The AI Won't Replace You. It'll Just Expose Who's Actually Doing the Work.
Let’s get this over with. Every week, another keynote speaker or VC-backed evangelist asks the same breathless question: "Will AI replace all the coders and architects?"
The short answer is no.
The longer, more useful answer is that it's going to ruthlessly replace the parts of the job that were never that valuable in the first place. It’s a tool that will separate the people who perform processes from the people who think.
We’re not here to sell you a dream or a nightmare. We’re here to make reality functional. So, let's talk about what's really changing.
Architecture Isn't About Diagrams Anymore
For years, the job of many IT architects has been to churn out diagrams. These documents are often just statements of belief—neat, tidy boxes that rarely match the messy reality of the system being built. The real logical architecture emerges from the people actually coding the thing.
This is where the job shifts. Architecture is becoming less about drawing fictitious diagrams and more about articulating and debating values. The crucial question we now face is: "Where do we want human decision-making in our systems?"
That’s a subtle, complex question that a Large Language Model (LLM) can't answer. An LLM can create "interesting" but deeply flawed concepts. It can give you a hypothesis, a thought-starter. But it's the human in the loop who has to squint at the output and ask, "Wait, why did it do that?" The model's job is to generate plausible-sounding nonsense; our job is to interpret, challenge, and fix.
Your Code Isn't Art. It's Furniture.
Some developers love to treat code like it’s artisanal craftwork, every brace hand-polished, every loop carved from a single piece of oak. But here’s the truth: most of the world doesn’t care. Most of the world runs perfectly fine on flat-pack.
Yes, you can whittle a bespoke chair out of mahogany. You can also waste six months of everyone’s time building something unique when what was needed was a seat that didn’t collapse. A lot of so-called “artisan” code is just that—pointless flourishes where a solid IKEA job would have done.
AI is becoming the IKEA factory for code. It churns out the flat-pack, the functional, the stuff nobody should be hand-carving anyway. And that’s not an insult—it’s an upgrade. We don’t need an army of artisans crafting endless vanity projects. We need systems that work, ethically and reliably, without bankrupting everyone in the process.
Your value isn’t in how pretty your chair legs look. It’s in knowing when the flat-pack is good enough, and when it isn’t. The AI will build the furniture. You’ll still be the one assembling it at 2am with the wrong Allen key, trying to work out why the instructions don’t match the parts in the box.
Taming the Legacy Beast: A Tool, Not an Answer
Nowhere is the gap between hype and reality wider than with legacy systems. The pitch from the evangelists is seductive: "Just point the AI at your legacy code, and it will create a new, modern environment!"
If you believe that, you’ve been listening to far too many VC pitches.
Here's what actually happens. You point an LLM at a complex legacy system, and it might reveal an architecture that's 60-80% correct. The problem is the 20-40% that's dangerously wrong, and you have no idea which bits are which. It’s "confidently wrong in a way that sounds plausible to someone who wasn't paying attention".
The smart approach isn't to ask the AI for the answer. It's to use the AI to help you build context-specific tools to find the answer yourself. This is the difference between asking a magic box for a fish and building a better fishing rod. One is a fantasy; the other is engineering. We’re not evangelists; we’re interpreters, and occasionally, firefighters.
So no, AI isn't going to take your job. But it is going to change it. It will automate the trivial and force a focus on the essential: judgment, values, and the radical act of asking better questions. The hype machine is designed to make you feel obsolete.
AI isn’t going to take your job. But if you cling to the parts of it that never mattered, you’ll hand it over willingly.
