AI Is Not Coming For Your Job. But Someone Using AI Might Be.
The headlines have been running the same story for three years now: "AI will take X million jobs." The number changes depending on the consultant, the week, and the desired reaction. The anxiety is real. But the framing is almost entirely wrong.
The Automation Panic Is Old
Every major technological shift has produced the same panic. The printing press was supposed to put scribes out of work—it did, but it created a vastly larger market for writers. The industrial loom threatened hand-weavers—it did displace them, but it made textiles accessible to everyone and grew the entire industry by orders of magnitude.
This is not to say disruption is painless. It is not. Real people lose real livelihoods in these transitions. But the macro narrative of technology destroying work has consistently proven false. Technology reshapes work.
What AI Is Actually Doing
Large language models are not replacing human cognition. They are compressing the time it takes to produce first drafts, to explore solution spaces, to synthesize information, to generate code that needs review.
A lawyer who used to spend three hours researching case precedent can now do it in twenty minutes. That lawyer is not unemployed. They are three times more productive. The question is: what do they do with those two freed hours?
This is where the real disruption lives—not in mass unemployment, but in productivity stratification.
"The pessimists sound sophisticated. The optimists build the future." — Kevin Kelly
The Two Futures
I see two futures unfolding simultaneously, and which one you inhabit is largely a choice you are making right now.
Future A: You treat AI as a gimmick, a threat, or something to wait and see about. You continue working the way you always have. You are not replaced by AI—you are gradually outcompeted by colleagues, freelancers, and small teams who use AI to produce more, faster, at higher quality.
Future B: You treat AI as the most powerful leverage tool you have ever been given. You spend time learning which tasks it does well, which it does badly, and how to combine its speed with your judgment, taste, and domain expertise. You become effectively superhuman within your specialty.
Neither future requires you to become a machine learning engineer.
The Skills That Compound
Here is what I have observed in practice: AI amplifies the skills you already have. A mediocre writer using AI produces mediocre content faster. A genuinely skilled communicator using AI produces excellent content at scale.
The skills that matter more than ever:
- Taste. Knowing what good looks like. AI can generate; it struggles to evaluate.
- Domain expertise. Knowing which AI outputs are plausible versus correct.
- Judgment. Deciding which problems are worth solving, which solutions are worth pursuing.
- Communication. The ability to articulate complex ideas clearly becomes more valuable as AI handles the mechanical parts of writing.
These are deeply human skills. They are learned through experience, feedback, and genuine intellectual engagement. They cannot be automated.
What To Do About It
Stop treating AI tools as a monolith. Experiment specifically. Pick the three most time-consuming, repetitive, or frustrating tasks in your work. Spend a week trying to use AI to help with each one. Most will be partially helpful. Some will be transformative. A few will not work at all.
That empirical knowledge—knowing exactly what AI can and cannot do in your specific domain—is itself a competitive advantage. Most people are still forming opinions based on fear or hype rather than hands-on experience.
The window for gaining that advantage is not infinite. The people who experimented with the web in 1996 built very different careers than those who "figured it out later."
This is one of those moments. What you do with it is up to you.
