AI Derangement Syndrome

I didn’t make up this term. People have been using it before. I’m describing it as a condition in which otherwise intelligent people lose the ability to think clearly about artificial intelligence and its impact on the economy and labor market. The symptoms are persistent, largely immune to data, and tend to worsen with each new model release.

After two years of tracking AI related discussions, I’ve classified the syndrome into five common variants.

Variant 1: The Tribal Reflex. The patient opposes AI not because of what it does, but because of who likes it. AI has become culturally coded — associated with tech billionaires and the current administration. For a certain kind of person, that’s enough. If Elon is excited, it must be bad. If the White House is promoting it, it is even worse. The analysis never begins, because identity did the work first.

Variant 2: The Ostrich Effect. The patient avoids engaging with AI’s capabilities because the implications are uncomfortable. If your job involves synthesizing information, producing text, or advising clients, the honest version of this conversation is unsettling. So the mind does what minds do: it looks away. “It can’t do what I do,” says the professional who has never once tested that claim. This isn’t stupidity. It’s self-preservation. But the labor market does not grade on coping.

Variant 3: The $20 Diagnosis. The patient has issued a confident verdict on frontier AI without ever actually using it. They tried a free-tier chatbot, got a mediocre answer, and declared the emperor has no clothes. The gap between free-tier and frontier models is enormous — and that’s before you get to tools like Claude Code, which turn AI from a chatbot into an actual working colleague. But to see any of this, you’d have to spend $20 a month. For reasons I still find baffling, highly paid professionals who drop $7 on coffee refuse to invest a couple of lattes a month in understanding the technology reshaping their industry.

Variant 4: The Moral Shield. The patient raises real concerns — energy use, privacy, bias, labor displacement — but uses them to avoid the more basic question: what can this technology actually do, and how fast is it improving? Yes, there are legitimate ethical debates. But when “AI is boiling the oceans” becomes a reason to dismiss its economic impact entirely, you’ve stopped doing analysis and started doing moral sorting.

Variant 5: The 2023 Fossil. The patient made confident predictions two years ago — it’s a fad, it’ll plateau, it’s just autocomplete — and has been unable to update since. Every new breakthrough gets processed through the original thesis rather than challenging it. This is anchoring bias with credentials.

Patient Profile

The typical sufferer is a knowledge worker: highly educated, often credentialed, often influential — employed in media, academia, policy, consulting, or professional services. Politically, the syndrome skews left and center-left, driven partly by tribal association and partly by the fact that the professions most exposed to AI disruption simply skew progressive. But you’ll find it across the spectrum.

The deepest irony: the left-leaning sufferers claim to care most about workers — but their refusal to engage seriously with AI means they’re the least equipped to help workers navigate what’s coming.

A quick human note: I don’t think people are “crazy” for feeling threatened. When the ground is shifting under your status, income, and identity, that’s genuinely scary. The goal here isn’t to mock that fear—it’s to push us toward a clearer read on reality, because clarity is what makes adaptation and good policy possible.

We’re early in what may be the most significant transformation of knowledge work in a generation, perhaps a century. Decisions being made right now — by companies, policymakers, educators — will shape whether this transition creates broad prosperity or concentrated pain.

Those decisions cannot be made well by people in denial about what’s happening.

The treatment is simple: Pay for a frontier model. Use it seriously for a month. Apply it to your actual work. Try research. Try drafting. Try analysis. Try the tools, not just the chat.

Then form your opinion.