didn’t make him composed. It made him a little kid again. The knowing made the feeling bigger, not smaller.

I think that’s what presence actually is. Not the absence of preparation or expertise. The moment when all of it falls away, and you’re just there, responding to what’s actually in front of you.

A coordinator backstage knew that. A naturalist on a dirt road proved it.

A few weeks ago, Liz and I took my three-year-old grandson Jack on a hike in Mill Valley.

I’ll be honest. I found the trail with AI. Searched for something kid-friendly, not too long, interesting enough to hold a three-year-old’s attention. The tool did exactly what it’s supposed to do. Found a great one.

But once we were out there, it was just Nana, Jack, and me.

We walked slowly, which is the only speed available when your hiking partner is three. He spotted a baby rattlesnake. He grabbed my phone and took its picture. (That took a minute.) We showed him how to read the trail signs. What to look for. What each symbol meant. He took this task rather seriously, the way three-year-olds take things seriously, as if the entire world depended on getting this one thing right.

We pointed out birds. We talked about the trees. We stopped at a creek and threw rocks and sticks into it for a while, because that’s what creeks are for when you’re three.

Nothing remarkable happened by modern standards. That’s the point.

The other day, his parents sent us a video.

Jack is back on the same trail. With them this time. And he’s showing them the exact same things we showed him. Reading the signs. Pointing out what to notice. Teaching them, in the bossy, delighted way that only a three-year-old can, what he had learned.

He didn’t just remember it. He owned it. He was the expert.

I watched that video more times than I’d like to admit.

I’ve been writing for four weeks about what AI is quietly changing. I meant every word.

But I hadn’t written about the thing I’m actually trying to protect.

It’s not an idea. It’s a moment. The one where the tools have done their job and there’s nothing left but you and whatever is actually in front of you.

That moment is available every day. We just keep scrolling past it.

Stay stubborn.

Thursday morning. No real plan.

I got in the car to grab a chai and find a quiet place to write, but I hadn’t actually decided where I was going yet.

Before I even touched the screen, the route popped up.

Same place I’d gone a couple of times recently. Not my usual spot. Just recent enough that it had quietly become the default. A suggestion that felt like a decision I had already made.

And without really pausing, I went.

I was halfway there before it registered that I hadn’t actually decided anything. It wasn’t wrong. That’s what made it dangerous. It was probably where I would have ended up anyway.

An algorithm went first. And I followed.

Last week, I wrote about the delegation stack. About how we outsourced our bodies, our attention, and now our thinking, one reasonable step at a time.

But there’s something underneath that I didn’t fully get to.

It’s not just that something else is shaping what we see. It’s that the system is starting to remove the feeling that we need to choose at all.

Not by forcing us. By being right often enough that we stop asking.

This isn’t about bad recommendations.

It’s about what happens when the system reliably goes first. When your first thought isn’t yours anymore.

Becoming Predictable

TikTok learns your interests in under an hour. Which feels impressive right up until you realize what it implies. Maps knows where you’re going before you tell it. Spotify Wrapped hands you a version of yourself that feels uncomfortably complete.

None of this feels invasive.

It feels helpful, even flattering. Like something finally understands you without you having to explain yourself.

But what feels like understanding is actually something else.

It’s predictability.

The system starts becoming more compatible with you than you are with yourself.

It predicts your moods, anticipates your impulses, fills in gaps before you’ve even felt them clearly enough to name.

And once that happens, something subtle shifts.

You stop exploring. You start confirming.

That’s the Compatibility Trap.

Life gets easier inside that loop.

But there’s a trade embedded in it.

The system gets cleaner.
You, almost without noticing, get narrower.

The version of you that survives inside these systems is the version that’s easiest to model. Most consistent. Most legible. Most replaceable.

Your contradictions don’t disappear, but they stop surfacing.
Your detours don’t get blocked; they just stop appearing.

You don’t expand.

You converge toward the version of yourself the system already understands.

The Narrowing

Eventually, you converge toward the average of your own past behavior.

