General Purpose Technologies, Innovation, and the Role of Venture Capital in Economic Transformation

This episode delves into the "new economic space" created by transformational technologies, focusing on the role of venture capital and the complex dynamics of General Purpose Technologies (GPTs). It highlights that while investors in foundational infrastructure (like railways or the internet) may incur significant losses, the social returns from these investments far outweigh private losses, as the infrastructure remains even if companies fail. The discussion emphasizes that it takes considerable time, often decades, to identify and exploit the commercially valuable applications (killer apps) of new technologies, maturing through levels of abstraction and trial-and-error experimentation. Tim Bresnahan's definition of a GPT is central, characterized by positive feedback between core technology improvements and the invention/deployment of novel application sectors, leading to "innovation complementarities."

The podcast explores critical features of GPT deployment, such as network construction (Metcalfe's Law, which is noted as incomplete), and the heterogeneous impact of GPTs across economic sectors. It distinguishes between social and private returns, noting that inventors (both core and application-specific) cannot capture all the value they create, leading to externalities. A key insight is that improvement in the core GPT enables new application sectors, and the ability to envision new markets is as crucial as reducing implementation costs. The discussion also touches on the challenges of complementary inventions, demand forecast errors, and how existing infrastructure can impede the deployment of new technologies, citing the example of electricity's impact on manufacturing productivity only after systemic changes like the electricity grid and unit drive motors.

The episode addresses the "productivity paradox" (Solow's wisecrack about computers everywhere but in productivity statistics), explaining that the full economic impact of GPTs diffuses over long periods and requires significant investment in intangibles like training, which are often understated in GDP numbers. It references Schumpeterian growth theory, suggesting that national income may temporarily fall during the transition between discovery and full implementation of a GPT. The podcast revisits historical analyses, contrasting Robert Fogel's neoclassical view of railroads with Homeback and Rottenberg's finding of a much greater impact due to radical reallocation of manufacturing and market access gains. It also notes the state's role, exemplified by the U.S. Department of Defense, in accelerating GPT development and fostering new application sectors through both incentives and pressures.

Finally, the discussion shifts to models of creative destruction, particularly Schumpeter Mark III, focusing on the division of innovative labor between frontier firms and established firms. A Duke University survey highlights the significance of external sources of invention, with customers driving incremental improvements and tech specialists providing more valuable, fundamental innovations. The episode concludes by validating the disproportionate role of startups and venture capital in providing a reservoir of new inventions that feed the overall innovation system, even if not all startups succeed commercially. The next lecture is previewed to cover the maturation of the digital economy, industrial concentration, inequality, and declining economic dynamism.

"the losses of investors in building the internet as in building railways a century before were outweighed by the social returns from what the investors financed"

"it takes time to identify and exploit the financially and commercially valuable uses of the new technology the killer apps"

"the key attribute of a gpt is the positive feedback between improvements in the core technology and invention and deployment of novel sectors"

"protecting intellectual property is not always socially optimal"

"social increasing returns are generated by positive feedback between advances in the core gpt technology and in the application sectors"

"the ability to envision new markets matters not just lower costs of implementing the technology"

"around 1989 we see the computers everywhere but in the productivity statistics"

"during the period between the discovery of a new gdp gpt and its ultimate implementation national income will fall"

"my customers wanted a faster horse they didn't want an automobile"

"startups in this division of innovative labor play a disproportionate role"