A software engineer's post climbed to the top of Hacker News, drawing 687 comments. The title was blunt: "LLMs are eating my career, and I don't know what to do." What the author described wasn't a layoff notice. He still has a job. The paychecks still arrive. But something has shifted—a sensation he couldn't shake—and 729 people who read it hit the upvote button.
It's worth looking closely at where that resonance came from. If this had been the vague dread of "AI is taking our jobs," the comment section would have looked different. There was consolation, and there was pushback, but far outnumbering both were firsthand accounts saying, in effect, "I know exactly what this feels like." The post didn't touch some distant future scenario. It named a concrete change people feel right now, in front of the tools they use every day—and that's what generated 687 responses.
What Changes When You Become the Person Who Reviews the Code
Using an LLM makes writing code faster. That fact itself isn't in dispute. The LLM drafts a substantial share of the engineering work, and the human shifts into the role of reviewing and revising it. On the surface, this is efficiency: more work completed in the same amount of time.
But the discomfort the author described started somewhere else. Writing code from scratch used to mean confirming with your own hands why something worked and where it might break. Deciding things line by line was tedious and slow—but that tedium was the foundation that let you handle more complex systems later. With the LLM producing the draft, that path got shorter, and the practice of designing and constructing logic yourself shrank along with it.
What happens after six months, a year? If a situation arrives where you have to write code from scratch without an LLM, can you do it at your old speed? Facing that question, the author wrote that he wasn't sure. He never used the phrase "my muscles have stiffened," but the sensation he described came close.
This phenomenon isn't confined to software engineering. Planners who've started handing planning documents to AI, and professionals who've started drafting reports with LLMs, end up standing before the same question: Do I still hold the logic of this work in my own head? If someone asked me why this sentence is here, why this structure took this shape, could I point to the answer?
Layered on top of this is a signal from the market. If one AI-assisted engineer can handle what three used to, companies have an incentive to resize their workforce. Starting in the second half of 2024, a string of tech companies—Google, Meta, Salesforce, and others—scaled back hiring or trimmed team headcounts, and a significant share of those moves were restructurings tied to AI adoption. Which is to say, the author's anxiety may not be merely a matter of personal sensitivity.
The Counterargument: LLMs Actually Raise the Value of People
There is an argument running in the opposite direction, and it can't be brushed aside.
As several engineers pointed out in the comments, LLM-generated code can't go into production without review by someone competent. Making architecture decisions, responding to shifting requirements, catching security vulnerabilities and edge cases—these remain human work. On this view, LLMs relieve people of repetitive tasks and free them to focus on harder problems. Not a crisis, but a reshuffling of roles.
History offers similar patterns. When Excel arrived, there were predictions that accountants would disappear; in reality, demand grew for people who could handle more complex financial analysis. When Photoshop arrived, there were worries that graphic designers would dwindle; instead, the design industry expanded. There is a recurring pattern in which technology automates certain roles while simultaneously creating demand for higher-level ones.
But for this counterargument to hold, one condition has to be met. It's true for people who already possess the capability to "focus on harder problems." For those still building that capability—or those who, in handing repetitive work to AI, handed over the practice of thinking for themselves along with it—the outcome looks different. The confession that earned 729 points resonated so uncomfortably precisely because it captured the latter sensation.
What Changes for Korea's Solo Business Owners and Middle Managers
Translate this debate into the perspective of working professionals in Korea, and the question narrows to one: as the share of my work I hand to AI grows, is the value I add visible from the outside?
Middle managers driving AI adoption keep running into the same situation: throughput goes up, but team members' roles blur. If the workload has shrunk while headcount stays flat, pressure builds to explain to those above what each person is actually doing. "The AI did it" doesn't relieve that pressure. What has to be visible is the direction you steered the AI, where you intervened, and what you added to the output.
For those running teams, a different worry emerges. If team members are working faster thanks to AI, where should the freed-up time go to put them on a growth path? Does processing a larger volume of the same work build their capability—or is it a more sustainable investment to spend that time deeply understanding the logic and context behind the work they've been doing? A manager's answer to this question steers the entire team's trajectory over the next year or two.
For solo business owners, the pressure takes another form. If you've sped up your work with AI, what did you fill that time with? If you moved toward processing more of the same kind of work, you may find yourself pulled into price competition or pressed for ever-shorter turnarounds. From a client's standpoint, when freelancers who use AI and those who don't are indistinguishable by their deliverables, the basis for pricing changes.
Hiring is shifting too. People who have run recruitment for years mention the same thing: the focus has moved away from which tools a candidate can use, toward how that person approaches problems and what context went into their work. Applicants who include—beyond the portfolio—a written account of "why I chose this direction in this project" are getting a different reception in interviews.
I don't think this is simply a matter of self-promotion skills. Being able to write that account is evidence that you still hold the logic of your work in your own head. If you can hand the draft to AI and still articulate where it fell short, how, and in what direction you revised it, that is the signal that your capability is alive.
One check helps. Of the work you handed to AI over the past month, for what percentage can you explain how the result differed from what you would have produced yourself? If you can't explain it, that means the value you added to that work is invisible from the outside. Recruiters, clients, team leads—they all ask the same question: what does this person add on top of the AI?
The confession that 729 people on Hacker News upvoted gave voice to a shared sensation among those who, in riding AI's speed, have gradually pared away their own thinking. The more AI accelerates the work, the more conspicuous the person becomes who can point to exactly where they intervened and what they changed.