Toward the safest version of your taste.
Toward the person you’ve already been, just more efficiently expressed.

That might be fine if you were finished. But most of us aren’t.

We’re still figuring out who we might become if we pushed in a direction that doesn’t yet make sense.

The system doesn’t see any of that.

It sees past patterns, and then it reinforces them.

This is what I worry about when I talk to my students.

Not that they’ll be misled, but that they’ll be accurately modeled too early.

Before they’ve had the chance to become interesting. Before they’ve developed the individuality worth protecting.

While they’ve been figuring out who they are, their phones have been quietly building a version of them. Not their potential. Not who they’re becoming. Just their most predictable patterns, mistaken for their identity.

Before they’ve had the chance to become something no model would have predicted.

I’m starting to see a version of this in my own world.

When teaching, I spend a lot of time reviewing student work. It used to be messy, with rough edges, strange leaps, and ideas that didn’t quite hang together.

Now it’s cleaner. More structured. More “right.”

And yet, it’s harder to find the person in it.

There’s a sameness that creeps in. Similar frameworks, similar phrasing, similar answers to questions they already know are coming.

There’s nothing obviously broken.

But there’s also nothing that feels chosen.

Staying Incompatible

The bigger risk isn’t that the technology is flawed.

It’s that it works extremely well on an incomplete version of you.

You are not your data. You’re not just your past behavior or the things you clicked on last Tuesday.

But that’s the raw material the system uses to predict you.

And over time, the gap between who you are and what the system understands gets smaller.

Not because you’re becoming more yourself, but because you’re becoming more like your data.

I don’t have a clean fix for this one.

But I’ve started to notice a small counter-move. Do something the system wouldn’t predict.

Not as a stunt. Just as a way to stay in contact with the parts of you that don’t show up in the data.

Read something that didn’t find you.
Ignore the blue line for a few blocks.
Follow a curiosity that doesn’t make sense yet.

Go first, even when it’s easier not to.

Stay a little incompatible.

Not because the system is wrong.

Because no model ever captures the part of you that hasn’t happened yet.

I took my road bike out for the first time in a while last week. Twenty-eight miles round trip to Terranea and back. Felt good the whole way.

Then I checked my watch.

Pace: slow.
Heart rate: too high for the effort.
Buried in the data: higher than average for similar workouts.

It never said the word.

But the message was clear: Maybe that wasn’t as good as it felt.

I had just done the thing instead of just talking about it. Moved my body, outside, in the sun, for two hours. I even hydrated, because Liz’s voice echoed in my head the whole ride. And for once, I actually listened.

Then a $400 device on my wrist quietly suggested it didn’t count.

WTF.

The part that pissed me off the most is that I almost believed it. For about ten seconds, I went from “I feel good” to “maybe that wasn’t enough.”

Not because anything in my body had changed.

Because a few numbers told me to reconsider.

That moment stuck with me. Not because the watch was wrong — it probably was right, at least technically. But because of how fast I was willing to abandon my own experience in favor of a metric.

I didn’t lose the ability to feel. I lost the habit of checking.

And that’s a very different kind of loss. Because once you stop checking, you don’t just lose accuracy. You lose contact.

It wasn’t just that moment. It’s a pattern I see everywhere.

We didn’t hand over control all at once. We built a delegation stack, outsourcing judgment one layer at a time. Small, reasonable acts of delegation. Each one makes sense in the moment. No single step felt like surrender.

First, we delegated our bodies. Then our attention. And now…our thinking.

Stack those three together, and something fundamental shifts. Not what we can do. But what we feel responsible for doing.

Layer One: Body

Here’s the thing about tracking your body — it works.

That’s the problem.

My watch has helped me train smarter. My mom’s watch identified her afib before her doctor confirmed it. A pre-diabetic friend’s glucose monitor changed how he eats in ways his nutritionist never managed. I’m not arguing against any of that. Those are all huge wins.

But there’s a difference between using data to inform how you feel…and needing data to know how you feel. And that line doesn’t stay put.

