In 2000, I was looking for a job. The dot-com bubble was leaking air, I had been in tech since I was eleven years old, and I had solid working knowledge of more things than I could name. Networks. Security. Systems. Business process. How organizations actually make money, which is different from how they think they make money. I walked into interviews with genuine depth across a dozen domains, and someone who knew more than I did about the world pulled me aside and gave me honest advice.
"Pick one thing," he said. "Learn SQL. Become a DBA. Or get your CCIE. Nobody hires a generalist."
He wasn't wrong about the market. He was right about what the market wanted. The advice was accurate, and the advice was also a trap, because what it actually meant was this. Shrink yourself until you fit the box they're hiring for.
I didn't take it. Partly stubbornness. Partly because I genuinely couldn't choose. And for years I felt the financial weight of that decision. While specialists built their one thing deep and commanded a premium, I was doing twenty things, some deep and some not, almost none shallow. Solid depth everywhere, confidence nowhere, because when you've only ever seen one thing, the person who did ten seems scattered. And when you've done twenty things alongside people doing one, you start to believe them.
The market said specialize or lose.
I'm here to tell you the market was wrong. And the timing of being wrong couldn't be better for you.
The thing nobody called it
What they called unfocused had a different name if you looked at it straight. It was synthesis. Pattern recognition across domains that don't usually talk to each other. The ability to sit in a room and hear not just what was said but what wasn't said, because the question that didn't get asked is almost always the question that matters most.
The generalist isn't the person who doesn't go deep. That's a misreading. The generalist is the person who can't stop going deep on anything that catches them. They're not shallow. They're stateful in every direction. Every rabbit hole they went down is still in memory, still load-bearing, still available the moment a conversation takes an unexpected turn.
In tech we talk about deterministic and non-deterministic systems. Stateful and stateless. And I've come to think those aren't just system properties. They're people properties too.
The specialist is deterministic. Same input, same output. Deep, reliable, fast within their domain. That is exactly what you want when the problem is known and the solution space is well-defined. When the fire is in the box they trained on, nobody puts it out faster.
The generalist is non-deterministic. They traverse. They synthesize from unexpected directions. They find the connection between something said in this meeting and something from a domain so removed that nobody else in the room would make the link. And that connection, the one nobody was looking for, is often the one that changes everything.
What this actually looks like
Last week I was in a conversation about early Black settlers in the United States who weren't brought as slaves. The Gullah people. Somewhere in that conversation I followed the word "juke" back to its origins. Juk, a Gullah word meaning rowdy or disorderly. Juke joints, the gathering places where people would dance, play music, be loud and alive. And from there, jukebox. The machine that filled the room with sound.
Here's the part I didn't look up. A friend of mine is the world's foremost expert on Rocknrolla machines, which are early predecessors to the modern jukebox. The name "Rock and Roll" itself may trace back through that chain. The chain I built in my head in real time from a conversation about history that had nothing to do with music technology.
Nobody asked me to make that connection. There was no strategic reason to pursue it. I pursued it because that's how I'm wired. The thing that interests me about a word is almost never what the word means. It's why that word exists at all, where it came from, and what the etymology reveals about how people framed the problem before there was a clean name for it.
That's not a skill you can teach in a course. It's a pattern of attention. An orientation toward the world that treats every subject as connected to every other subject if you pull the right thread.
The AI inflection
Here's the question that actually matters right now. When AI can answer the what and the how, when any specialist's knowledge is available on demand in seconds, what becomes scarce?
The question.
Not the answer. The question. More specifically, the assumption behind the question. The frame that was chosen before anyone started answering. Because AI optimizes for answers. It is extraordinarily good at giving you the best available response to the question you asked. It cannot tell you that you asked the wrong question. It cannot notice the assumption everyone carried into the room without checking. It cannot be the doubting Thomas who leans back and says, "Wait. Why are we sure about that?"
That's the generalist. That's always been the generalist. We just didn't have a world that made the value obvious, because the value was invisible when answers were expensive and questions were cheap.
The equation has flipped. Answers are now nearly free. The scarce resource is the person who knows which question to ask before the analysis begins. The person whose Wikipedia chain from Byzantine architecture to supply chain logistics is load-bearing when the conversation turns to resilience. The person who heard "juke" and followed it without being asked.
The world just built the generalist an unfair advantage. The question is whether you can see it.
Can you become one
I don't have a clean answer here. I think some people just have the hardware for it.
People with ADD and ADHD know exactly what I'm describing. The inability to concentrate when you want to, paired with deep fugue states of hyper-concentration when something catches you, means you rabbit hole compulsively, then surface, then find something else and do it again. That's not a deficit. That's a feature with terrible marketing. The same wiring that made school miserable and career advice useless is the thing that built a library across a hundred domains that nobody else in the room has read.
But I also think it can be taught. Or at least, encouraged. Mentored. Uncorked.
There's no stupid question. That's not a platitude. It's a discipline. You have to be willing to ask the question that reveals you don't know something obvious, because the person unwilling to do that is the person who eventually stops learning at the edge of their known territory. The generalist is the person who gets to the edge and keeps going. They ask again. They fail. They ask again. They are not afraid of looking slow, because they've noticed that the person who asks the question nobody wanted to ask is almost always the person who finds the thing nobody found.
If you have the instinct, protect it. If you've been told it's a liability, ignore that. If you've been burying it to fit into a job description, now is the time to stop.
The uncomfortable reality
The generalist was told for decades that the market didn't have room for them. In a world where specialist knowledge was scarce and hard to access, that was defensible. You needed the deep expert because the alternative was years of self-directed learning and most problems didn't wait that long.
That world is gone. The deep expert's knowledge is now a prompt away. What remains scarce is what was always scarce in the generalist: the synthesis, the pattern across domains, the question before the analysis, the connection built in real time from twenty rabbit holes that intersect at exactly the right moment.
The person who spent 2000 looking unfocused is now the person the room needs most.
You were not scattered. You were building context nobody else had the patience to build.
To the lost generalist
Stop apologizing for your breadth. Stop pretending you're working on "focus." Stop taking advice from people whose entire career has been one domain, about how to navigate a world that no longer rewards single domains the way it used to.
Your Wikipedia chains are not a waste of time. Your etymology obsessions are not a quirk. Your inability to walk past an interesting question without following it is not a problem to manage.
It is the thing. In an AI world, it might be the only thing.
Find the rooms where people are asking questions nobody else thought to ask. Add your chain. Make the connection. Let them wonder how you got there.
You've been getting there your whole life. You just didn't know the destination had a name yet.