Is There an Optimal Way to Live?

Most people I know can tell you where they'll be in five years. They know which role comes next, which skills will get them there, which move makes sense on paper. Every step is calculated, and every step points toward the same destination: success.

I'm on a path too, and I take it seriously. But I've never been able to stay entirely on it.

By day, I work with data and models. After hours, I wander. I started a podcast with a friend because I can't stop turning questions over in my head. I taught myself video editing because I wanted to make things, not just consume them. And lately I've been using AI to write code and build my own products, because there's nothing quite like watching an idea turn into something real.

On paper, none of this has much to do with my day job.

"Focus," people say. "Pick one thing and go all in." I believe that. Focus is powerful. But a question kept nagging at me: is all this wandering a waste of time? Does every step need a clear purpose to count?

Put more simply: is there an optimal way to live?

Two questions hiding in one

The longer I sat with it, the more it split in two:

1. The destination. Is there such a thing as an ideal life, an optimal solution?
2. The route. Even if there is, which paths actually get you there?

The finish line that keeps moving

Start with the destination.

Psychologists have a name for a trap most of us fall into: the arrival fallacy. Fudan University economist Lan Xiaohuan brought it up in a talk on career choices, and it stuck with me.1 We tell ourselves we'll finally be satisfied once we land the offer, get the title, buy the apartment. Then we get there. For a moment it feels great. Then a new want shows up, and the finish line quietly moves.

If "optimal" means a final destination, it might not exist. The destination keeps moving.

The trouble with always climbing

Now the route. Even with a clear goal, walking toward it one sensible step at a time doesn't guarantee you'll end up in the best place. Computer scientists have wrestled with this exact problem for decades.

Picture this: you've been dropped somewhere in a vast, foggy mountain range. You can only see a few feet in any direction. Your goal is to reach the highest point.

The most obvious strategy is called hill climbing. Look around, take a step in whichever direction goes up, and repeat until every direction goes down.

It's simple. It's efficient. And it has one fatal flaw: where you end up depends almost entirely on where you start. Land near the base of a small hill, and the algorithm will march you confidently to its peak, then stop. Every direction leads down now, so as far as you can tell, you've made it.

That's called a local optimum: the best spot in the neighborhood, but not the best spot overall.

A lot of seemingly perfect life plans look like this. Every step is the highest option in sight. No single step is a mistake. And you still end up on whichever hill happened to be closest to where you began.

The investor Chris Dixon once used this very metaphor as a warning to people early in their careers.2 We all feel the pull to make the next step an upward one, he wrote, and ambitious people feel it most. That pull is how smart people end up climbing higher and higher on a hill they never really wanted to be on.

How algorithms get unstuck

So how do computers get around this?

Counterintuitively, by adding randomness on purpose. One common approach is to add a small random nudge to every step:

It looks intimidating, but it's simpler than it seems. x is where you're standing. Δ is your step uphill. ε is a random jump: you don't know which way it will go, or how far. And σ² is the variance of that jump, which is really just a measure of how wild you're willing to be.

With low variance, you mostly just climb. With high variance, every so often you leap somewhere unexpected.

The trick is timing. Early on, you keep the variance high. You roam. You explore. You accept that some steps will take you lower, because you don't yet know where the tallest mountain is. Once you've spotted it, you dial the variance down and climb with focus.

Explore first, then commit. That's how the algorithm escapes the local optimum. Try it yourself:

Maybe life works the same way. Without a map you can read in advance, a little more variance early on isn't a bad thing. Try a few more directions. Go somewhere unfamiliar. Talk to people who aren't like you. Leave room for possibilities you can't see yet.

A step that seems to take you down might be the one that reveals the mountain in the fog.

The best things weren't planned

Some of the most remarkable things in history came out of exactly this kind of randomness, often while people were chasing something else entirely.