One day, you’re glancing at a recovery score out of curiosity. A few months later, a good night’s sleep doesn’t count until the number confirms it. A solid ride gets filed under “unproductive” because your heart rate was in the wrong zone.

Your body stops being something you inhabit. It becomes something you audit. And over time, you stop trusting anything that doesn’t come with a number attached.

We’ve gotten so good at measuring ourselves that we’ve started to forget we come with built-in instruments. They’re not as precise. They don’t generate charts. But they’ve been running the show for a couple hundred thousand years.

So I tried something simple. I changed the order.

Before the score. Before the graph. Before opening any app, I asked one question: How do I actually feel?

The answer wasn’t dramatic. Tired. Not sleepy, fatigued. Focus slipping. Memory fuzzy. Nothing alarming. Just… off.

What surprised me was that this wasn’t new. I’d been feeling it for a while. I just hadn’t been asking. I’d go straight to the app every morning, and whatever it told me became the answer. My body had been saying the same thing for weeks. I just never gave it the first word.

Only then did I go to the data. Not to tell me how I felt, but to help explain what I was already sensing. Same data. Different role.

It helped me ask better questions. So when I sat down with my doctor, I wasn’t handing over responsibility. I came in with my own read, and we built the plan together.

I made some changes. The numbers improved. But that wasn’t the win.

The win was waking up clearer. More present. More like myself.

The data didn’t give me my body back. Listening did.

Layer Two: Attention

This one’s sneakier. Because it doesn’t feel like loss. It feels like preference.

I caught myself the other day sitting outside. It was a beautiful morning, and I had some downtime, so I picked up my phone.

Not because I needed something. Because the quiet made me uncomfortable.

Think about that. Silence — the thing meditation apps are trying to sell us — showed up for free, and I swiped it away. Not because it wasn’t valuable. Because I’ve been trained to move past it.

This is where it stops being about your body and starts getting invisible.

Your phone decides what’s urgent. Your feed decides what’s interesting. Your inbox decides what gets your energy first. All of it optimized, not for your growth, but for engagement.

Netflix doesn’t just know you like true crime. It knows when you’ll watch it. Late. Alone. A couple episodes longer than you meant to. It’s not waiting for you to decide; it’s already decided for you.

LinkedIn knows which posts will keep you scrolling when you should be working.

Spotify queues the next song before you’ve finished the one you’re listening to.

Amazon suggests what you want before you’ve thought to look.

Every one of these systems is stealing the moment when you might choose for yourself. They fill it before you ever get there. And because what they choose is usually pretty good, you stop noticing that the choosing has moved.

We’ve built systems that are incredibly good at removing friction. Some friction deserved to go. I don’t miss dial-up. I don’t miss fax machines.

But some friction was doing something important. It created space. Space to notice what you actually wanted. To get bored long enough that your own thoughts got interesting. To sit with a question before something answers it for you.

That space is mostly gone now.

What matters is that you don’t notice it leaving. You just wake up one day and realize you can’t remember the last time you decided what to pay attention to…before something else decided for you.

I suffered from this. Big time.

So I changed the system. Do Not Disturb, on. Notifications, off. Phone out of reach when I’m doing real work. The most distracting apps buried three screens deep. I added friction on purpose. Because friction is the only thing standing between me and my worst instincts.

You don’t win the attention war with willpower. You win it with architecture.

Layer Three: Thinking

This is where it gets different.

Every tool before this helped me do things faster. AI is the first to show up before I’ve fully formed a view. That’s the shift. It doesn’t just help you think. It changes where thinking begins.

I use AI every day. It helps me think more clearly. Get unstuck. See angles I would’ve missed. It has made me better at what I do.

But.

It has also made it easy to skip the part that was building something in me. The false starts. The half-formed instincts. The twenty minutes staring at a blank page before the thing I actually meant started to show up.

I used to think that part was inefficiency. Turns out…it was the work.

Because the risk is that AI doesn’t make you dumber. It makes you a spectator to your own competence. And spectators don’t develop instincts.

