AI skeptic Ed Zitron claims generative AI is a 'scam' with overhyped capabilities and financials.

TEXT START: Ed Zitron, founder of PR firm EZPR and a well-known AI skeptic, stated on the podcast The Diary Of A CEO that he believes generative AI is essentially a "scam," as the industry has vastly overstated what AI can currently do, what it might achieve in the future, and the true financial realities behind it.

The Dissection

The text performs a legitimate financial autopsy, then smuggles in a technological obituary. It identifies overclaiming, loss-making providers, dependence on major-company funding, and opaque revenue metrics, then treats those weaknesses as proof that the entire AI trajectory is fraudulent.

It collapses three separate questions: whether current vendors are profitable, whether current products are reliable, and whether AI can eventually dominate the economics of cognitive labor.

The Core Fallacy

The core error is presentism and category confusion. A loss-making, unstable product can still be the early form of a technology that later becomes cheaper, embedded, and strategically unavoidable.

Under the Discontinuity Thesis, the decisive test is P1: durable AI cost and performance superiority across cognitive work. The supplied argument establishes neither the failure of P1 nor anything decisive about P2 or P3. It attacks the capitalization layer and mistakes it for the automation layer.

Current financial weakness may disprove inflated valuations and marketing claims. It does not disprove eventual labor displacement.

Hidden Assumptions

• Future capability is bounded by present product performance.
• Current losses imply eventual failure rather than subsidized scaling or transition spending.
• Revenue disclosure is a sufficient proxy for productive impact.
• Expensive cloud software cannot become a general labor substitute.
• Exaggerated claims about replacing all jobs or curing cancer invalidate the narrower automation thesis.
• The vendors’ current business models are the only route through which AI can reshape production.

Social Function

Partial truth with an ideological-anesthetic function. The critique punctures fantasy about present profitability, but offers a convenient escape hatch: if the financial story is fraudulent, observers can conclude that the labor threat is fraudulent too. That postpones adaptation while leaving capital owners free to deploy the technology once the economics harden.

The Verdict

Zitron’s critique is useful as a forensic report on hype and current AI finances, but "AI is a scam" is structurally lazy. The evidence can kill inflated narratives and weak vendors; it cannot kill the Discontinuity Thesis. At most, it shows that the transition is expensive, subsidized, and narratively inflated—not that the employment-to-consumption circuit survives. The salesman may be fraudulent. That does not prove the machine is imaginary.