Why I Build This: A Question From Sixteen

People ask why a behavioral scientist builds AI systems. The honest answer goes back to when I was sixteen and a health crisis with real risk took months out of my life. I'll leave the medical details out. What it taught me, I've never been able to put down.

The question that didn't fade

For a while I was angry. That part is normal, and it burns off. What stayed was something quieter and harder to shake: a life can be rewritten overnight by something you never chose. One week you have plans. The next week your plans have you.

When you're young and well, you assume your life is the things happening to you. Then something happens that you didn't sign up for, and the assumption breaks. You start watching your own reactions like a third person in the room. Why did I spiral there. Why did that small thing flatten me for two days. Why did the same news barely touch the person in the next bed.

The question I came out with wasn't dramatic. It was almost embarrassingly plain. How do you actually live well? Not how do you achieve, not how do you optimize a morning. How do you live in a way that holds when the circumstances turn on you, because eventually they do.

That question pulled me through a psychology degree, then a master's in clinical health psychology and wellbeing, then twelve countries and thousands of conversations. I kept thinking the answer would be complicated. It wasn't.

It was never the circumstances

Here is what surprised me after years of looking. The people who lived well were not the people with the easiest circumstances. Some of them had been handed brutal ones. And the people quietly coming apart often looked, from the outside, like they had everything sorted.

The difference wasn't what happened to them. It was the system running underneath the life. By system I mean four things working together, mostly out of sight: where attention goes by default, how the body handles stress, the story a person tells about what events mean, and the habits that run without permission.

Most of that machinery is automatic. You didn't write it. A lot of it was installed before you could vote, by your family, your school, the first few times something hurt. Then you spend decades acting it out and calling it your personality.

That's the part that's easy to miss. We treat the system as fixed and the circumstances as the thing to fight. So we change jobs, cities, partners, and within a year the same patterns reassemble in the new setting. The set changes. The script doesn't. Change the script and even an ordinary life starts to feel different from the inside.

Why the underneath is the real lever

This isn't a feel-good framing. The mechanism is well documented. Stress isn't only an emotion; it's a physiological state that narrows attention, shortens your time horizon, and pushes you toward old defaults. Under load, the brain spends less on deliberate thinking and more on the fast, automatic stuff. That's why you make worse decisions exactly when the stakes are highest, and why willpower fails on the bad day rather than the good one.

Habits run on the same logic. Research on behavior change keeps landing on the same finding: lasting change comes less from motivation and more from how the environment and the cue-routine loop are arranged. James Clear built a whole book on that one idea. Motivation is a guest. Structure is what stays.

And the story you tell shapes the body's response before any action happens. The way you interpret an event changes the stress reaction to it. Same event, different reading, different cortisol, different choice. That's not mindset fluff; it's measurable.

So when you work on the system underneath, you're not decorating. You're moving the thing that decides hundreds of small outcomes a day. That's the highest-leverage place to stand.

Why the usual advice quietly fails

Most self-improvement targets behavior directly. Wake up at five. Cold plunge. Journal. Twelve habits before breakfast. The advice isn't wrong, exactly. It's aimed at the wrong layer.

You can force a behavior for a few weeks on willpower. Then a hard month comes, the system underneath hasn't changed, and the behavior collapses back to baseline. People read that collapse as a character flaw. It isn't. It's an architecture problem. You bolted a new habit onto an old system and the old system won.

The other failure is treating circumstances as the cure. Get the promotion, hit the number, move somewhere with better weather. The relief is real and it's brief. Then the underlying patterns catch up, because they traveled with you. There's a name for this in the research, the hedonic treadmill, and most people walk it their whole lives without noticing the floor is moving.

The fix isn't more discipline or a better location. It's to go one layer down and change the thing that generates the behavior in the first place. Slower. Less impressive on the surface. The only version that holds.

What this looks like in practice

You don't need a crisis to start. You need to make the invisible system visible, one piece at a time.

Start with attention, because it's the easiest to catch. For two days, just notice where your mind goes the second it's free. The phone. A worry on loop. A grievance. You're not fixing anything yet. You're reading the defaults. Most people have never actually looked.

Then stress. Find the early signal in your own body, the one that shows up before you've named the feeling. Tight jaw, shallow breath, a short fuse. Catching it early is most of the work, because once the stress state is fully on, your good intentions are already offline. A slow exhale, longer than the inhale, will start to settle the nervous system within a minute or two. Simple, and it works on the physiology directly.

Then story. When something knocks you, write the sentence you're telling yourself about it. Read it back. Half the time the sentence is doing more damage than the event.

None of this is fast. You're rewriting code that took years to install. But the first time you watch yourself not react the old way, you'll feel the floor shift.

Why I build with machines now

So, the AI question. People assume there's a contradiction in a behavioral scientist building software. There isn't. It's the same project.

The insight from sixteen was that the system underneath a life can be changed, and that almost no one gets the help to do it. The understanding has mostly lived in clinics, books, and the heads of a few good practitioners. It doesn't scale. One person can sit with a few thousand people in a career. That's the ceiling, and it's a low one against the size of the need.

What I'm building at MARSA is a way to put that work in more hands. Not to replace the human part of change, which is the whole point, but to make the steady, daily support of it something a person can actually reach. The machine handles the patient, repeatable layer. The living stays yours. That's the line I won't cross.

The teenager who asked how to actually live well wasn't asking about technology. She was asking for a way out of feeling like a passenger in her own life. Everything I make is still an answer to her. If any of it gives you back the controls, it did its job. That's at marsa.ai, and it's the only sell I'll make.