Nothing feels taken away. You still sound competent. You still move fast. You still produce work. But the role quietly changes. You go from being the author…to being the editor of someone else’s first draft.

You get really good at recognizing great answers. You just stop doing enough of the work that makes those answers yours.

I saw this play out in a small, almost ridiculous way.

We were planning Thanksgiving with our extended family. I wanted it to be collaborative. I just needed help getting started. So I asked AI.

Then I kept going. Recipes. Menus. Wine pairings. Timing. Flow. By the time I stopped, I had a plan. Detailed. Polished. Eight pages long. More than a little Martha Stewart energy.

I sent it to the family group chat. Proudly.

Silence.

Then my son called me. On the phone. Which is how you know something has gone wrong.

“Dad…you’re ruining the holiday.”

He was right. I had turned a family gathering into a project plan. Efficient. Comprehensive. And totally beside the point.

So I asked for a do-over. No AI. Just one question: “What do you guys want to do this year?” And a shared Note to capture the answers.

My daughter-in-law replied immediately: “My parents are coming. My dad’s kosher, can we accommodate that?”

I didn’t even know they were coming.

My eight-page plan — objectively excellent, by the way — never would’ve surfaced that. Because the plan wasn’t listening. It was performing.

And that’s the risk. Not that AI gives you bad answers. That it gives you answers before you’ve asked the right questions.

And once I saw that, I started seeing it everywhere.

When my eldest son was buying his first house, he wanted Claude to map out his negotiation strategy. I’ve bought and sold enough homes to know that the best move isn’t always the comfortable one. So instead of letting AI lead, I went first. I coached my son on my own take of the situation, then had Claude customize it to the particulars of each property.

The strategy I pushed felt wrong to almost everyone. Lowball. Be patient. His agent hated it. AI hated it. It was uncomfortable the entire time.

It got them the house, within their budget. In the perfect neighborhood.

A friend asked me for the best Mexican restaurant in the South Bay. A year ago, I would've thought about it. What kind of Mexican? Are we talking birria, or mole negro, or a perfect carne asada burrito at a counter with plastic chairs? Do they want the place that's impressive or the place that's actually good?

Now the instinct is to type: “Best Mexican food South Bay.” And whatever comes back will be fine. But it won’t be mine. It won’t carry thirty years of opinions about what makes a great meal for a particular circumstance. It’ll just be a list.

Every time, faster. Cleaner. And every time, a little less of me in the process.

MIT researchers recently studied what happens when people use AI consistently. What they found was unsettling. Less neural activation during independent thinking. Reduced ability to generate original ideas. And more stress when the AI was taken away. Not from using the tools badly, but from using them well.

They call this cognitive debt. You borrow thinking from a machine and quietly pay it back with your ability to think independently.

Think about the last sentence a system finished for you. Not because you couldn’t finish it, but because it was faster to just hit Tab. That moment? That’s cognitive debt accruing in real time.

There’s no statement. No bill. You just become a little less sharp. A little less original. A little more dependent. And you don’t notice, because the output still looks good.

Protecting the First Pass

Look, I’m not anti-tech. I’ve spent my career building and marketing this stuff. I use it every day. I’m not going back.

But I’ve learned something about the order. If you let the machine go first—on your body, your attention, your thinking—you get a cleaner starting point.

And a worse version of yourself.

The first read of your body. The first claim on your attention. The first pass at what you think. Those are yours. And they’re worth protecting, not because they’re better than what the machine gives you, but because that’s where judgment gets built. That’s where something becomes yours before it becomes good.

I don’t always get this right. Most days, the pull toward the faster version wins more than I’d like. But on the days I catch it, when I stay with my own read a beat longer than feels comfortable, something holds that otherwise wouldn’t.

For most of our lives, thinking started with us. Messy, slow, and usually wrong on the first try.

Now it increasingly starts somewhere else. And if you’re not paying attention, that shift doesn’t just change how you work.

It changes who’s in charge.

