The new math of capability

Headcount is a vanity metric now. The number your function is proudest of, how many people it employs, is quietly becoming the best evidence that something is off.

For most of the modern corporation’s life, that number meant something real. Work was made of human hours, so more humans meant more work. Budget season ran on that arithmetic. A bigger team meant a bigger mandate. Span of control was a career ladder. “How many people report to you” stood in for “how much do you matter.”

That arithmetic just broke. Not because people got worse. Because systems got dramatically better, and most planning processes haven’t noticed yet.

The unit of capability is moving

Here is the claim, plainly. The unit of capability is shifting from people to systems. A small team with sharp systems will outperform a large team with smart slides, and the gap widens every quarter.

I’m not saying this from a conference stage. I’m saying it from inside the experiment.

I run an enterprise software company, Vidloft, where I am the entire engineering department. My co-founder handles delivery. One editor runs editing operations. That’s the roster. And the engineering output, measured in features shipped, systems maintained, integrations live in production, is what I would have staffed fifteen or twenty people to produce a few years ago. I also carry sales and strategy, because someone has to.

The trick is not that I type fast. The trick is that I mostly stopped writing code and started building the systems that write it. Agentic AI, chained into pipelines that plan, build, test, and review their own work, with me sitting exactly where the judgment goes.

I’ve built the other kind of company too. My first business was a video agency. Over ten thousand videos, Fortune 500 rebrand work, crews on planes. I know what a big-headcount operation feels like from the inside. The coordination tax. The meetings about meetings. The slow realization that a meaningful share of the payroll exists to manage the rest of the payroll. That model worked because it was the only model available. It isn’t anymore.

What the leverage actually feels like

Let me give you the honest version, because the internet is full of the dishonest one.

It is not magic. Nobody presses a button and receives a company. My mornings look like dispatch. Five or six agent runs in flight at once, each working from a spec I wrote carefully, because a vague spec produces confident garbage at machine speed. My afternoons look like review. Reading output the way an editor reads a rough cut, catching the subtle wrongness before it compounds into expensive wrongness.

The work didn’t disappear. It changed shape. The old constraint was hands. The new constraint is judgment. Taste. The discipline to decide what not to build, which turns out to be the most expensive decision in the whole building, because a system this productive will cheerfully produce ten times more of the wrong thing.

Some days the leverage is exhilarating. An idea at breakfast is a working feature by dinner. Other days it feels like conducting an orchestra that never gets tired and never asks whether the piece is worth playing. Every mistake in my thinking gets amplified with perfect fidelity. When I choose well, one person moves like a department. When I choose badly, I generate a department’s worth of rework and there’s no one downstream to catch it.

That’s the part the headcount worldview cannot see. When execution gets cheap, discernment gets scarce. The bottleneck in my company is no longer labor. It is the quality of the decisions I feed the machine.

Slides versus systems

I’ve now sat across the table from enough enterprise AI initiatives to notice the pattern. The forty-person function arrives with a beautiful deck. Transformation roadmap, center of excellence, three horizons, a slide with gears on it. Eighteen months of runway requested, mostly for hiring.

Meanwhile, somewhere else in the same company, four people have working pipelines quietly doing the job. No roadmap slide. Just throughput.

The deck describes capability. The pipeline is capability. And organizations keep funding the description, because the description comes with headcount attached, and headcount still reads as seriousness in the budget process. That’s the enemy here. Not large teams as such. The reflex that treats team size as proof of commitment, when it is increasingly just proof of lag.

The tell to watch instead is the ratio. Output per person, function by function. When that ratio is climbing, someone inside is building systems. When it’s flat while the headcount grows, you are paying for the org chart’s self-esteem. A four-person team outshipping a forty-person team used to be an anomaly worth investigating. Now it’s a preview. The anomaly is the forty.

None of this means firing your way to greatness. People still matter enormously. They just matter for different things than they did. The scarce human contributions now are judgment, taste, context, relationships, and accountability. Everything that is pure execution is migrating into systems, whether your planning cycle acknowledges it or not.

The half-and-double question

So here’s the practical reframe, and I’d encourage you to actually use it this planning season.

For every function requesting budget, ask one question. What would this function look like at half the headcount and twice the systems budget?

Not as a cost cut. Cost cutting is the boring, defensive version of this idea, and if that’s all you take from it you’ll do damage. Ask it as a capability question. If the answer is “roughly the same output,” you’ve learned the function is carrying structural padding. If the answer is “we’d collapse,” you’ve learned it runs on manual effort with no leverage underneath, which is its own kind of risk. And if a leader comes back with a real answer, a specific list of what they’d systematize, what they’d stop doing, and where the remaining humans would concentrate their judgment, you have found someone worth giving more resources to. Possibly more headcount, even. But headcount they can multiply rather than headcount they merely occupy.

The leaders who can answer that question are doing the new math of capability. Output over headcount, judgment over volume, systems as the asset and people as the intelligence that directs them. The leaders who can’t answer it will keep bringing you bigger teams and smarter slides, and for a while that will keep looking like seriousness.

It isn’t. I watch one person’s judgment, routed through good systems, outbuild rooms I used to need. The question was never how many people you can command. It’s how much capability you can direct. Those used to be the same number. They’re not anymore, and the organizations that internalize the difference first are going to make everyone else’s org chart look like a museum exhibit.