In an era of breathless headlines predicting a 20% unemployment rate, Nominal News offers a necessary, data-driven reality check: the current wave of generative AI is likely an incremental evolution of automation, not the sudden apocalypse of work many fear. By grounding the debate in decades of economic modeling rather than sci-fi speculation, the piece argues that the labor market's resilience is not a fluke, but a predictable outcome of how technology actually interacts with human tasks.
The Task, Not the Title
The article's most valuable contribution is its refusal to treat job titles as monolithic blocks. "Jobs are usually very vaguely defined or entail a broad set of activities," the piece notes, pointing out that a title like "engineer" or "consultant" hides the specific activities that make up the role. Instead, economists must look at "tasks"—the granular actions like "finger dexterity" or "setting limits, tolerances and standards"—to determine what can truly be automated.
This distinction is crucial because it explains why past predictions of mass unemployment failed to materialize. The editors highlight the work of Frey and Osborne (2017), who famously predicted that 47% of US jobs were at risk of computerization. "As can be seen from the US labor market – this has not happened," the article observes, noting that unemployment remains near historic lows. The failure of that prediction wasn't a lack of technology, but a flaw in the model's assumptions about what constitutes a "routine" task versus one requiring complex social or physical intelligence.
"Predicting the future outcomes to the economy are highly dependent on the modeling assumptions we make, and it is important that we are aware of them when considering the impacts to the economy."
This framing effectively dismantles the panic around self-driving trucks or automated customer service. While the article acknowledges that some roles are vulnerable, it emphasizes that tasks requiring "perception and manipulation" or "social intelligence" remain stubbornly resistant to current algorithms. This aligns with the historical context of the Jevons paradox, where efficiency gains in one area often expand demand in others, rather than simply erasing the need for labor.
The Displacement and the Productivity Effect
The commentary then shifts to the economic mechanics of automation, introducing a nuanced view that avoids the binary trap of "robots take jobs" or "robots create jobs." The piece explains that automation creates two opposing forces: a "displacement effect" that reduces labor demand, and a "productivity effect" that increases it.
To illustrate this, the editors cite the classic example of the Automated Teller Machine (ATM). "Even though ATMs completely replaced banking tellers, more banking 'tellers' were actually hired!" the article reports. The logic is counterintuitive but sound: by lowering the cost of operating a branch, the technology allowed banks to open more locations, which in turn required more humans to handle complex services like loans and financial advice.
However, the piece warns that this productivity effect is not guaranteed. It hinges on whether the technology is "brilliant" enough to drastically lower costs or if it is merely "so-so" AI that offers marginal improvements. "If the AI does very complex tasks... this would be preferable, than if the AI were to do only basic tasks," the editors argue. This distinction is vital for understanding why the current generation of Large Language Models might not trigger the massive job creation seen in previous industrial shifts.
Critics might note that this optimistic view relies heavily on the assumption that new tasks will emerge fast enough to absorb displaced workers. History shows that the transition can be painful and prolonged, even if the long-term net effect is positive. The article touches on this by referencing Daron Acemoglu's work on job polarization, where middle-income jobs disappear, pushing workers into either high-skill or low-skill sectors, often with a net loss in wages for the displaced.
The Reality of Current AI Capabilities
Perhaps the most sobering section of the piece is its assessment of the current state of artificial intelligence. Rather than assuming a path to Artificial General Intelligence (AGI), the editors lean on expert skepticism. "The key founders behind machine learning are also skeptical that the current AI method... will result in true AI," the article states, citing doubts about the ability of current models to replicate human reasoning.
Empirical evidence supports this cautious stance. The piece references a study by Noy and Zhang, which found that workers often use AI as a "replacement for effort rather than a complement to their effort." In other words, the technology is being used to cut corners, not to augment human capability in a way that expands the pie. Furthermore, a study on entrepreneurs in Kenya found that AI advice only helped the already skilled, while hurting those who lacked the expertise to ask the right questions.
"All of this underlines that the new AI technologies are more likely to continue the job polarization trend described by Autor and Dorn."
This suggests that the immediate future of work is not a sudden collapse, but a gradual reshuffling where expertise becomes the primary differentiator between success and failure. The "reinstatement effect," where entirely new tasks are created, will likely be slow and generational, not immediate.
Bottom Line
Nominal News delivers a compelling, evidence-based argument that the panic over generative AI is largely misplaced, rooted in flawed models that ignore the complexity of human tasks. The piece's strongest asset is its reliance on historical precedents and specific economic mechanisms, such as the ATM example, to demystify the automation process. Its biggest vulnerability, however, is the assumption that the "productivity effect" will inevitably dominate; if the current AI proves to be merely "so-so" technology, the displacement of middle-skill workers could outpace the creation of new roles, leading to significant short-term inequality even without mass unemployment. Readers should watch for how quickly the "reinstatement effect" actually materializes in the next two years, as that will determine whether this is a moment of adjustment or a period of structural pain.