By Christopher Combs, Chief Investment Officer, Silicon Valley Capital Partners
February 26, 2020
Artificial intelligence has become a focal point for modern economic anxiety. At global forums, from Davos to Silicon Valley, the debate is no longer whether AI will reshape the economy, but whether it will undermine it. The prevailing narrative is dark: jobs eliminated at scale, white‑collar work hollowed out, and wealth concentrated ever more tightly among those who own the machines.
This concern is understandable. Every major technological transition produces dislocation before it produces prosperity. Yet history suggests that today’s dominant framing may be missing something essential.
What if artificial intelligence follows a far more familiar path, one resembling electricity, the microprocessor, or the internet itself? What if AI proves to be a general‑purpose technology that raises productivity, expands margins, lowers barriers to entry, and ultimately creates more opportunity than it destroys?
That possibility is not speculative optimism. It is consistent with centuries of economic experience.
Corporate Margins and the Quality of Growth
Consider the software‑as‑a‑service sector, often cited as an early beneficiary of artificial intelligence. SaaS companies already operate with structurally high gross margins and near‑zero marginal distribution costs. AI does not disrupt this model; it reinforces it.
Machine‑learning applications, from automated customer support and predictive maintenance to dynamic pricing and personalized onboarding, allow firms to reduce costs while simultaneously increasing customer lifetime value. McKinsey estimates that AI could add approximately $13 trillion to global economic output by 2030, with the majority of gains coming from productivity improvements rather than pure labor substitution (McKinsey Global Institute, 2018).
Crucially, AI‑enabled features are not merely defensive cost controls. Automated analytics, anomaly detection, and real‑time decision‑support tools deepen customer integration and increase switching costs. These characteristics historically translate into higher margins and more stable revenue streams.
Economic history shows that when general‑purpose technologies emerge, early adopters enjoy a period of expanded profitability before competition erodes excess returns. Steam power, electrification, and computing all followed this pattern (Brynjolfsson and McAfee, 2014). Artificial intelligence is unlikely to be an exception.
If those higher margins are reinvested into research, development, and service quality, the resulting efficiency gains can cascade through entire industries, logistics, healthcare, manufacturing, and finance, producing compounding benefits across the economy.
Innovation at Lower Cost
Technological revolutions rarely deliver immediate productivity gains. Instead, they move through stages: invention, overinvestment, disappointment, and finally broad diffusion. The electrification of American manufacturing illustrates this well. Factories had to be redesigned around electric motors before productivity surged in the 1920s (David, 1990).
Artificial intelligence now appears to be entering its diffusion phase.
By dramatically lowering the cost of experimentation, AI compresses innovation cycles. Software can be prototyped in days rather than months. Drug‑discovery simulations, supply‑chain optimization, and operational stress testing can be conducted virtually at a fraction of historical cost.
Goldfarb, Agrawal, and Gans argue that AI’s defining economic impact is its ability to make prediction cheap (2018). When prediction costs fall, organizations restructure how decisions are made. Activities that were once uneconomic become viable, and entirely new business models emerge.
History is unambiguous on this point: when the cost of prediction, communication, or transportation declines sharply, innovation does not slow. It accelerates.
Employment, Automation, and Adaptation
Few concerns surrounding AI are as emotionally charged as employment. The fear that machines will permanently displace human labor has accompanied every major wave of automation.
Yet the historical record tells a more nuanced story. In 1900, roughly 40 percent of the U.S. workforce was employed in agriculture. Today, that figure is under 2 percent. The jobs did not vanish; they shifted into manufacturing, services, healthcare, education, and entirely new sectors.
Autor (2015) demonstrates that automation tends to replace tasks rather than whole occupations. Often, it increases demand for complementary human skills. The introduction of automated teller machines reduced routine teller functions but ultimately led to an expansion in bank branches and sustained growth in teller employment as roles shifted toward customer relationships.
The OECD reaches a similar conclusion, estimating that only a minority of jobs face a high risk of full automation. Most will undergo task reconfiguration rather than elimination (OECD, 2019).
Artificial intelligence will almost certainly displace routine analytical work. But it will also create demand for data governance, model supervision, ethics oversight, human‑machine interface design, and hybrid roles that combine technical literacy with domain expertise. The labor market has demonstrated remarkable adaptive capacity over time. There is little empirical evidence that AI fundamentally alters this dynamic.
Competition and the Democratization of Capability
At first glance, AI appears to favor large incumbents with vast data resources. In practice, cloud infrastructure and open‑source frameworks are democratizing access to advanced tools.
