I spent about 13 years turning ideas into code. I haven’t written a line in about two years. I read less code every month, and I have never been more effective.

Senior engineers who are still skeptical of AI coding tools aren’t being rigorous. They’re being nostalgic. And in a technical field, nostalgia is a career risk.

We have seen this movie before. Java developers who refused to reckon with modern languages. Infrastructure teams that dismissed the cloud as a fad. The pattern is always the same: a new approach arrives that is faster, cheaper, and good enough. The holdouts argue about quality and control. The adopters ship products. History doesn’t wait for consensus.

The Craft Argument Doesn’t Hold

When factories started mass-producing shoes, cobblers didn’t disappear. They had to become specialists. Today, a master cobbler still exists, but the customer paying for bespoke footwear is buying something beyond function. They’re buying a story, a feeling, a thing they can point to and say “that was made for me.”

The same happened with furniture. Ikea didn’t kill woodworking. It commoditized tables. The master furniture maker survived by serving customers who attach meaning to the object itself; its provenance, its maker, the craft behind it. People form memories around that kitchen table. They tell guests about the woodworker who built it.

Can you say the same thing about software?

Software serves a purpose. The button needs to do the thing. When a user clicks “submit,” they don’t care whether the handler was written by a 10x engineer with elegant abstractions or generated by Opus 4.6 in 90 seconds. They care that it works. There is no pitch that goes: “Our code was hand-crafted by senior engineers who loved every line. That’s why we charge $499 a month instead of $29.” Maybe a handful of enterprise buyers find comfort in that story, but it is not a business model.

Software is not a table. It has no grain, no weight, no smell. It carries no story the user can see or touch. What users feel is the product. The experience. The outcome. Not the process that produced it.

The Farming Numbers

The cobbler still exists. The furniture maker still exists. There is real demand for both, but it is a fraction of what it once was, and it serves a niche willing to pay a steep premium for the story behind the object.

Farming is the starker example. In the 1700s, nearly 80% of Americans were farmers. By 1900, that had dropped to around 40%. As of 2015, fewer than 2% of the workforce works in agriculture. And those 2% feed the whole country, because they embraced the machines. Total agricultural output has more than doubled since 1948, while the labor required has fallen to a quarter of what it was (source). The farmers who survived didn’t do so by defending the old way. They did so by letting go of it.

The ones who didn’t embrace the machines had to become artisan farmers selling heritage grain to fine restaurants. Most of them just stopped farming.

There is a small market for the “farm-to-table, hand-harvested by people who love the land” story. Someone is buying it. But you are not building a software company on “our engineers loved every line of code.” That story doesn’t transfer. Software has no grain, no weight, no smell. Your users click a button. They don’t care who wrote the handler.

The Job Has Changed

The engineers who will matter in five years are not the ones who write the best code. They are the ones who design the best systems, ask the best questions, and know how to give an agent a goal and a set of guardrails and get out of the way.

That is not a lesser skill. It is a harder one. You still need to understand architecture, tradeoffs, failure modes, and user needs. You need to know when the agent is going in the wrong direction and why. But you no longer need to be the one typing the implementation. That part has been automated, and pretending otherwise is covering your ears.

I earn a respectable living because for ~13 years I understood how to turn ideas into working software. That skill got me here. But the job I do today looks almost nothing like it did then. I spend my time designing systems, setting goals, defining constraints, and watching agents build things better and faster than I ever could on my own. The implementation is not the job anymore. The judgment is.

The cobbler still exists. The furniture maker still exists. The artisan farmer still exists. But there are far fewer of them, and the market for their work is small. Most people who once held those jobs didn’t gracefully transition into niche premium specialists. They were displaced, and the industry moved on without them.

Senior engineers are not immune to this. The question isn’t whether AI changes the job. It already has. The question is whether you are going to adapt before the market makes the decision for you.