The Vanishing Apprenticeship: What Happens When AI Learns the Job Before People Do

Simon Carver

Welcome back to The Human Workforce Podcast.

Lachlan Reed

Good to be here, mate. And today we’re getting into something that feels small at first... but I reckon it’s massive. Not flashy AI demos, not robot apocalypse gear. Just the quiet vanishing of the beginner’s lane at work.

Simon Carver

Yeah. The missing first rung on the ladder. The part of work where you used to be allowed to be new.

Lachlan Reed

Picture a pretty normal office now. Or honestly, a Slack channel and a bunch of tabs. The AI writes the first draft. It summarizes the meeting. It cleans up the email. It suggests the code. It turns messy notes into something neat as a pin.

Simon Carver

Which sounds wonderful... until you imagine the youngest person in the room. The new analyst, the junior marketer, the first-year associate, the fresh developer. They’re handed something polished. Tidy. Fast. Maybe even better than what they would’ve done.

Lachlan Reed

But they didn’t do the ugly bit. They didn’t wrestle with the blank page. Didn’t write the clunky opening line. Didn’t sit there thinking, “Hang on, what am I even trying to say here?” And that ugly bit, weirdly, is where your brain starts growing muscles.

Simon Carver

I keep thinking about old apprenticeships. Not in some romantic, sepia-toned way. I mean the practical thing. You watched. Then you tried. Then you got it wrong. Then someone better than you said, “Not quite. Here’s why.” That rhythm mattered.

Lachlan Reed

Exactly. Whether you were learning to weld, write a client memo, edit audio, whatever. You didn’t start by approving perfect work. You started by making slightly dodgy work. [laughs softly] Bit wobbly. Bit embarrassing. Like my first attempt rebuilding a carburetor in the shed. Looked like a spider had assembled it.

Simon Carver

And now a lot of workplaces are saying, maybe without meaning to, “We’ll skip that part.” Because the machine can do the beginner work faster. So the junior person becomes... what? A reviewer? A prompt refiner? A passenger with edit access?

Lachlan Reed

Yeah, and that’s the trap. People say, “Well, they’re still involved.” Maybe. But involvement isn’t apprenticeship. Watching a magician pull a rabbit out of a hat doesn’t teach you sleight of hand.

Simon Carver

Or to put it more plainly: if your first years of work are spent curating machine output, there’s a real question underneath that. When do you become someone who can generate judgment of your own?

Lachlan Reed

That’s the bit that gives me the wobbles. Because this isn’t just tasks disappearing. It’s the training ground disappearing. The messy practice field. The shallow end of the pool where you learn not to drown.

Simon Carver

And organizations may not notice at first. Everything still looks productive. The documents are cleaner. Turnaround is quicker. People seem efficient.

Lachlan Reed

But five years later, you look around and go, “Hang on... why is nobody ready?” Why is there a gap between the senior few and everybody else? Why can people ship work, but struggle to explain it, defend it, or remake it from scratch when the system gets weird?

Simon Carver

Because the ladder wasn’t climbed. It was deleted.

Lachlan Reed

You know what’s funny? Most of us were pretty ordinary at the start. Not tragic, maybe, but definitely rough as guts. First presentation, first client call, first strategy memo... you could hear the gears grinding.

Simon Carver

[laughs] Oh, absolutely. My early stuff had that very specific energy of someone trying to sound smart instead of trying to be clear.

Lachlan Reed

Mate, same. I used to think if I just chucked in enough polished words, no one would notice I was still figuring it out. But they do notice. And that’s not a disaster. That’s the lesson.

Simon Carver

That awkward phase teaches things that are hard to measure. Judgment, for one. Not just, “Can I produce an answer?” but “Is this the right answer for this person, in this moment, with these stakes?”

Lachlan Reed

And confidence. Real confidence, not fake swagger. Real confidence comes after you’ve been wrong a few times and survived it. You learn, “OK, I can recover. I can ask a better question. I can spot the weak bit before it goes out the door next time.”

Simon Carver

And taste. That’s the slippery one. Taste is built by making mediocre things and slowly learning why they’re mediocre. Why one sentence lands and another clunks. Why one recommendation is brave and another is just noisy.

Lachlan Reed

AI kind of scrambles that sequence. It shows up looking instantly competent. Maybe not wise, maybe not deep, but smooth. Smooth enough that a beginner can mistake fluency for understanding.

