Let’s talk about the person who said you could do it before you actually could.
Most people can name one. Not a mentor in the LinkedIn sense. A person who put your name forward for something you were not ready for — and then had to sit there while everyone watched how it went.
The teacher who entered you for the thing you had no business entering. The manager who handed you a project meant for someone ten years older. The uncle who let you drive the bakkie down the farm road at eleven, hand hovering near the wheel, saying nothing, because he could see you were doing fine.
Look at what those people have in common. They didn’t teach you much. Some of them couldn’t have. What they did was spend something on your behalf — their name, their judgment, their face in a room where you weren’t standing. And then they waited to find out whether they’d been stupid about you.
That last part is the whole thing. Keep it in mind.
Now the part I actually have to say honestly, because I’m not here to pretend the machine is useless.
It can teach you anything. It’s patient. It doesn’t sigh at the fifth time you ask the same question. It can go as slow as you need, in your language, at two in the morning, and it never once makes you feel like the stupidest person in the room. That is a real gift and a lot of people never had it.
But it cannot believe in you.
And not because it’s cold. That’s the lazy version of the argument and it’s wrong. It can’t believe in you because belief requires a person who can be wrong about you — and pay for it.
Encouragement is free. We’ve confused the two. Encouragement costs a sentence. Belief costs a reputation.
When someone with standing puts their standing behind you, they’ve made a bet with their own money. If you fall flat, it’s their name that takes the damage, not yours. That’s why the sentence lands the way it does. It’s not a compliment. It’s a transfer of risk.
And here’s what that risk actually buys: permission.
It doesn’t add a single skill to you. It changes what you’re willing to try out loud. Almost nobody attempts the big thing because they privately decided they were ready. They attempt it because someone with authority made it socially survivable to fail at it. The skill was already there. The permission was the missing part.
Think about the last risk you took. There was a person somewhere upstream of it.
Now look at the arithmetic, because this is where it gets uncomfortable.
Capability is being produced at an absurd rate and dropping in price every quarter. Meanwhile the number of humans willing to stake something on a specific person is quietly going down. Nobody announced that. It just follows.
The junior developer whose code nobody has ever personally vouched for. The first-year hire whose work gets reviewed by a model, so no senior ever has to make a judgment call about them out loud. The manager who hasn’t backed a bet on a person in two years, because the tooling means they’ve never had to.
Each of those is a small mercy at the time. Multiply it across a company and you get an organisation where nobody has ever lost anything by trusting an unproven person — because nobody has ever trusted one. And where nobody has ever been right about a rising talent either, because being right and being wrong both need you to have put something on the table first.
That’s a bench that quietly disappears. In five years you’ve got a layer of people who were extremely well served and never once believed in. Nobody learned how to back a person. It’s not a skill you can pick up from a screen.
Because here’s the part the machine cannot fix, no matter how good it gets.
Your career is not your skills. Skills are cheap now, and getting cheaper. Your career is who answers when your name comes up and nobody has the file in front of them.
That asset does not come out of a tool. It gets built before you need it, in small asks and small deliveries, with people who watched you do a thing you promised you’d do. There’s no export button for it. You can’t migrate it to a different vendor.
So two moves, and I’ll keep them blunt.
First: go ask someone specific for something small. Not a favour. A real ask, where they can say no and you’ll still be fine. Being believed in starts with being visible, and visibility is a choice you have to make out loud.
Second: if you’re holding any standing in your field at all — even a little — spend some of it on one person. By name. In a room they’re not in. Say the actual sentence: this one’s good, give them the thing, I’ll answer for it. Then answer for it.
That’s the principle worth defending. The right to be believed in — and the obligation, once you’ve got anything worth risking, to risk it on someone else.
Belief is the last thing with a price attached. The machine can generate every other form of support at zero marginal cost, and it will. This one it can’t, because belief requires a person who can be wrong about you and carries the consequences when they are.
I think that’s a good thing. It means the most valuable thing you can give somebody is the one thing a model will never hand out. And unlike everything else we’ve been optimising, it doesn’t dilute when you give it away.
Be the person who said they could. Somebody’s waiting for that sentence.
You’ve got this.