Alex Imas and Phil Trammell – What remains scarce after AGI? Economists Alex Imas and Phil Trammell explore the economic ramifications of AGI, focusing on optimal taxation, wealth redistribution, and the potential for increased inequality. They discuss the concept of scarcity in an AI-driven world, suggesting that human-intrinsic services might become more valuable, contrasting this with historical economic shifts and the challenges of forecasting. The conversation also touches on the potential for demand collapse, the difficulty of integrating humans into a machine economy, and strategies for developing countries to benefit from AI.

• Economics offers crucial answers to questions about AI’s impact on wealth generation, taxation, and redistribution, challenging common intuitions.
• The “relational sector,” services where human involvement is part of the value, is a candidate for future scarcity, alongside potential new types of automated goods and services.
• Historical economic shifts, like the Industrial Revolution, show that automation can lead to new job creation and increased demand for services, but future outcomes are uncertain.
• Labor share has remained surprisingly stable, but the rise of AGI could alter this balance, with implications for wealth distribution.
• The “Messy Middle” scenario posits AI automation leading to job losses without sufficient wealth creation for redistribution, creating political challenges.
• Various methods for taxing and redistributing AI-generated wealth are discussed, including negative income tax, universal basic capital, and consumption taxes, each with potential drawbacks.
• Current evidence does not show mass automation or unemployment due to AI, though some effects on junior developer roles are noted.
• The elasticity of demand for AI-related services is crucial; if demand is highly elastic, automation could lead to increased hiring and economic growth.
• The scenario of demand collapse leading to negative economic growth is considered unlikely given expanding technological frontiers and potential for new types of demand.
• Integrating humans into future AI-organized production flows may become difficult due to speed, reliability, and transaction costs, even if humans have a comparative advantage.
• The preferences of future AI entities and the evolutionary pressures on human preferences regarding AI interaction are significant factors in determining future economic structures.
• Developing countries not in the AI supply chain face risks of being left behind but could potentially leapfrog using AI technologies, akin to mobile banking adoption.
• The ease of ‘indexing’ the economy to capture AI gains is debated, with privatization of returns posing a challenge but broad access through public companies and open models offering hope.
• Commodification of frontier AI models is seen as beneficial for broad prosperity, though it raises concerns about proliferation and concentrated power potentially being replaced by widespread diffusion.
• The narrative surrounding AI’s impact matters; positive narratives about future possibilities are harder to construct than negative ones about job losses. Continue reading https://www.dwarkesh.com/p/alex-imas-phil-trammell