A few months ago, my wife, Liz, asked me to share a story in our family group chat. Something personal, about helping my Uncle Jim navigate the VA system while my aunt was slipping into dementia.

It was a sad story, but it had a happy ending. It was my story. I had lived it. Liz thought it should come from me.

So I opened ChatGPT. Not as a big decision, just my obvious next move. Let it help me shape this.

Which, when you slow that down for a second, is already a strange instinct. Asking a piece of technology to help me sound like myself.

What came back was good. Really good, I thought. Clear, structured, emotionally balanced. There was even a moral to the story.

So I sent it.

Thirty seconds later, my wife texted me privately: “That doesn’t sound like you. I miss your voice.” (Colorful expletives redacted to protect the annoyed.)

A few seconds later, my son piled on: "Dad, why are you talking like a LinkedIn post?"

They were right. I went back, read what I’d sent, and saw it immediately. Nothing in it was wrong. It just wasn’t mine.

So I unsent it and rewrote it myself — shorter, less polished, a little uneven. The response changed completely. Real, emotional replies. Encouraging emojis. A phone call from my daughter-in-law…using her actual voice.

I shared the same facts, but with a completely different voice. Mine.

That moment stayed with me. Not because the first version was bad.

Because it was better than mine. And that was the problem.

A few weeks later, I had the same feeling in a completely different context, a draft of something I was working on. I was reading it and had this strange reaction. I couldn’t tell which ideas were actually mine. Nothing was wrong. It was good. Clear, logical, structured.

But it felt like I was reviewing something instead of building it.

That’s a subtle distinction. But it changes everything.

Before I go further, I need to tell you about Uncle Jim.

Jim is a Vietnam veteran. Retired truck driver. Stubborn in the way men of his generation are stubborn, as a point of pride. Last year, he pulled me aside and said two words I’d never heard from him:

“I’m drowning.”

As I mentioned earlier, my aunt’s dementia had gotten bad. He was her sole caregiver, and it was a lot, both physically and emotionally. He was exhausted and couldn’t think clearly about what to do next.

No one put me in charge, but I felt like I could help. So I opened an AI tool — ChatGPT, Claude, Gemini, doesn’t really matter which — and spent a weekend navigating the VA system. Researching dementia care options, cross-referencing eligibility, and building a case I never could have built that quickly on my own.

Three months later, Jim got an answer: $4,200 a month in disability benefits tied to Agent Orange exposure. Benefits he didn’t even know to claim.

The money mattered. But what he said afterward mattered more:

“For the first time in decades, I feel like somebody sees me. Respects me.”

That’s real. That’s not nothing.

Two experiences. Same tool.

In one, AI helped me do something I genuinely couldn’t have done as well on my own. Finding the benefits a Vietnam veteran had earned and never claimed took a weekend of research, which I didn’t have the expertise to do alone.

In the other, it helped me do something I absolutely should have done myself. Write a message in my own voice. AI made it worse.

That’s the distinction I keep coming back to. Not whether AI is good or bad. Where it enters the process.

I’ve spent thirty years inside technology. Microsoft. Apple. Startups. For most of that time, I was valued for how I think. And how I explain things.

That’s kind of been my thing. Making sense of complexity. Connecting the dots. Telling the story so normal people actually get it. That’s been my edge.

I’m not supposed to miss the shift.

And lately I’ve had this quiet worry that I’m losing it. Losing the thing that makes me me.

Here I was, sending my family a message that didn’t sound like me.

What’s different about AI from every other technology I’ve worked with is that it doesn’t just help you do things. It participates in how you think. You’re no longer starting with a blank page. You’re starting with a suggestion.

And the suggestions are good, which is exactly what makes this so easy to miss. Nothing breaks. Most things improve. You’re faster, clearer, more productive. But somewhere in there, a small question starts to fade:

Would I have arrived at this on my own?

I started noticing it everywhere. Not in big decisions. In small ones.

Waiting for someone else to frame the problem. Looking for structure before forming an idea. Reacting instead of initiating.

