AI Didn't Take My Job. It Just Made Me Feel Worthless At It Anyway.

Asked an AI tool to write something recently that would've taken me a genuinely proud afternoon to draft a couple of years ago.

Got a decent first version back in about eleven seconds.

Sat there for a moment afterward feeling something I didn't immediately have a name for.

Not relief that the work was done faster.

Something closer to a quiet, sinking feeling that whatever skill I'd built up over years, the thing I used to feel genuinely good at, had just been made to look kind of unremarkable.

Nobody took my job.

That's worth saying clearly, because the feeling doesn't actually match the reality most of the time.

What actually happened was smaller and stranger, and I think a lot of people are quietly experiencing some version of it without having language for what it actually is.

This isn't really about job security.

It's closer to imposter syndrome wearing a new costume.

Worth separating these two things clearly, because they get tangled up constantly and they're not actually the same feeling.

Genuine job insecurity is about whether a role will exist.

That's a real, practical concern for some industries and roles right now, and it deserves to be taken seriously on its own terms.

What a lot of people are describing, though, isn't quite that.

It's a quieter, more personal version — a sense that the specific thing they were proud of being good at, the skill that felt like it belonged to them, has been made to look ordinary or replaceable, even in situations where the job itself is completely secure.

That's imposter syndrome's actual shape.

Not "am I going to lose my job" but "was what I thought was skill actually just... something a tool could do the whole time, and I never really needed to be as good at it as I thought."

Naming that distinction matters, because the two feelings need genuinely different responses, and treating an identity wobble as if it were a practical employment threat tends to make people spiral in a direction that doesn't actually lead anywhere useful.

Why this hits certain kinds of work particularly hard.

Not evenly distributed, this feeling.

It tends to land hardest on people whose sense of professional competence was built specifically around producing a certain kind of output quickly and well — writing, first-draft thinking, certain categories of analysis, anything where the finished product itself was the visible proof of skill.

When a tool can produce something structurally similar to that output almost instantly, it doesn't just speed up a task.

It quietly undermines the specific evidence a person had been using, often for years, to feel confident in their own competence.

The actual underlying judgment, taste, and experience that made someone good in the first place — knowing what's actually true, what's actually useful, what a specific audience actually needs — that stuff hasn't gone anywhere.

But it's suddenly harder to see or feel proud of, because the visible output that used to prove it now looks achievable by something that took eleven seconds and no experience at all.

The part almost nobody says out loud, that the AI output usually isn't actually as good as it first feels.

Worth sitting with this honestly, because the initial reaction — that sinking, deflating feeling — tends to happen in the first few seconds, before anyone's actually properly evaluated the output against what they know.

Most AI-generated first drafts, in most fields, are competent but generic.

They lack the specific judgment calls, the awareness of what actually matters to a particular audience or situation, the small decisions a skilled person makes almost invisibly because the skill is so internalised it doesn't feel like a decision anymore.

The gap between "looks similar at first glance" and "is actually as good" is usually much bigger than the initial deflating feeling suggests, and most people, once they slow down and actually compare the two properly, find that gap again fairly quickly.

The deflation tends to happen in the reflexive moment, before that proper comparison gets made.

Which means a lot of this feeling is responding to a first impression rather than an accurate, considered assessment of what actually changed.

What's actually shifted, and it's worth being honest that something real has.

Not trying to argue nothing's changed, because something genuinely has, and pretending otherwise doesn't help anyone.

The value of pure output — just producing a competent first version of something — has genuinely dropped, because that specific thing is now much faster and cheaper to get from a tool.

What hasn't dropped, and arguably matters more now than it used to, is judgment.

Knowing whether an output is actually right, actually appropriate, actually going to land with the specific person or situation it's meant for.

Knowing what question to even ask in the first place, because a tool can only respond to the framing it's given, and framing a problem well is still an entirely human skill.

Catching the subtle thing that's wrong or missing, which requires actually understanding the domain rather than just being able to generate plausible-sounding content within it.

The skill that mattered didn't disappear.

It shifted, from "can produce the thing" toward "can tell whether the thing produced is actually good, and knows how to make it genuinely good rather than just plausible."

That's a real shift, and it's worth taking seriously, but it's not the same as being made obsolete, whatever the eleven-second gut reaction suggests.

What actually helps when this feeling shows up.

Separate the feeling from the fact, deliberately, the way most imposter syndrome eventually needs to be separated.

The feeling says "I'm not needed anymore."

The fact, in most cases, is closer to "the specific task got faster, and my actual value has shifted rather than vanished."

Those are genuinely different statements and worth treating as such rather than letting the emotional one quietly stand in for the accurate one.

Actually compare the AI output properly, rather than reacting to the first-glance deflation.

Most of the time, on real scrutiny, the gap between "fast and plausible" and "actually good" reappears, and that comparison itself tends to be reassuring in a way the initial reaction wasn't.

Notice what you're doing differently now, rather than only noticing what got faster.

If reviewing, refining, and applying real judgment to an AI-assisted output is taking up more of your time than pure generation used to, that's not evidence of being less valuable.

It's often evidence that the more valuable part of the job was always that judgment, and it's just more visible now that the easier part got automated around it.

The bigger thing worth remembering.

A tool getting faster at producing a first draft doesn't mean the years spent building actual judgment, taste, and domain knowledge were wasted or unnecessary.

It means the part of the job that was always the more replaceable part became visibly replaceable, which is uncomfortable, but isn't the same thing as being told the whole skill was worthless all along.

Worth naming this feeling clearly when it shows up, rather than letting it sit there unexamined and slowly corrosive.

It's imposter syndrome, mostly, wearing a very current costume.

And like most imposter syndrome, it tends to say much less about actual competence than it feels like it does in the moment it hits.