What this covers

Azeem Azhar and Russ Roberts discuss the mechanisms and consequences of technological acceleration in modern economies. Azhar traces how multiple technologies—solar panels, batteries, semiconductors—are improving at compounding exponential rates, driven by Wright's Law (the principle that per-unit costs decline with cumulative production) and by information technology's ability to codify and share knowledge across firms and industries. The conversation moves from these cost declines into their economic and social consequences: how rapidly adopted technologies like smartphones and large language models are reshaping markets and human behavior faster than institutions and cultural norms can adjust. Roberts and Azhar probe what Azhar calls the exponential gap—the widening mismatch between the pace of technological change and the speed at which humans and their governance structures can meaningfully respond.

The discussion spans several distinct territories. It covers the mechanics of exponential improvement: Learning by Doing, the role of modularity in solar's cost collapse, and how information compression through large language models represents a qualitative shift in how economies process data. It then turns to adoption speed, examining why TikTok and ChatGPT reach billions of users far faster than technologies of previous eras. The hardest terrain concerns smartphone harms and engagement incentives—both acknowledge documented social harms alongside unfounded moral panics, and debate whether structural economic fixes (like solar's price advantage) or institutional design are viable approaches to harm reduction. Azhar sketches potential benefits in developing economies, where large language models and mobile infrastructure might inject capability independent of institutional decay. Throughout, the speakers wrestle with whether the problem is simply keeping pace with change, or something more fundamental: learning to govern at human speed vs silicon speed rather than chasing technological velocity.

Sharpest takeaway

Azhar argues that multiple technologies are improving at compounding exponential rates driven by learning curves and information compression, and that humanity now faces an 'exponential gap' where the pace of change may become objectively too fast for human institutions and norms to handle, requiring us to learn to govern technology rather than keep up with it.

  • Exponential cost declines are appearing across computing, solar, batteries, and manufacturing, not just chips
  • Wright's Law (learning by doing) plus IT-accelerated knowledge sharing explains and predicts these declines
  • Our norms, regulations, and social institutions have not adapted to the speed and volume of change, especially around smartphones and immersive tech

The argument · threads3 threads · 33 claims
0.85

Smartphones represent a qualitative leap in social isolation that society's norms have failed to constrain or resist.

2 pointscentrality 5/5
  • Society has not adjusted its norms to cope with smartphones; if anything, the social acceptability of ignoring people around you to devote yourself to your screen has increased steadily over the last 10-15 years, and immersive devices like Apple's Vision Pro threaten to accelerate isolation by hiding the eyes and smile that mark our humanity.

    “I would argue that we haven't adjusted our norms very much. And, if anything, we just continue to accelerate into isolation.”

  • Humanity has coped reasonably well with accelerating quantitative change—as with the car, despite debates over whether suburbs are anti-human—but the smartphone represents a ratcheting up not just in quantity but in quality of how life has changed, raising the question of whether this is the point at which the proverbial frog gets boiled.

    “we've coped pretty well with the car... But, the cell phone seems to be a ratcheting up of--it's not just quantity, it's quality, of how things have changed.”

0.56

Information volume from modern technology overwhelms human capacity, requiring institutional and normative solutions we may not develop.

1 pointcentrality 4/5
  • The problem with information technology is not merely time compression (learning of an event in seconds rather than days) but the volume—like drinking from a fire hose—which humans are not good at handling, so we need institutions and norms to redirect, slow, or selectively access the flow, even though it is unclear whether we can develop them.

    “It's not just the compression. It's the volume. It's like drinking from a fire hose. And I don't think human beings are good at drinking from a fire hose.”

0.29

Specialization and market competition create continuous improvement cycles poorly captured by static economic models.

1 pointcentrality 2/5
  • Adam Smith's principle that the division of labor is limited by the extent of the market is, at its core, as much about learning by doing as about market size: specialization lets individuals focus on and continually improve a process, and combined with globalization-driven competition creates a dynamic feedback loop of falling prices and rising quantity that is poorly captured by static supply-and-demand pictures.

    “Smith writes that the division of labor is limited by the extent of the market... it's more about--its as much about learning by doing.”