What does it look like when AI takes your Job

For developers, there's the tendency to imagine a showdown. A senior developer on one end, and a Sophisticated AI on the other, racing to complete a Jira ticket. Whoever completes the work first is the clear winner. It has to be high quality code, both reusable and scalable. This is where real human programmers like to believe they will make a difference. This is pure fiction. AI doesn’t compete with you. It dissolves your job into the system.

The reality is entirely different. We compare our modern days to the era of the horse and car. Of the assembly line worker and machine that replaced him. It was the physical work that was being replaced. When AI takes your job, it does not hijack your desk in the office. Instead, your job disappears. It becomes a flag in the system.

Imagine the job of a proofreader. It’s not just that computers are better at it, it’s that proofreading no longer exists as a distinct task. In modern text editors, spellcheck happens in real time, seamlessly integrated into the act of writing. When you misspell recieve, it’s corrected without interrupting your flow. The writer becomes the proofreader, not because they’ve gained new skills, but because the system has dissolved the boundary between the two roles.

This isn’t about AI ‘stealing’ jobs, it’s about dissolving the boundaries between roles. Just as proofreading evaporates into the act of writing, management evaporates into data.

Which job is at stake?

We often think of customer service as a prime target for AI disruption. In fact, I worked in a startup where we successfully automated customer service for several companies. From the outside, people thought the AI agent will replace the human customer service agents. But the job at stake was not that of the agent. It was the manager's.

What we found is the most sophisticated of AI models can only handle about %40 of issues. Any gain grows exponentially in complexity. The human agents are still needed after using AI. However, the managers become superfluous. With AI systems tracking metrics like refund rates and exception approvals in real time, the ‘gut feeling’ decisions managers once made are now automated. Agents no longer need oversight; they need autonomy. The role of the manager isn’t replaced—it’s bypassed.

In my experience, customer service managers get promoted out of the job after a successful integration.

The issue is not unique to AI, it's technology in general. Automated metrics don’t just replace managers, they flatten hierarchies. A CVS store run by two employees isn’t a triumph of efficiency; it’s evidence that data has turned ‘management’ into a set of rules embedded in software.

The purpose of the manager was to resolve uncertainty. But when AI quantifies every exception such as tracking returns, refunds, and risk, then uncertainty becomes a spreadsheet cell. Decisions once entrusted to human judgment are now dictated by systems. The manager’s job doesn’t vanish because AI is better at it. It vanishes because the system no longer needs their permission to act.

Software Engineering Dissolves Into Systems

This is the paradox of progress: the more a job relies on managing complexity, the faster it dissolves into the tools it creates. Just as spellcheck dissolved proofreading into writing, and metrics dissolved management into software, the act of coding dissolves into the systems engineers once painstakingly built. The developer’s role isn’t to compete with AI. It’s to design the systems that make their own expertise obsolete.

In 2011, my value as a developer hinged on a seemingly superhuman skill: ensuring JavaScript code ran flawlessly across every browser and device. Internet Explorer 6’s quirks demanded ancient incantations in CSS. Mobile Safari’s touch events required polyfills. My days were spent wrestling with inconsistencies, patching gaps in standards, and writing tribal knowledge into brittle workarounds.

By 2025, this expertise sounds like folklore. Modern browsers are near-homogeneous, and build tools like Babel and Webpack automatically transpile, optimize, and polyfill code. What once required a human specialist, debugging IE6’s box model or Android 2.3’s JavaScript engine, is now a checkbox in a configuration file. My old job didn’t disappear; it dissolved into the infrastructure.

The role of the software engineer hasn’t been replaced. It’s been redefined. We’ve shifted from writing code to curating systems that write code. The job title remains, but the work is unrecognizable to past versions of ourselves. With the current wave of LLMs, you won't notice a small shift at your job. Before you know it, you'll find yourself in a new position. And your old job will only exist as a setting in the system.

---