Will AI solve the productivity puzzle? With Nick Bloom

Episode Summary

Executive Summary: Samaya Keynes speaks with Nick Bloom about whether productivity growth is turning a corner after decades of slowdown. Bloom explains his shift from pessimism to cautious optimism: AI may boost productivity, but work-from-home could also be contributing through higher labor supply, better focus, and reduced commuting. He also warns that benefits are uneven, concentrated in a few sectors and geographies, and may face backlash if gains aren’t widely shared.

Main Topics: Long-run productivity slowdown and Bloom’s changing outlook (Priority: 5/5): Bloom revisits his earlier pessimism about the declining trend in productivity growth since the mid-20th century and explains why recent evidence has made him more optimistic. AI as a possible inflection point (Priority: 5/5): The discussion centers on whether AI could reverse weak productivity growth, with Bloom citing CFO surveys showing firms expect meaningful productivity gains over the next three years. Diffusion, concentration, and inequality of technological gains (Priority: 4/5): The episode explores how new technologies spread slowly across firms, occupations, and regions, with early gains concentrated among high-skilled workers and leading firms. Economics profession’s response to AI (Priority: 3/5): Bloom describes how AI is reshaping research workflows, accelerating analysis, changing publication incentives, and raising concerns about slop and coding errors. Work from home as a productivity and labor-supply force (Priority: 5/5): Bloom argues that remote/hybrid work may be boosting output through quieter work environments and less commuting, but especially through higher labor supply. Organizational design, monitoring, and office attendance (Priority: 4/5): The conversation covers why employers and employees disagree on office attendance, and why some firms still require in-person work despite evidence supporting hybrid models. Unexpected social effects: fertility and family life (Priority: 3/5): Bloom highlights a possible fertility boost from work-from-home arrangements, arguing that flexibility makes it easier to manage childcare and family decisions.

Key Arguments: Productivity growth has slowed for decades, but recent data and surveys suggest the direction may be changing. The 2020 paper on ideas becoming harder to find found innovation outputs worsening relative to inputs, though later work suggests diffusion constraints may also explain the slowdown. CFO surveys are a strong guide to near-term AI effects because executives have good visibility into implementation bottlenecks and business realities. AI’s gains are highly concentrated: many firms report no current effect, while a smaller share of adopters report large spending and expected benefits. Technology adoption usually begins in a few places and occupations, then diffuses over 10-20 years, so initial gains are uneven and politically sensitive. Work from home can raise productivity by reducing commuting and improving concentration, but its larger macro effect may be increased labor supply rather than higher output per hour. Hybrid work appears to outperform full-time remote or full-time in-office for many jobs, with some office presence improving learning, connection, and retention. AI is affecting teaching and hiring by making take-home work and remote interviews less reliable, forcing more in-person assessment. Work-from-home flexibility may also increase fertility by making childcare and family logistics easier, especially for dual-earner households.

Data Points: US productivity growth in the 1950s: about 3% - Bloom cites long-run US productivity growth as very high in the postwar era. US productivity growth in the 1980s: about 2% - He describes a continued long-term slowdown by the 1980s. US productivity growth pre-pandemic: closer to 1% - Bloom says growth had fallen further before COVID-19. Bloom’s optimism about productivity a couple of years ago: 2 or 3 out of 10 - His self-assessed optimism before recent AI and work-from-home evidence. Bloom’s optimism today: 6 or 7 out of 10 - He says his outlook has improved substantially over the last 24 months. AI-related expected boost to productivity growth: about 1% a year - Average forecast from surveyed executives looking forward to 2029. Time horizon in CFO survey: next 3 years - Executives were asked about AI effects over a near-term horizon. Share of firms reporting positive AI impact: about half - Bloom says roughly half report positive numbers and half report zero. Median AI spending per person in US firms: close to zero - Most firms spend very little despite some very high-spending adopters. Average AI spending per person in US firms: a couple of thousand dollars - Mean spending is pulled up by a small number of heavy users. Office worker time saved by work from home: about 1.5 hours per day - Bloom says commuting time saved is substantial on WFH days. Estimated number of people with disabilities enabled to work by WFH in the US: about 500,000 - Bloom cites a paper on labor-force participation gains from remote work. Productivity increase from one day a month in office: 8% - Result from Bloom’s randomized trial in a company that was initially fully remote. Quit-rate reduction from one day a month in office: one-third lower - The same RCT found improved retention. Fortune 500 hybrid share: about 75% - Bloom says most large firms allow hybrid work. Fully remote Fortune 500 share: about 7% - A smaller subset is fully remote. Estimated fertility increase from WFH: 5-10% - Bloom’s best guess for the effect of allowing a couple of WFH days per week. Discussion time at NBER Summer Institute devoted to AI: about 25% - Bloom says AI dominated conversation among economists.

Pivotal Quotes: "when you start thinking about growth, it's hard to think about anything else. And now I'd replace the word growth with AI." — Nick Bloom: He describes how AI has become the dominant topic in economics and research. "The thing that's surprising for me is take the current US government. It's on this bender to ban work from home." — Nick Bloom: He contrasts anti-WFH policy with government goals to raise fertility and growth. "you can't get pregnant by email." — Nick Bloom's friend, cited by Bloom: A shorthand explanation for why remote work may support fertility by making family life easier.

Implications: AI and hybrid work may both lift growth, but the benefits are uneven. Firms, workers, and policymakers should expect concentration in leading sectors first, while broader diffusion, management practices, and labor-market rules will determine how much of the upside becomes economy-wide.