Understanding the AI Revolution: Insights from Matt Shumer’s Viral Essay
In a compelling essay titled “Something Big Is Happening,” Matt Shumer, founder and CEO of OthersideAI, draws an intriguing parallel between the current state of artificial intelligence (AI) and the early days of the Covid-19 pandemic. Shumer posits that AI has evolved beyond being merely a helpful assistant to potentially becoming a general cognitive substitute. The ability of AI to not only assist but also enhance its capabilities signifies a monumental shift that could soon yield systems rivaling many human experts.
Shifting Perspectives on AI
While experts in the field are well aware of the transformative changes on the horizon, the average person—referred to as “normies”—might find themselves unprepared. To use a pandemic-era metaphor, one might say that “Tom Hanks is about to get sick,” indicating that widespread public realization could arrive too late.
This sense of urgency is underscored by the recent resignation of Mrinank Sharma, who led the safety team at Anthropic. His farewell letter expressed grave concern, warning of “interconnected crises” and highlighting the pressures companies face in pursuit of massive valuations—Anthropic is gunning for a lofty $350 billion. Such developments have intensified fears among those already wary of AI’s rapid advancement.
The Road to Artificial General Intelligence (AGI)
The question on many minds is whether AI will soon fulfill various definitions of “weak AGI.” Plenty of technologists believe it could happen, while others, like Google DeepMind CEO Demis Hassabis, indicate that we still need significant technological breakthroughs before achieving true AGI.
Yet, rather than focusing solely on technological advancements, it’s essential to consider the economic factors that could impact the deployment of AI. The journey from “impressive AI models” to widespread integration is often slower and more complex than anticipated.
The Economic Bottlenecks
Historically, it has taken decades for new technologies to reshape industries. For example, electrification redesigned factories, and the internet gradually transformed retail. Currently, less than 20% of U.S. businesses have integrated AI into their operations. Large, risk-averse institutions require significant investments in data infrastructure, process redesign, compliance measures, and worker retraining to effectively deploy AI. Economists describe this phenomenon as the productivity J-curve, indicating that initial investments can lead to a dip in measured output before benefits materialize.
Richer Societies, Leisure Choices
Even if we accept the optimists’ view of quick advancements in AI capability, an important point remains: increased output does not equate to more work. Richer societies often opt for more leisure time, resulting in earlier retirements and shorter workweeks. Economist Dietrich Vollrath notes that higher productivity does not automatically lead to faster economic growth if households choose to supply less labor, potentially leading to moderate overall growth despite substantial welfare improvements.
The Constraints of Slow Sectors
The sectors that tend to grow the slowest can set limits on overall growth. Even if AI dramatically reduces the costs of certain services, demand in areas resistant to automation—such as healthcare and education—will not expand infinitely. This phenomenon is encapsulated by the “Baumol effect,” which explains how labor-intensive sectors can consume more income as wages rise, despite slower productivity growth.
The Narrowest Pipe Principle
As economist Charles Jones explains, in a system composed of interconnected components, the narrowest element dictates the flow. AI may enhance various processes such as coding and drafting, but if other vital components — energy infrastructure, regulatory approvals, or human decision-making — proceed at a conventional pace, these will pose constraints on economic growth.
Gradual Change in Complex Systems
Economies are intricate, dynamic entities capable of adaptation. They create tangible products that embody complex information, as economist Cesar Hidalgo describes them. When economies evolve, they do so through gradual reorganization rather than abrupt collapses or instant transformations. This understanding should shape our expectations moving forward.
Conclusion: Embracing AI with Caution
While there is certainly a case for urgency surrounding AI adoption—as Shumer advises to leverage advanced AI tools as they emerge—there’s no need for panic-inducing rhetoric reminiscent of early 2020’s pandemic response.
This article draws on insights from the original piece, which can be found Here.
Image Credit: www.vox.com
