I've spent most of my working life moving between things. Every time I got deep enough in one, something else pulled me sideways. Each jump felt like starting over.

The advice was always the same: pick one. Go deep. Stop dabbling. So I kept feeling guilty that I didn't want to.

Here's what I've noticed over the past couple of years, as AI tools have gotten genuinely capable: the people who went deep in one thing are starting to look uneasy. And the ones who move across disciplines, the ones who never fit cleanly into a single title, they're suddenly very hard to replace.

That line has rattled around in my head for a while now. It gets at something most career advice won't say out loud. You don't need to be world-class at one thing. You need to be capable enough across enough things that you can move without asking for permission. See a problem, understand its shape, act on it. That's what dangerous means.

AI just made that quality worth a lot more.

What AI is actually replacing

There's a lot of noise about AI and jobs. Most of it is either too dramatic or too vague. The real answer is more specific than that: AI is excellent at going deep inside clear boundaries. Hand it a well-defined task and it will perform. Sometimes better than a person would.

Where it falls apart is at the edges. The places where one domain bleeds into another and the right answer depends on context that doesn't fit in a prompt. It doesn't know when the brief itself is the problem. It can't tell when a technically correct solution is going to create a political mess with the people involved. It doesn't carry the instinct you get from having worked across five different areas and recognising a pattern none of them would surface alone.

That's the generalist's territory. Not execution. Judgment.

The dangerous person in the room

In most teams, the person who actually moves things forward isn't the deepest specialist. It's the one who understands enough about everything to know what to ask next. Who can sit with the designer, read the technical constraints, feel the business pressure, and make a decision without waiting for three more meetings.

That person has always existed. AI just multiplied their leverage. That same person can now build a working prototype alone. Ship a campaign without an agency. Stand up an internal tool without filing a ticket. Not because they're the best coder or the best designer or the best strategist. Because they're dangerous enough at all three.

Specialists still matter, obviously. When the system breaks at depth, you need someone who actually understands what's happening underneath. But for the vast majority of day-to-day work, the generalist with sharp tools is doing what used to take a whole team. That's a real shift.

Where generalists still fall short

The generalist weakness has always been finishing. Knowing enough to start something, but not always enough to ship it at a level that holds up against someone who does that one thing for a living.

AI doesn't fully close that gap. But it raises the floor, meaningfully. A generalist who would have shipped something rough can now ship something genuinely solid. Not expert-grade, but good enough that the speed advantage makes the trade-off obvious.

The self-check still matters though. Knowing where your “good enough” actually holds, and where it doesn't. AI won't tell you that. It just means the line where you genuinely need a specialist keeps moving further out.

So what does this actually look like

If you've spent years feeling like your range was a liability, the ground is shifting under your feet. The market for people who can do a lot of things at a decent level and understand how they connect is genuinely expanding.

You don't need to master everything. Build a working vocabulary. Enough to ask the right questions, sense when something's off, and keep moving without waiting for someone else to unblock you. Then lean on the tools for the depth.

The point isn't to know it all. It's to be the person who walks into any room and is immediately useful. Who sees the shape of a problem before it's been fully explained. Who doesn't freeze.

You don't have to be the best. You just have to be dangerous.

Sacha