Simon Carver

Yes. The danger isn’t only bad output. It’s borrowed competence. The feeling of “I know how to do this,” when what you really know is how to select among polished options.

Lachlan Reed

That’s a bloody good way to put it. Because low-stakes work used to do something important. It let you fail where the cost was manageable. Draft the internal note. Build the rough spreadsheet. Sit in on the meeting and take a crack at the summary. Nobody expected genius. They expected progress.

Simon Carver

And from that came pattern recognition. You start noticing what matters. The client says one thing but means another. The data point that looks dramatic is actually noise. The teammate who goes quiet is the one you should check on after the call.

Lachlan Reed

AI can hand you an answer. It can’t hand you the memory of almost getting it wrong. And weirdly, that memory is gold. It’s like touching a hot exhaust pipe once. You don’t need a second workshop on the concept.

Simon Carver

There’s also a dignity in being taught through effort. In earning your way into a craft. Not because suffering is noble. I don’t buy that. We shouldn’t keep pointless drudgery just because previous generations had it hard.

Lachlan Reed

Yeah, no medals for unnecessary pain. If AI can kill some nonsense work, beauty. Let it. But if it removes every beginner rep, every small moment where someone learns by doing, then we’re not freeing people. We’re flattening them.

Simon Carver

And then we act surprised when fewer people develop range. We wanted speed. What we may lose is depth.

Lachlan Reed

That’s the heartbreak, eh? Not that work gets easier. It’s that people may never get the chance to become excellent in the old human way: slowly, awkwardly, with someone nearby saying, “Try again.”

Simon Carver

So the obvious question is: what do we protect? If AI is going to stay in the workflow, and it is, what has to remain human on purpose?

Lachlan Reed

First thing for me is shadowing. Real shadowing. Not “Here’s the final deck, have a peek.” I mean sit in the room, watch how a hard conversation gets handled, listen to the hesitation, the phrasing, the little judgment calls. That’s where the trade gets passed down.

Simon Carver

And coaching. Specific coaching. Not vague praise, not annual feedback written like it was generated by a polite toaster. I mean somebody saying, “This part worked. This part didn’t. Here’s what I think you were missing.”

Lachlan Reed

[laughs] A polite toaster is bang on. But yeah, leaders have gotta preserve room for imperfect work. That’s the one that feels most endangered. If every deliverable has to be machine-polished before a junior touches it, then they never build the internal wiring.

Simon Carver

Which means designing jobs differently. Not just around consumption of AI output, but around growth. Let the new person draft first sometimes, then compare. Let them explain why they disagree with the machine. Let them make a recommendation before they see the synthetic one.

Lachlan Reed

Exactly. Use AI like a spotter at the gym, not the bloke doing the entire lift for you. Terrible analogy—nah, actually, it works. If the weight never hits your arms, your muscles don’t care that the rep got completed.

Simon Carver

And there’s a leadership courage here. Because teaching slows things down in the short term. It does. There’s no magical way around that. Apprenticeship is inefficient if all you worship is this quarter’s output.

Lachlan Reed

But if you want a bench—if you want people who can one day carry the hard stuff without panicking—you have to let them practice while the stakes are still small. Otherwise you end up with a company full of operators and not enough craftspeople.

Simon Carver

I think that’s the deeper choice in front of organizations. Do you want workers who can merely supervise tools? Or do you want people who are becoming wiser, steadier, more capable over time?

Lachlan Reed

And that becoming bit matters. Work isn’t just a machine for extracting output. It’s one of the main places adults are formed. Where they learn judgment, restraint, courage, standards. Strip out the learning stage and you don’t just change productivity. You change people.

Simon Carver

So yes, keep the intelligence. Use the systems. Let them reduce waste. Let them help. But don’t confuse assistance with formation.

Lachlan Reed

That’s it. A healthy workplace needs brains, sure. Fast ones, even. But it also needs initiation. A way for someone new to cross the bridge from observer to contributor to trusted adult. If that bridge disappears, all the clever tooling in the world won’t save you later.

Simon Carver

Because eventually every organization is handed a messy moment with no template. And on that day, what matters is whether you have people who were actually taught.

Lachlan Reed

[softly] Not just assisted. Taught. Simon, that feels like the whole thing, honestly.

Simon Carver

Yeah. I think we’ll leave it there.

Lachlan Reed

Good yarn, mate. Catch you next time.

Simon Carver

See you soon, Lachlan. Bye, everyone.