Episode Summary
Executive Summary: The episode argues that AI is best understood not as sci-fi “intelligence” but as a major general-purpose and cultural technology reshaping economics, labor, and social life. Cameron Abadi and Adam Tooze debate whether AI is merely another productivity tool or a potentially unprecedented innovation that could accelerate innovation itself. They also stress the hidden labor, energy, and data extraction behind AI systems and compare AI’s societal impact to the printing press, bureaucracy, and the Industrial Revolution.
Main Topics: AI as a General Purpose Technology (Priority: 5/5): The hosts frame AI alongside steam power, electrification, semiconductors, and the internal combustion engine as a technology whose effects are broad, diffuse, and time-lagged but transformative. Hyper-technology vs. Normal Transformation (Priority: 5/5): Adam Tooze explores whether AI is just a standard GPT or something more radical—an R&D-enhancing system that could speed up innovation itself and potentially produce runaway growth. Scaling Laws and Industry Strategy (Priority: 4/5): The conversation turns to whether the current data/compute/parameter scaling model will continue to work, or whether firms like Meta need to pivot toward world-modeling, memory, and richer data inputs. AI as Extractive Infrastructure (Priority: 5/5): The episode emphasizes the material and labor inputs behind AI: energy, water, land, human labeling work, content moderation, and disputed training-data extraction from books and online texts. AI as a Cultural and Social Technology (Priority: 5/5): The hosts endorse viewing AI as a symbol-generating mechanism akin to writing, print, computers, and bureaucracy—something that can reshape language, creativity, emotion, and politics. AI and Economic Planning / Market Coordination (Priority: 4/5): Tooze revisits the calculation debate in economics, arguing AI may blur the old market-versus-planning divide because modern firms already use algorithmic systems to manage preferences and logistics. Historical Analogies for Work and Society (Priority: 4/5): The Industrial Revolution, the printing press, and the rise of the office/bureaucratic state are used as precedents for understanding how AI may change white-collar work more than manual labor.
Key Arguments: AI is already widely accepted as a general-purpose technology; the real question is whether it is a normal one or a qualitatively unprecedented one. If AI enhances research and development, it could raise the pace of innovation itself, creating accelerating rather than diminishing returns to technological progress. Current AI systems may face diminishing returns if scaling laws hit limits, especially because they rely heavily on text, compute, and imperfect memory-like architectures. The hidden labor behind AI is substantial: human workers label data, perform alignment work, and anchor systems in 'ground truth.' AI is extractive not only because it consumes energy and water, but because it draws on human labor and disputed training data from books, news, and other copyrighted materials. The apocalyptic vs. normal-technology framing is too stark; AI should instead be seen as a powerful social and symbolic technology that will reshape institutions and culture. AI may alter creativity from authorship to curation, prompting, and machine control—similar to how the printing press and search engines changed knowledge work. The old socialist calculation debate is less relevant in a world where markets and bureaucracies are already heavily algorithmized and data-driven. AI’s biggest impact may be on white-collar and office work, not just factory labor, because it targets symbolic and informational tasks. Historical comparisons to the Industrial Revolution are useful, but AI’s distinctive feature is that it intervenes in mental and symbolic labor rather than manual labor.
Data Points: U.S. tech firm AI infrastructure spending in 2024: $400 billion - The episode’s opening data point for the AI boom. Projected AI infrastructure spending by 2028: $3 trillion or more - Estimated future spending level if the boom continues. BetterHelp therapist pool: 30,000 therapists - Sponsor read describing the platform’s scale. BetterHelp global reach: Over 5 million people globally - Sponsor read describing user base. BetterHelp average live-session rating: 4.9/5 - Sponsor read citing client reviews. BetterHelp client reviews: 1.7 million reviews - Sponsor read supporting rating claims. Amazon share of American retail spending: 13% - Tooze cites Amazon’s algorithmic influence in retail. Online retail share in the U.S.: Around 25% - Used to illustrate how much commerce is already algorithmically mediated. Online shopping share in China: Closer to 25%-30% - Used as a comparison for even greater algorithmic mediation.
Pivotal Quotes: "AI clearly is of the type of general purpose technologies." — Adam Tooze: Tooze argues there is broad agreement that AI is already a GPT, with debate only over whether it is something more. "What we're doing here is applying technology to thinking, which is the source of technology." — Adam Tooze: Tooze explains why AI raises the possibility of a second-order or unprecedented technological break. "AI is neither artificial nor intelligent." — Cameron Abadi citing Kate Crawford: Used to emphasize the material, labor-intensive, and extractive realities behind AI systems.
Implications: Listeners should expect AI to reshape office work, creativity, commerce, and institutions more than replace all labor outright. The biggest uncertainty is whether current scaling methods can keep delivering gains or whether the industry must reinvent AI’s core architecture.
From the Transcript
AI is there isn't really very much disagreement. I think practically everyone agrees that AI clearly is of the type of general purpose technologies. In other words, it's changing things very dramatically. It's very pervasive. It may take some time for its full effects to be felt because what is typical of general purpose technologies is it takes time for entire infrastructures to be built, societal arrangements to be shifted to accommodate and ultimately make best and full use of. You know, digitization. You know, it's a long way from the first microchip to the entire logistical system built around Amazon, for instance. But in due course, one does enable and make possible the other. The question with II is not so much is it a general purpose technology, but is it something more than just a normal general purpose technology? Is it something hyper? Is it in some sense the end point of all technological development? Because it's a technology about technology.
Because what we're doing here is applying technology to thinking, which is the source of technology. And so that then raises a bunch of other questions. I think most economists are agreed that this is a normal general purpose technology. The question is, could it be some sort of unprecedented general purpose technology? And if it were, it would again impact at the very heart of economic thinking about technological change. And this goes to a weird aspect of economic thinking about economic development, which is You might think that would be at the heart of all economic models. But if you look at standard classical, neoclassical, mid-20th century, 1950s, 1960s, neoclassical growth models, in which capital and labor are combined to produce output, which grows over time because labor becomes more productive because you add more capital, or machines are more productive because you add more labor, those models don't actually predict sustained growth. They predict convergence to some kind of level of GDP after which growth subsides.
Ways extraction is perhaps less ethically sound as a mode of economic activity than production, I guess, or manufacturing, or maybe commerce, even. You know, there's something violent that's inherent in this. And I mean, on the one hand, find that maybe illuminating. It's difficult to see why that makes AI distinctive, like it's a thing in the world. And so, you know, I mean, the phrase from Kate Cawford's book from 2022 is: AI is neither artificial nor intelligent. made from natural resources and human labor. I mean, which is, I think, as much as to say artificial is intelligence is in the world. I mean, it's made of things like natural resources and labor. Of course, it's not just about intelligence as a disembodied thought process, but, you know, as a fueled by material inputs. And we're going to need a lot of energy and water for cooling and some land to install these facilities in. I mean, I think the two aspects of this where it's worth digging.
About Ones and Tooze
Foreign Policy economics columnist Adam Tooze, a history professor and a popular author, is encyclopedic about basically everything: from the COVID shutdown, to climate change, to pasta sauce. On our new podcast, Tooze and FP deputy editor Cameron Abadi will look at two data points each week that explain the world: one drawn from the week’s headlines and the other from just about anywhere else Tooze takes us. Check out Adam Tooze’s column at https://foreignpolicy.com/author/adam-tooze/.