The most important skills in today’s workplace are judgment and discretion.
This is because the cost of output has never been cheaper. With models like Fable producing yet another step-function increase in capabilities, it’s easier than ever to run a million miles per hour in any direction, so that means you better be running in the right one.
Productivity at work used to be measured largely by output. Yes, results and impact have always been the flagship markers of success, but you could get pretty far in yesterday’s world by producing a high volume of just-fine output. That’s because output used to be expensive—you had to think and devote energy towards what you were creating. The effort itself signaled that you cared enough to think through what you were putting into the world, and visibly caring about your job was typically good enough to keep you from getting fired.
So because of that, most of us, to one degree or another, are hard-wired to believe that output equals good. Or at least productive, and productive equals good.
But that mental model starts to break in a world where you can chat with Claude, paste in a co-worker’s proposal for a new strategy, type “Nah, let’s try a different direction,” and get 1,000 words of modestly well-reasoned prose. The language is coherent. It’s probably more formally correct than most people can write. And it took nine seconds.
That shift creates a problem for companies like Amazon, JPMorgan, Meta, and Disney that are trying to drive AI adoption by deploying AI leaderboards that reward pure quantity of output. One Disney employee reportedly hit Claude over 400,000 times in nine days. When output is no longer scarce, rewarding more of it may not optimize the thing that creates the most value.
Now, at this point, most people will say, “Well Nick, you just have to use AI correctly.” I agree! Better prompting produces better output, and the people who are good at this are getting meaningfully more done.
But that’s not the problem I’m worried about.
Because when output is cheap, and you put it out into the world without substantial discretion, you’re not being productive. You’re shifting effort that should be yours onto your colleagues. You did the easy part, and you left them the hard part, figuring out whether any of it is true, relevant, or worth doing anything about.
That’s what judgment is. Deciding which ideas deserve attention—yours, your colleagues’, your organization’s.
You’ve seen what happens when nobody exercises it. The 10-page AI-generated strategy memo that could have been two. The Slack update that buries no substance under ten beautifully formatted bullets. In volume, they suffocate a company, and they drain exactly the velocity AI promised to create.
Effort is cheap
Before AI, effort itself was the signal. There was an implicit social contract in reading and writing: if you were reading someone’s work, you knew it was worth reading, because the act of devoting time and energy to it signaled that the writer cared enough to work on it. If someone spent a few hours on something, it was probably worth a quick skim on your part.
Now that output is cheap, quantity isn’t a signal for “I care.” And the problem is, we don’t know what the new signal is. I only want to read things my colleagues have deemed important enough to spend real mental energy on. That used to be easy to spot, but today it’s much harder to tease out.
The playbook
Until we figure it out, we need some ground rules for what we put into the world:
- Think first; use AI second.
- Every sentence should earn its place, and the meaning of every sentence should be clear.
- Bias toward producing content that only you could have written—proprietary data, your actual opinion, your voice. If I wanted to ask Claude, I would have!
- You’re probably sending slop if it takes more time for someone to read the output than it took you to think about your perspective and do the input.
- Assume the quantity of AI output can be cut by 50% to 75% and cut it.
I’m fortunate to have received a decent education in writing, and most of the rules above are fundamentals that were drilled into me in high school. Reflecting on those lessons, I now appreciate William Faulkner’s line as more true than ever: “In writing, you must kill all your darlings.” The exercise has always been agonizing.
But it’s more relevant now than it’s ever been. What changed is that this homicide has become mission-critical to a high-functioning workplace. Deciding which ideas deserve to exist is one of the most important competencies of modern work.
We should spend at least a fraction of the time AI saves us deciding what’s actually worth putting into the world. To do this, especially in the face of increasing pressure to move faster than ever, requires not just the skill to do so, but a work environment and leaders that champion slowing down to speed up. Spend 15 minutes with a blank canvas before touching Claude. Ruthlessly edit output. Rewrite it in your voice.
AI can generate almost anything—that doesn’t mean almost everything deserves to exist. Today’s winners won’t be defined by their ability to produce quickly. Instead, the breakout companies of today’s generation will be defined by their judgment about what should be produced in the first place. And if we do that, I’m confident we’d move faster than ever, and in exactly the right direction.