Mid‑sized manufacturers can now deploy optimization algorithms once available only to global conglomerates. Startups can access machine‑learning models through cloud APIs without building proprietary infrastructure. Research on technology diffusion shows that productivity gains are often strongest among laggard firms once adoption spreads (Andrews, Criscuolo and Gal, 2016).
This dynamic also applies internationally. Emerging economies that leapfrog legacy infrastructure, adopting mobile banking rather than building branch networks, cloud systems instead of on‑premise data centers, can experience disproportionate growth. Economic theory and historical evidence both suggest that late adopters often converge faster by importing frontier technologies (Barro and Sala‑i‑Martin, 2004).
Rather than entrenching inequality, artificial intelligence may narrow productivity gaps across firms and nations.
Productivity and Living Standards
The most consequential question is macroeconomic. Global productivity growth has slowed since the early 2000s, fueling concerns about secular stagnation. Yet history shows that general‑purpose technologies often take decades to register in national statistics.
The productivity resurgence of the late 1990s followed years of incremental computer adoption (Jorgenson, Ho and Stiroh, 2008). Brynjolfsson, Rock, and Syverson describe this phenomenon as a “productivity J‑curve,” where heavy upfront investment depresses measured productivity before transformational gains emerge (2017).
If artificial intelligence follows this pattern, today’s investment wave may be laying the foundation for sustained productivity acceleration in the coming decade.
Historically, productivity growth has been the single most powerful driver of rising living standards. It supports wage growth, corporate profits, public revenues, and social investment (Gordon, 2016). The resulting wealth effects are not theoretical abstractions; they define modern economic progress.
Education and Human Capital Formation
Education inequality remains one of the most persistent structural challenges facing the global economy. AI‑enabled learning systems offer a potential breakthrough.
Adaptive tutoring platforms can tailor instruction to individual learning speeds, languages, and styles. Early evidence suggests measurable improvements in standardized test performance, particularly among disadvantaged students (Escueta et al., 2020).
If artificial intelligence reduces the marginal cost of high‑quality instruction, access to world‑class educational content could become largely independent of geography or income. Historically, expansions in mass education have preceded periods of rapid economic growth. If AI lowers the cost of delivering that education, the long‑term human‑capital dividend could be substantial.
A Longer Historical Perspective
Every major technological transition has produced disruption alongside progress. Steam power displaced artisans. Electricity rendered entire industries obsolete. The internet dismantled classified advertising. Each transition generated inequality, regional dislocation, and political tension.
Yet over time, each expanded economic capacity and improved average living standards.
Artificial intelligence is unlikely to defy this historical pattern. To believe otherwise requires an extraordinary claim: that innovation no longer creates complementary demand; that human ingenuity has become obsolete; that productivity gains cease to translate into shared prosperity.
History offers little support for such pessimism.
The more interesting question, then, is not whether AI will destroy the job market, but whether we are underestimating its capacity to elevate it.
References
Andrews, D., Criscuolo, C. and Gal, P.N. (2016) ‘The best versus the rest: The global productivity slowdown, divergence across firms and the role of public policy’, OECD Productivity Working Papers, No. 5.
Autor, D.H. (2015) ‘Why are there still so many jobs? The history and future of workplace automation’, Journal of Economic Perspectives, 29(3), pp. 3–30.
Barro, R.J. and Sala‑i‑Martin, X. (2004) Economic Growth. 2nd edn. Cambridge, MA: MIT Press.
Brynjolfsson, E. and McAfee, A. (2014) The Second Machine Age. New York: W.W. Norton.
Brynjolfsson, E., Rock, D. and Syverson, C. (2017) ‘Artificial intelligence and the modern productivity paradox’, NBER Working Paper No. 24001.
David, P.A. (1990) ‘The dynamo and the computer: An historical perspective on the modern productivity paradox’, American Economic Review Papers and Proceedings, 80(2), pp. 355–361.
Escueta, M., Quan, V., Nickow, A.J. and Oreopoulos, P. (2020) ‘Education technology: An evidence‑based review’, NBER Working Paper No. 23744.
Goldfarb, A., Agrawal, A. and Gans, J. (2018) Prediction Machines. Boston: Harvard Business Review Press.
Gordon, R.J. (2016) The Rise and Fall of American Growth. Princeton: Princeton University Press.
Jorgenson, D.W., Ho, M.S. and Stiroh, K.J. (2008) ‘A retrospective look at the U.S. productivity growth resurgence’, Journal of Economic Perspectives, 22(1), pp. 3–24.
McKinsey Global Institute (2018) Notes from the AI frontier: Modeling the impact of AI on the world economy. San Francisco: McKinsey & Company.
OECD (2019) The Future of Work: OECD Employment Outlook 2019. Paris: OECD Publishing.