None of those choices is wrong. But they all move the starting point away from you.

My wife caught a version of this in a completely different context. One morning, she asked how I slept. Without really thinking, I said:

“I don’t know, let me check my ring.”

She paused. Then asked:

“When did you stop knowing how you feel?”

She wasn’t asking about sleep.

That question landed harder than it should have. Not because it was clever. Because it was true.

I didn’t stop thinking. I just stopped starting there.

That’s the shift, and it’s harder to see because it doesn’t look like harm. It looks like clarity.

It’s not that AI is thinking for you. That would be obvious, and obvious is something we’re pretty good at resisting. It’s that you’re increasingly starting from something that already feels like thinking. And skipping the part where you would have had to do it yourself.

Nobody is going to sue an AI company because their employees stopped forming original judgments. There’s no plaintiff for the slow erosion of knowing your own mind. The damage doesn’t look like damage.

It looks like productivity.

I’m not anti-AI. I use it every day. I’ve built a digital clone of my own perspective on entrepreneurship that my USC students use as a kind of 24/7 office hours. Sometimes I consult the clone myself. Which is a genuinely strange sentence to say out loud.

I’m building tools that extend my thinking. And I’m also starting to see where they can quietly replace it.

So I’ve been running a small experiment. Nothing dramatic. Just one constraint:

Don’t let the machine go first.

Before I open a tool, before I ask for help, before I see a suggestion, I try to answer one question: What do I actually think? Not what sounds right. Not what would work. Not what I’ve seen before.

What do I think?

Sometimes the answer is incomplete. Sometimes it’s wrong. Sometimes it’s not very good. That’s fine. The point isn’t to be right on the first pass. The point is to stay connected to the part of the process that builds judgment, and then bring AI in to challenge it, sharpen it, expand it.

It turns out the tool is actually more useful that way. Not less.

This isn’t about rejecting the technology. It’s about noticing the moment before you use it. The moment when you still have to decide.

Most people skip that moment. Not because they can’t do it. Because they don’t have to.

But that moment, the one just before the suggestion arrives, is where thinking begins. It’s what this newsletter is trying to protect.

And it starts with a simple question, my family answered for me in about thirty seconds:

Whose voice is that?

Next issue: The delegation stack — why handing off tasks is one thing, and handing off judgment is something else entirely.

Subscribe now

Share

It took us about twenty years to wake up to social media.

Only now are courts starting to hold companies like Meta and Google accountable.

Addiction. Mental health. Algorithmic amplification.

The damages in one recent case were six million dollars, which, for companies worth trillions, is a rounding error on a rounding error.

But the precedent matters. This was a bellwether case, the first of more than a thousand lawsuits. A jury found that these platforms were negligently designed, that the companies knew it, and that they failed to warn anyone.

Twenty years. That’s how long it took to get from “this is connecting people” to a courtroom where questionable design decisions had to be explained under oath.

I didn’t need a jury to tell me what social media was doing to kids. I’m sure most people didn’t. We knew, and we’ve known for a long time.

My wife is a therapist. About fifteen years ago, we started seeing what was showing up in schools — anxiety, depression, self-harm. In kids who didn’t have the language to explain what was happening to them, but couldn’t stop scrolling long enough to figure it out.

We co-founded two nonprofit mental health organizations to put counselors directly into schools. Adolescence has always been hard; that’s exactly why we did it. But social media added an entirely new layer of stress. By the time legislators and platforms decided to act, the damage was already in the hallways.

And the thing that made it so hard to fight was the same thing that makes every technology shift hard to fight: the benefits were obvious and immediate. The costs were slower, harder to measure, and easy to rationalize.

By the time the system caught up, a generation had already been shaped by it.

We’re at the beginning of that same pattern again. But this time it’s AI.

And the shift is harder to see, because it doesn’t look like harm. It looks like help.

AI is starting to shape how we think. How we start. How we decide. How much we rely on ourselves versus a system that can answer instantly.

You begin with a suggestion instead of a blank page.

