Erik Brynjolfsson on the Second Machine Age

Episode Summary

Executive Summary: Eric Brynjolfsson argues that digital technology is advancing rapidly, but its benefits are unevenly distributed: it boosts productivity, creates “superstar” markets, and lowers the cost of many free goods while also threatening jobs, wages, and economic mobility. He is optimistic if society adapts through education, entrepreneurship, immigration, and work-supporting policies.

Main Topics: Acceleration of digital technology (Priority: 5/5): Brynjolfsson highlights recent breakthroughs in autonomous driving, Watson, and conversational interfaces as evidence that machine capability has advanced faster than expected. Driverless cars and automation limits (Priority: 5/5): He argues autonomous vehicles are already viable in constrained settings like highways, but dense urban driving still requires human judgment; future gains may come from fleets of driverless cars coordinating with each other. Jobs, wages, and inequality (Priority: 5/5): The conversation centers on whether technology raises overall prosperity while depressing median wages, increasing inequality, and creating winner-take-all outcomes for highly skilled workers and superstars. Measuring welfare beyond GDP (Priority: 4/5): Brynjolfsson stresses that GDP misses much of the value from free digital goods such as Wikipedia and the Internet, so consumer surplus and quality-adjusted measures matter more for welfare. Education and MOOCs (Priority: 4/5): He sees education as a critical policy lever, especially through digital tools and MOOCs that can scale instruction, improve measurability, and adapt teaching to student behavior. Entrepreneurship, immigration, and labor market dynamism (Priority: 4/5): He calls for policies that encourage startups, reduce regulatory barriers, reform health insurance portability, and welcome skilled immigrants to sustain job creation. Long-run adaptation and work (Priority: 3/5): Rather than a universal basic income, he prefers policies like the Earned Income Tax Credit and educational reform to keep people working while the economy transitions toward more abundance.

Key Arguments: Technology has moved from doing routine tasks to handling complex ones like driving and language, so the boundary between human and machine capability is shifting quickly. Driverless cars will likely arrive first in limited environments; urban driving remains hard because exceptions and judgment calls are difficult to automate. If multiple cars are autonomous, highways could become much more efficient through coordinated platooning and reduced spacing. Technology can increase total wealth while leaving median workers behind; growth in the pie does not guarantee broad distribution of gains. Inequality is driven by skill-biased, capital-biased, and superstar-biased technical change. Consumer surplus from free digital goods is large and understates real living standards when GDP is used alone. Education has lagged other industries in digitization, but MOOCs and data-driven learning could transform both scale and measurement. A dynamic economy needs entrepreneurship and immigration because new jobs are created by people who start companies, not by bureaucrats or professors. The best response is not to stop technology but to speed up adaptation through institutions, training, and work-supporting transfers.

Data Points: Conversation date: January 16th, 2014 - Opening introduction by Russ Roberts Consumer surplus from free Internet goods: about $300 billion per year - Brynjolfsson estimates the value of free goods online U.S. wealth: $77 trillion - Brynjolfsson cites record wealth as evidence the pie is growing Highway pavement unused in congestion: about 90% - Used to argue autonomous vehicles could greatly raise highway capacity Autonomous speed example: 55 mph - Brynjolfsson describes the Google car as driving smoothly and obeying the law Alternative historical benchmark: 1972 - Year when average and median income begin to diverge in his chart Employment share in agriculture in first machine age: 90% to less than 2% - Illustrates how old jobs disappeared and were replaced by new industries Human capital trend: The current graduating generation may be the first in U.S. history not more educated than their parents - Used to explain slowing educational progress Technology adoption example: Hundreds of thousands of people - Sebastian Thrun’s AI MOOC reached far more learners than a classroom Time horizon for some driverless applications: within a decade - Super cruise control and self-parking applications are expected to become common

Pivotal Quotes: "Technology is advancing incredibly fast. We are in a very innovative period, but there’s no economic law that says that everybody’s going to benefit evenly from improvements in technology." — Eric Brynjolfsson: Core thesis on innovation and unequal distribution of gains "We’re mindful optimists because we think if we take action and we adjust our skills, our organizations, even our economic institutions, and the way we organize work, we can adapt to technology." — Eric Brynjolfsson: His final stance on how society should respond "Work solves three great ills: boredom, vice, and need." — Eric Brynjolfsson: Why he prefers policy that preserves work rather than simply transferring income

Implications: The future may bring greater abundance, but without reforms in education, entrepreneurship, immigration, and labor policy, gains could concentrate at the top while many workers stagnate. The key challenge is adapting institutions fast enough to share technological progress.