In Why Greatness Cannot Be Planned, AI researchers Kenneth Stanley and Joel Lehman describe an online experiment called Picbreeder.3 Users could "breed" images the way you might breed animals: choose two parents, get a child with small random mutations, or pick up someone else's half-finished lineage and keep going. Everything started from simple shapes. Generations later, the site was full of striking images, a car and a dolphin among them, and almost none were what anyone had set out to create.

Stanley's own car started as someone else's alien face. A few generations in, the alien's eyes slid down until they looked like wheels, and he followed the accident. As he half-jokingly put it, without the alien there would be no car, and no book either.

You see this pattern everywhere. The vacuum tube wasn't invented for computers, yet it became the heart of the early ones. Before the iPhone, toughened glass, multi-touch screens and lithium-ion batteries already existed. None were built for a phone. All of them ended up inside one.

Steve Jobs told the most famous version of this story at Stanford in 2005.4

He dropped out of Reed College after just six months. Tuition was draining his working-class parents' savings, and he couldn't see what it was for. But he didn't leave. For another year and a half or so, he hung around campus as a drop-in, sitting in on whatever classes caught his interest.

One of them was calligraphy. Reed was known for its calligraphy teaching back then, and the campus was covered in hand-lettered posters and labels. So he signed up. He learned about serif and sans-serif typefaces, how the space between letters changes the feel of a word, what separates beautiful type from merely legible type. He loved it. And as far as he could tell, none of it would ever be useful.

Ten years later, his team was designing the first Macintosh, and all of it came back. The Mac shipped with multiple typefaces and proportionally spaced letters, one of the first personal computers to treat type as something beautiful. If he hadn't dropped out, he said, he never would have dropped in on that class.

Then came the line that has been quoted ever since.

You just have to trust that, somehow, they will connect.

My own dots

Looking back, I can see some of mine starting to connect.

In March 2024, a friend and I launched a podcast called Another Me, with a new episode every two weeks. To have something worth saying each time, I started reading and listening far more than I ever had. To keep it all from slipping away, I built a "second brain" in Notion: books, podcasts and stray ideas, each with my own notes attached. The raw material for this essay came from there.

Then that habit turned into a product. What I really wanted was one place to save anything I found online and come back to it later, with my own thoughts alongside. So I used AI to write the code and built it myself. It's called Slowkeep. And the storytelling and visual sense I'd picked up making podcasts and videos? They found their way into my products, and into this website.

Here's what surprised me most: none of it pulled me away from my main path. If anything, it kept feeding back into it. The podcast taught me to explain complicated things clearly. Building products made me fluent with AI tools and taught me to think from the user's side. And the data-driven thinking from my day job came with me into every side project.

Wander, but with a compass

Of course, adding variance isn't the same as drifting. There's a difference between wandering aimlessly and following your curiosity on purpose. Paul Graham's essay How to Do Great Work helped me see it.5

He describes two kinds of ambition: the kind you have before you've found what interests you, and the kind that grows out of that interest. Too much of the first, he argues, can make it harder to find what you actually want to do.

Since finding your direction depends so much on luck, the best you can do is improve your odds. The way to do that is to stay curious: try more things, meet more people, read more, ask more questions.

When in doubt, he says, optimize for interestingness. If a field doesn't get more interesting the deeper you go, it probably isn't yours. Instead of mapping out your whole future, at each stage do whatever seems most interesting and keeps the most doors open. He calls this staying upwind.

He also warns against thinking of time invested as a cost, a framing that will scare you off before you start, and against underestimating compounding: a page a day doesn't sound like much, but after a year it's a book. School trains us to answer questions someone else wrote. Real work starts with finding the questions worth asking.

And if he had to sum up great work in a single word, he'd bet on curiosity.

So, is there an optimal way to live?

Probably not. At least, not one you can calculate before you set out.

Honestly, I find that freeing. If there's no optimal path, the detours aren't waste. If the dots only connect looking back, then my job right now is simple: walk my main path well, stay curious, and keep collecting dots.

I've stopped needing every step to come with an answer.

Maybe the search is part of the answer.