The Edge Moved: What "Being Human" Actually Means When Machines Do the Work

For about a decade, the advice was the same and it was right: produce more, move faster, out-work the room. It stopped being right sometime in the last two years, and most of us are still running the old playbook. The cleanest explanation I know comes from 1971, from an economist who never saw a language model and described this exactly.

The advice that quietly expired

Herbert Simon wrote a line in 1971 that has aged better than almost anything in his field: a wealth of information creates a poverty of attention. His point wasn't poetic. It was structural. When something becomes abundant, it starts to consume something else, and whatever it consumes becomes scarce. Scarce is where value collects.

For most of working life, output was the bottleneck. The person who could write more proposals, ship more code, make more calls, close more rooms, won. So we built whole identities around throughput. Wake earlier. Add a system. Squeeze another hour out of the day. None of it was wrong.

Then output got cheap. Not a little cheaper. A single person can now generate a hundred versions of a document, a campaign, a plan, a design, before lunch. The thing we spent ten years optimizing is now the part a machine does in seconds.

Which means the bottleneck moved. It always moves. And almost nobody updated their behavior, because the old advice still feels virtuous. Producing more still feels like work. It just stopped being the work that matters.

Where the value physically went

Run Simon's logic forward. If output is now abundant, what does all that output consume?

It consumes judgment. The decision about which of the hundred versions is worth your name. The taste to know that nine of the ten ideas are fine and one is actually good, and the ability to say why. The machine will hand you options forever. It has no opinion about which one is true.

This is why "being human is the edge" isn't a comfort blanket. I used to roll my eyes at the phrase too, because it usually shows up as reassurance, the thing people say so they feel less replaceable. It isn't that. It's a flat description of where the value moved when the cost of producing collapsed.

Think about what judgment actually is underneath. It's pattern recognition built from consequences you personally lived. The machine has read everything and felt nothing. It has never lost a client, never shipped the wrong thing and watched it fail, never sat across from someone whose voice changed half a sentence before they admitted the deal was off. Judgment is compressed experience plus stakes. That's the part that didn't get cheap. If anything it got rarer, because fewer people are building it. When the first draft is free, the muscle that decides what the draft should have been is the only one still doing real work.

Why working harder now makes it worse

Here is the trap, and it's a cruel one. The instinct, when you feel yourself falling behind, is to produce more. It's the move that worked for a decade. So people respond to an abundance problem with more abundance.

They generate fifty pieces of content instead of five. They let the machine draft every email and send most of them. They build more, ship more, automate more, and the output graph goes up and to the right. And the results get worse, or flat, and they can't see why, because the dashboard looks like effort.

What's actually happening is they've automated the cheap part and starved the expensive part. They've removed the friction that used to force a decision. When writing a proposal took three hours, you thought hard about whether to send it at all. When it takes thirty seconds, you send ten, and you've outsourced the judgment along with the typing.

The market notices. Audiences can feel content nobody decided to make. Clients can feel a plan that was generated rather than chosen. The flood of cheap output doesn't raise your value. It buries it, including yours, in the same sameness as everyone else who took the easy lever. More is now the commodity. The scarce move is restraint with a reason behind it.

What it actually looks like in practice

The disappointing part is how ordinary the edge looks once you find it. Nobody's doing anything cinematic.

It's the founder who reads the ten ideas the machine produced, deletes nine, and can tell you exactly why the tenth survived. Not a vibe. A reason.

It's noticing a client has gone quiet and tense before they've said a word about it, and acting on it.

It's the discipline to decide what not to build this quarter, when building everything is finally, technically, possible. The constraint used to be capacity. Now it's choice, and choice is harder.

So here's how to actually move. First, let the machine do the producing. All of it, shamelessly. Drafts, options, variations, the boring scaffolding. That's settled.

Second, spend the time you just freed on the decision, not on more production. Read the output like an editor, not a manager approving a queue. Ask one question of everything that crosses your desk: would I put my name on this and defend it in a room? If not, why exactly. Make yourself name the reason out loud. That naming is the skill.

Third, protect the experiences that build judgment in the first place. Stay close to customers. Keep making real calls with real consequences. Don't automate yourself out of the feedback that taught you taste, or in five years you'll have nothing left to judge with. We build this thinking into the agent layer at MARSA on purpose: hand the machine the work, keep the human on the decisions that have stakes.

The honest objection

The obvious pushback: won't the machines get good at judgment too? They're improving fast. Fair.

They'll keep getting better at the kind of judgment that's really pattern-matching at scale, the stuff that looks like discernment but is closer to averaging. Probably better than most people at that, soon.

But judgment about what matters is a different thing. It depends on values, on context only you hold, on knowing this specific client, this specific market, what you're actually trying to build and why. The machine can tell you what most people would choose. It cannot tell you what you should choose, because it doesn't have your stakes, your reputation, your particular bet on the world. Taste isn't an average. Often it's the deliberate refusal of the average.

There's a second, quieter objection worth naming: this can read as elitist, as if only the decision-makers survive. I don't think that's it. Judgment isn't a job title. The person closest to the actual work usually has the most of it, the warehouse lead who knows why the system fails on Fridays, the support rep who can hear churn three messages early. The shift isn't fewer humans. It's which part of every human's day is the valuable part. The doing shrinks. The deciding grows. That's available to almost anyone willing to build it.

The frontier was always you

Step back from the tactics. Something larger is happening here, and it's mostly good, even though it feels like loss while you're inside it.

For a long time, a lot of human effort went into being a slower, more expensive machine. Typing, formatting, copying, generating the obvious next version. We called it work and some of us were proud of how much of it we could do. The machines are taking that, and they should.

What's left is the part that was always more human and that we never had time for: deciding, discerning, caring about which thing is true, reading the person in front of you. The skills that don't scale and don't automate because they were never really about output to begin with.

The people who'll do well in the next decade aren't the ones who out-produce the machines. That race is already lost and it wasn't worth winning. They're the ones who let the machines produce and reinvest every freed hour into judgment, taste, and the relationships that build both. The edge didn't disappear. It moved up the stack, into the part of you that a model can't borrow. We built MARSA around that one bet. If it's right, the frontier was never the technology. It was always the human deciding what to point it at.