The Great Decoupling: How AI Is Rewriting Work

Simon Carver

[calm] Welcome back to The Human Workforce Podcast. Today’s not some far-off thought experiment. It’s happening now, quietly in boardrooms and budget meetings. Over the last year or so, a line’s been crossed. Companies aren’t just trimming spend anymore. More and more, they’re replacing people so they can fund machines.

Lachlan Reed

Yeah. And let’s call it what it is. This isn’t just “efficiency,” and it’s definitely not a cute little restructuring. This is the Great Decoupling. Bit of a clunky name, I know, but it fits. Businesses are separating human labour from value creation, then wiring that value straight into compute. Payroll out, processing power in. Simple as that. Brutal as that.

Simon Carver

What makes this moment different is that layoffs and hiring are happening at the same time. That can sound contradictory until you look closer. Some roles are being cut, while other roles are being created around AI systems, data pipelines, infrastructure, and model operations. So the company may not be shrinking in the way people assume. It may be reallocating human value.

Lachlan Reed

Exactly. They’re not always shutting the shop. They’re changing who they reckon matters inside it. Analysts, operations staff, middle layers, people who move information from one screen to another, from one meeting to the next, those roles are exposed. Meanwhile, AI engineers, data specialists, infrastructure people, the folks building the rails, they’re suddenly gold. It’s like ripping the old engine out of the ute and dropping in a new one while the car’s still moving down the highway. Unsafe? Maybe. But they’re doing it.

Simon Carver

And underneath all that is a deeper shift in how businesses define value. The old model often rewarded effort, presence, repetition, reliability at the task level. The new model is increasingly asking: what can be automated, accelerated, or scaled through machines? If software can produce most of the output at a fraction of the time, leadership will be tempted, maybe compelled, to choose that route.

Lachlan Reed

Which means a lot of people are waking up to a nasty surprise. The job they thought was secure was actually a bridge job. It existed because moving work around used to need humans at every step. Now, not so much. If your role is mostly execution without interpretation, mate, you’re standing on thin ice. Even a kangaroo could trip over that one.

Simon Carver

I think that’s the hard truth. We’re watching the workforce split into two broad groups. First, the builders, the people who create, maintain, tune, and connect these systems. Second, the context experts, the people who guide the systems, validate the results, interpret what matters, and understand the consequences of getting it wrong.

Lachlan Reed

And the squishy middle? That’s the bit getting squeezed. Not because those people aren’t smart. Not because they don’t work hard. But because the market’s saying, “We don’t need as many humans just pushing the process along.” We need people building the machine, or people telling us when the machine’s gone off the rails. The middle layer, the human spreadsheet era, that’s copping it.

Simon Carver

There’s a human cost to that, and we shouldn’t talk about it like it’s abstract. These are careers, identities, routines, mortgages, families. So this episode isn’t about cheering displacement. It’s about seeing the structure clearly. Once you see it, you can respond to it.

Lachlan Reed

Yeah, because pretending it’s not happening is no good. You don’t beat a storm by calling it a light drizzle. Companies are trading payroll for processing power. They’re moving money from people to systems, and then hiring a smaller set of people who can make those systems useful. That’s the board-level logic. Cold, clean, and very real.

Simon Carver

So the question becomes: if this is the new map of work, where do you stand on it? Are you building intelligence, or are you supplying context to it? Because the space between those two is getting smaller by the month.

Simon Carver

So how do you survive this without slipping into panic or denial? I think we start with a correction. AI can replicate output. Sometimes impressively. It can draft, summarize, classify, respond, sort, predict. But output is not the whole job. Not even close.

Lachlan Reed

Yep. This is where a lot of companies got a bit too excited, a bit too early. They saw fast output and thought, “Beauty, problem solved.” Then reality wandered in wearing steel-capped boots. Because AI can spit out an answer, but it doesn’t actually carry context the way a human does. It doesn’t own the consequence. It doesn’t stand there when the regulator asks, “Who approved this?” It doesn’t get that sinking feeling in its guts when something looks right but is dead wrong.

Simon Carver

That’s the key distinction. AI gives speed. Humans provide judgment. And judgment is not some vague, magical quality. It’s the ability to weigh nuance, understand stakes, recognize exceptions, and decide under uncertainty. Accountability lives there too. A machine doesn’t absorb blame, explain intent, or repair trust after a failure.

Lachlan Reed

Right. It can mimic the surface of competence, but not the full stack of responsibility. That’s why some companies that rushed to replace people with AI ended up quietly bringing humans back into the loop. Not because the tools were useless. More because the tools weren’t enough on their own. Fast isn’t the same as safe. Cheap isn’t the same as complete. That’s a lesson people keep learning the hard way.

Simon Carver

Which leads to the three shifts professionals need to make. The first is moving from task execution to problem framing. In other words: stop defining your value as doing the task itself. Start defining your value as understanding what problem actually needs solving. AI is often capable at producing responses, but much less reliable at identifying the right question in a messy real-world situation.

Lachlan Reed

That one’s massive. Anyone can open a tool and have a crack. But knowing where the real bottleneck is, knowing what matters, knowing which question unlocks the whole thing, that’s where the money is now. It’s like working on an old trail bike. If you just keep polishing the mirror when the carb’s clogged, you’re busy, sure, but you’re not solving the problem.

Simon Carver

The second shift is from tool use to system thinking. Lots of people can use AI tools. Far fewer can design a workflow around them, connect inputs and outputs, spot failure points, and build a process that a team can trust. Knowing a tool matters less than understanding the system it sits inside.

Lachlan Reed

Exactly. Prompting alone isn’t a career moat. Harsh, but true. The real value is stitching things together. What comes in, what gets checked, what goes out, where a human signs off, where risk piles up. If you can think at that level, you stop being just a user. You become someone the business leans on.

Simon Carver

And the third shift is from output to accountability. Be the person who can say, “This looks right, but it isn’t.” Or, “This is good enough for a draft, but not for a decision.” That requires domain knowledge, yes, but also courage. Accountability means attaching your judgment to the outcome.

Lachlan Reed

And here’s the blunt version: if you’re not supervising intelligence, you’re competing with it. That doesn’t mean everyone has to become an engineer overnight. Thank goodness, because I’d probably solder my own thumb to the desk. It means you need to own the layer above the task. Design it. steer it. validate it.

Simon Carver

That’s why I think the most useful mindset now is agentic, maybe even managerial in a very personal sense. You are no longer just an employee completing assigned tasks. You are the product manager of your own role. You decide what should be automated, what must remain human, where quality breaks down, and where your judgment creates disproportionate value.

Lachlan Reed

Yeah. Offload the routine stuff. Keep the decisions. Lift your impact. Make AI your intern, not your replacement. Because the future probably belongs to people who understand systems, risk, and consequence, people with skin in the game.

Simon Carver

[firm] Don’t fight the machine. Learn to steer it.

Lachlan Reed

And don’t just work harder. Work where you matter. Good chat, Simon.

Simon Carver

Good chat, Lachlan. We’ll see you next time.