AI, creative destruction, and digital advertising

The Royal Swedish Academy of Sciences announced today that its annual Nobel Prize in Economics has been awarded to economists Joel Mokyr, Philippe Aghion, and Peter Howitt. From the prize announcement (emphasis mine):

This year’s prize relates to the explanations for sustained growth based on technological innovation. Economic historian Joel Mokyr is rewarded with one half of the prize for his description of the mecha­nisms that enable scientific breakthroughs and practical applications to enhance each other and create a self-generating process, leading to sustained economic growth … The other half of the prize is awarded to the economists Philippe Aghion and Peter Howitt. In a joint publica­tion from 1992, they constructed a mathematical model of how companies invest in improved pro­duction processes and new, better-quality prod­ucts, while the companies that previously had the best products are outcompeted. Growth arises through creative destruction, leading to sustained economic growth.

The award, formally named the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel, amounts to roughly $1.2MM: one-half was awarded to Mokyr, a Northwestern University Economics professor, and the other half was awarded jointly to Aghion, from INSEAD and the LSE, and Howitt, from Brown University.

All three economists’ research focuses on the conditions for economic growth. Mokyr, an economic historian, is the author of A Culture of Growth: The Origins of the Modern Economy, which explores the origins of the Industrial Revolution in the late 18th century as a function of a “culture of growth”: a pervasive set of ideas, norms, and attitudes that allowed for technological progress to take root, giving rise to dramatic economic growth. Mokyr argues in the book that the Industrial Revolution instantiated a paradigm in which growth becomes self-sustaining through an interrogative process related to the origins of that growth: that a scientific understanding of why technological innovations serve as catalysts for economic growth is critical in replicating that growth. Aghion and Howitt jointly authored A model of growth through creative destruction, a 1992 paper that proposes an analytical framework for modeling the process of creative destruction as theorized by Schumpeter.

It’s difficult to separate this award from the broader, systemic influence of AI, which is the prevailing economic innovation of this moment, accounting for up to $400BN in CapEx this year from American “Big Tech” firms alone and an estimated cumulative total of $3TN by 2028. From a Schumpeterian perspective, AI is poised to deliver economic growth by replacing older, less efficient working methodologies with new approaches empowered by technological innovation. This is systemic and disruptive but ultimately expansionary in a way that benefits not just the firms that deploy it but society broadly, in the aggregate.

I invoked Schumpeter’s creative destruction in a recent podcast, Commerce at the limit, in which I argue that digital advertising is both a critical component of economic growth and provides demonstrable welfare benefits for society. Which is why I find the endless hand-wringing about whether AI can create or has created any positive economic impact to be so misplaced. American Big Tech firms are principally responsible for much of the progress that has been made in AI to date. Google famously invented the concept of self-attention and the transformer model in neural networks; two of the most popular deep learning development frameworks, PyTorch and TensorFlow, were created by Meta and Google, respectively; researchers at Microsoft Research Asia created ResNet, which paved the way for modern computer vision; Meta’s FAIR helped advance and scale the use of convolutional neural networks (CNNs) for computer vision. These paradigm shifts were financed and developed by scaled companies that pursued them for their own economic interests: in the cases of Meta and Google, the commercial focus of these investments is improved advertising efficiency.

And as I point out in AI is not the Metaverse, these investments have already produced real returns. From that piece:

AI is not the Metaverse. The Metaverse was an alluring, if fanciful, hypothetical extrapolation of engagement patterns that only existed during COVID and which has failed to coalesce into a scaled consumer use case. Even if the superintelligence initiative is mostly a marketing campaign, AI is generating real, material commercial value for Meta and Google currently. The application of AI to advertising doesn’t represent a far-flung, fundamental overhaul of computing: it’s an accelerant to the existing advertising structure that benefits — right now, as I write this — from compounding effects and expanded economic activity.

What the Nobel committee’s decision in this award affirms, implicitly, is that sustained economic growth — with obvious caveats and thoughtful optimizations — is its own reward. In this moment, AI is a technological catalyst for growth, and digital advertising, as the engine of the internet economy, is the instrument with which its impact has already been (and will likely continue to be) realized.