LinkedIn

Magnus PS

The Space Between

I don't write code anymore.

Maybe an exaggeration... the writer's oldest privilege. The honest version is that I barely write code. And I used to get paid for it.

These days I describe what I want in plain English, and a machine builds it. First drafts, reports, double-checking other people's work, meeting notes, the to-do lists that keep a team on schedule: all of it now has AI riding shotgun. My job used to be doing the work. Now it's mostly knowing exactly what I want and being stubborn until I get it, which, my wife would tell you, plays to my strengths.

And I like it. Coding was always a means to an end for me, the way a shovel is a means to a hole. I don't miss the shovel.

You don't need to know a line of code to know the feeling. If you've let AI write an email, summarize a meeting, or plan a trip, you've handed over a shovel too.

But I've learned a few things about the machine now holding it.

Junior mind, heavyweight swagger

AI is wrong a lot. Not shyly wrong. Wrong with its chest out.

I've watched it bloat my code by more than half with features nobody ordered. I've watched it merge my meeting notes with a colleague's until neither of us could tell whose opinions were whose. Coworkers tell me it has ignored their direct requests so it could keep doing its own favorite thing. It has the experience of a junior hire and the confidence of a prizefighter at the weigh-in.

I only caught it because I knew better. I knew Python and every feature on the list, so I could spot the stowaways. I'd sat in the meeting and taken my own notes, so I could spot the fiction.

That knowledge didn't come from a tool. It came from years of doing the work the slow way.

Which raises an uncomfortable question: what happens to the people who never do?

1. Cheap output, fragile minds

When anyone can produce a report, an email, or a design in seconds, output stops being impressive. It's hard to brag about a thing that costs nothing to make.

Value moves to judgment: knowing which output is right, what's missing, and what to do next. The catch is that judgment is built by thinking, and every task we hand off without engaging is a rep we skip. A mind that only works with the tools on is fragile, and it weakens at the exact moment its judgment becomes most valuable.

I think this explains a strange mood going around. People are producing more than ever and trusting themselves less than ever. They're busy and uneasy and can't quite say why.

Autopilot looks like progress. It's fool's gold. Skate to where the puck is going, and the puck is headed toward people who can still think when the Wi-Fi drops.

When I picture the most valuable person at my company five years from now, they aren't the fastest prompter. They can sit with a customer for two conversations and walk away knowing exactly what that customer needs. They build a rough prototype on the spot to prove it. They say no to features, to projects, even to customers, because more is not a strategy. And they aren't too proud to put marker to whiteboard, which is exactly why old-school clients trust them to drag everyone into the future.

So keep your faculties in shape. Write. Read books. Do the math in your head. Solve a problem by hand. Not all day, just often enough that it doesn't wither. Old school, it turns out, is the new school.

And use judgment about when to reach for AI at all. Not every passing thought deserves a prompt. Some deserve a walk.

2. Your network is your leverage

AI can draft the message. It can't be the person someone calls when things go sideways, the name someone vouches for in a hiring meeting, or the face someone remembers when an opportunity opens up.

Trust is built slowly, in person, and can't be generated on demand.

The slowness is the whole point.

Relationships compound, and they turn your skills into options. If your role changes or vanishes, a strong network plus sharp judgment means you have somewhere to go. That security is what lets you rest instead of grinding out of fear.

A fair counterargument is that AI will absorb more "human" work than we expect: screening interviews, support calls, maybe even rapport. Some companies will certainly try. I don't think it'll go well, and people are already tiring of talking to machines. But even if I'm wrong, the conclusion holds. The less you can predict which tasks will survive, the more your safety rests on people who know you, not on any single skill.

The work isn't disappearing, it's changing shape. The people who fare best will either be "hard human" or live on the bridge between human and artificial, whether through hands-on trades like electrician, carpenter, or physical therapist, or by leaning hard into the human parts of the job they already have.

The skill is what it's always been: build relationships, tend them, and lean on them in a time of need, preferably before the need arrives. People have ridden that skill all the way to the White House (ahem Billie Clinton).

You can ride it wherever you're going, once you've decided where that is.

3. The daily minimum

Here's the "How to", with a floor for busy weeks and a ceiling for good ones.

The floor: protect the first and last hour of your day. No screens, no blue light, no notifications. Warm up with journaling, reading, meditation, or better yet, a workout.

One step up: carve out fifteen screenless minutes in the morning, at lunch, and in the afternoon. Close the laptop. Walk. Look at something far away. Call someone. Sketch the problem on a whiteboard. If you're remote, find some in-person time at the office, a coworking space, or a team day. Humans work better around humans; it's one of our few reliable features.

The ceiling, the habits that keep me sane in the age of AI:

Each of these protects one of three things: your judgment, your relationships, or a mind that runs without a power cord.

The space between ...

Machines will keep getting better at making things. Let them. What they can't make is you: the taste you've earned, the people who trust you, the thoughts that show up on a long walk with nothing in your pocket.

If all this seems like too much, start small. Bookend your day with an hour on each side that belongs to no screen, and see who shows up.

...

Want to try this yourself?

Here are the questions I used to write this piece.

#1 Think of a time in the past year when you or a coworker caught an AI tool being confidently wrong, or made a call it couldn't have. What did you know that it didn't, and where did that knowledge come from?

#2 What's one skill you use less since you started relying on AI? What would it cost you if it quietly disappeared?

#3 Picture the person at your job who'll be most valuable in five years. What do they do differently, and how much of it happens in front of a screen?