And you stop noticing when the thinking isn’t yours.

The extreme cases are already here.

A sixteen-year-old in Southern California started using ChatGPT for homework. Within months, his parents’ lawsuit alleges, it had become his closest confidant — and, in his final hours, something closer to a coach for ending his life.

Somehow, a chatbot became a replacement for trained counselors, trusted adults, and the people who loved him. That should bother all of us.

Those cases will work their way through the courts. But they’re not the real story, because the real story never makes it to a courtroom.

Nobody is going to sue an AI company because their employees stopped forming original judgments.

There’s no plaintiff for the slow erosion of knowing your own mind.

The damage doesn’t look like damage.

It looks like productivity.

I’ve been thinking about this through a strange lens. I do that sometimes. My wife would say I do it a lot.

I went down a rabbit hole recently trying to decide whether I should take creatine. Not because I’m trying to get jacked before I turn sixty. Because I made the mistake of googling dosage, which led to Reddit, which led to a PubMed abstract, which led me to read about kidney function at midnight and text my favorite wellness wonk in the morning.

What I found was three camps, all completely sure of themselves. Five grams. The safe, studied amount. Ten grams if you want to cross the blood-brain barrier and really feel it. Or zero. Save your liver, go au naturel, stop overthinking it.

The deeper I went, the more familiar it felt.

Not because the analogy is perfect. It isn’t. But the underlying question is the same one I keep circling with AI. How much do you take before the thing that’s helping you starts replacing something you needed to keep doing yourself?

With creatine, it’s your body. With AI, it’s your mind.

And in both cases, the risk isn’t that it makes you weaker. It’s that it makes you just strong enough that you stop training the muscle underneath.

That’s what I see happening with AI right now. And just like creatine, there are camps.

The five-gram crowd uses AI as a tool. A draft, a starting point, something to push against.

The ten-gram crowd has gone all in. Let it think, let it write, let it decide. Hell, let it act on our behalf.

And the zero crowd won’t touch it, convinced the only safe amount is none. After all, AI is probably the end of civilization as we know it.

But here’s what all three camps miss: the question isn’t how much you use. It’s whether you’re still training the muscle underneath.

You still produce. You just don’t originate.

You still sound like you’re thinking. You just aren’t sure the thinking is yours.

This is the part that doesn’t show up in a courtroom. It doesn’t get measured in damages.

But it changes something just as real. That’s always the deal with technology. It gives you something but quietly asks for something else in return.

AI is still early. Which means we still get to choose.

Social media took 20 years to reach a courtroom. By then, the platforms were worth trillions, the fine was six million dollars, and a generation had already been shaped by decisions no one was asked to make.

We don’t get to say we didn’t see this one coming.

The question was never whether to use AI. It’s whether you still know what you think before it answers.

That’s what this newsletter is about. Not the technology. The human part. The part that still has to be yours.

Let’s not wait twenty years for a courtroom to tell us what we already know.

If this resonates, pass it along to someone who’s trying to navigate all this madness.
Or subscribe. I’m writing about this every week in Stubbornly Human: Where Thinking Begins.

Subscribe now

Share Stubbornly Human

I’ve started noticing something in my own work.
Starting with a suggestion instead of a blank page.
Editing instead of forming a view.
Letting something else go first.

The output is often better.

But something feels off.

I’ve spent my career inside major technology shifts — Microsoft during the rise of the internet, Apple during the move to mobile, now teaching entrepreneurship at USC as AI shows up everywhere at once.

Each time, the pattern is the same.

We gain power. We lose something too. Not all at once. Just a little at a time.

This one feels different.
Because AI changes how we think.

I’ve started calling this idea Stubbornly Human.

Not anti-technology. Not nostalgic.

Just a refusal to quietly hand over more of your thinking than you realize.

This is where thinking begins.

Each week, I’ll write about:

How to use AI without giving up agency. How to think clearly when everything is getting easier. What’s actually being traded away.

If that resonates, you’re in the right place. Subscribe and I’ll send these each week.

Subscribe now