Alex Imas and Phil Trammell – What remains scarce after AGI?
Source: Dwarkesh Patel Podcast Guests: Alex Imas (Director of AGI Economics, Google DeepMind; Professor of Economics, University of Chicago), Phil Trammell (Head of Economics, Epoch; Research Scholar, Stanford) Date: June 4, 2026
Episode Topics
- Will capital share increase after AGI?
- The relational sector and what remains inherently human
- The "Messy Middle" scenario
- Taxation and redistribution of AI-generated wealth
- Demand collapse and labor market realities
- Human integration into the machine economy
- Intrinsic wealth accumulation by humans or AIs
- Advice for developing countries
Key Arguments
Will Capital Share Increase?
- Imas points to the "relational sector" where human involvement inherently adds value.
- Historically, economists like David Ricardo feared automation would cause mass unemployment, missing the eventual shift toward new service jobs.
- Trammell notes that while entire supply chains may eventually become fully automated, the overall impact on capital share remains ambiguous due to diminishing marginal utility.
- Historical analogy: predicting future scarcity is limited by our current imagination of goods.
- Imas stresses the urgent need for better data on consumer demand elasticities.
The "Messy Middle" Scenario
- A scenario where AI automates jobs without generating enough new wealth to easily support displaced workers.
- Trammell and Imas consider this a narrow possibility; typically, such advanced automation implies rapid economic growth.
- Imas warns that a gradual "drip scenario" of workers shifting to lower-paying roles could be politically worse than a sudden shock that prompts immediate government intervention.
Taxation and Redistribution
- Imas evaluates policies like universal basic income (UBI) and universal basic capital.
- Imas warns that direct cash transfers create dangerous political dependencies.
- Capital endowments struggle with targeting specific investments.
- Trammell recommends separating revenue generation (e.g., consumption taxes) from distribution mechanisms (e.g., purchasing and dispersing broad market shares).
- Negative income tax as an alternative: already provides a floor, but has political economy implications.
- UBI concern: "If people are just dependent on a check, it really matters who's in power. Right now, we're endowed with labor that can turn into income. When that is no longer the case and we are at the mercy of the elected official for basic needs, that feels like a power-sharing arrangement that's really dangerous."
Demand Collapse and Labor Market Realities
- Imas argues that demand collapse outcomes require highly improbable conditions, like hard bounds on demand and a total halt in investment.
- Current labor data shows little evidence of a "white-collar bloodbath," though junior developer hiring is slightly below trend.
- The elasticity of demand and the O-ring theory of jobs suggest that automating specific tasks can actually increase the value and demand for remaining human responsibilities.
Integration into the Machine Economy
- As AI systems become faster and more reliable, Trammell suggests that integrating humans into future production flows will become difficult due to quality mismatches and transaction costs.
- Regulatory frictions currently mandate human involvement in fields like law, but Trammell views these requirements as transitional.
Intrinsic Wealth Accumulation
- Future entities (immortal humans or autonomous AIs) might possess unsatiating preferences for capital accumulation.
- If certain actors continuously reinvest without consuming, they could eventually dominate global wealth.
- Historically, "dissipation shocks" like heirs squandering fortunes prevented this concentration, but life extension or non-human entities could alter this dynamic.
Advice for Developing Countries
- Lack of economic research regarding developing nations in the AI era.
- Trammell suggests high interest rates and falling prices for capital goods could naturally benefit those with even modest savings.
- Rather than focusing solely on retraining, developing countries should attempt to index into the global AI economy.
- Debate: Will AI resemble electricity (broadly distributed benefits) or social media (rents remain highly concentrated)? Open-source models and public market listings could help ensure wider prosperity.
Episode Timestamps
- (00:00:00) – Will capital share increase?
- (00:19:36) – Messy Middle scenario
- (00:25:57) – How to tax and redistribute AI wealth
- (00:30:02) – Why demand collapse is unlikely
- (00:39:26) – Human employees would be hard to integrate into the machine economy
- (00:43:08) – What if some humans (or AIs) value wealth accumulation intrinsically?
- (01:01:28) – What should developing countries do?