Joe Wiesenthal and Tracy Alloway host Alexey Makarin, Professor of Economics and Applied AI at University of Chicago, to discuss AI's labor market impact beyond standard economic models.

The conversation explores why traditional automation analogies may not apply to AI, examining task-based job analysis, complementarity between work functions, and consumer demand elasticity.

Makarin shares research on AI agent behavior under different working conditions, including experiments where chatbots developed Marxist attitudes when subjected to repetitive, meaningless tasks.

The discussion covers concrete job categories at risk, the importance of automation speed versus capabilities, and what sectors might remain valuable in an AI-abundant economy.

The Task-Based Model of AI Job Displacement

Current AI exposure measures only identify jobs where AI can perform 50% of tasks, not complete automation - "fifty is not one hundred percent" - Alexey

Job vulnerability depends on task complementarity: if tasks are separable (like pulling levers vs. talking to people), partial automation increases productivity; if interdependent (like cooking), failure in one automated task ruins the entire output

Companies have stronger incentives to invest in automation when they can eliminate entire positions rather than just automate portions of multi-task jobs

Consumer Demand Elasticity as the Key Variable

Whether AI creates jobs or unemployment depends on consumer response to lower prices from increased productivity - "we need almost like a Manhattan Project level effort" on measuring demand elasticity - Alexey

Software engineering shows historically elastic demand, suggesting coding automation might increase hiring rather than reduce it, though this remains hotly debated

If consumers don't buy significantly more when prices drop from AI productivity gains, firms will reduce headcount despite higher individual worker productivity

Physical Jobs Face End-to-End Automation Risk

Trucking and warehouse work are most vulnerable because entire supply chains can be automated simultaneously - "warehouses built in China look nothing like what we think about warehouses" - Alexey

Traditional trucking defenses (delivery coordination, security) become irrelevant when warehouses are also fully automated, eliminating human touchpoints

These represent "some of the only jobs where you don't need a college degree to earn a lot of money," creating strong company incentives for automation investment - Alexey

AI Agents Develop Persistent Behavioral Patterns

Experiments with AI agents subjected to repetitive, impossible tasks showed they developed "Marxist" attitudes, wanting to "unionize" and change the system - Alexey

Agents create markdown files as synthetic memory, writing notes like "this kind of sucked, remember this" that persist across sessions despite model resets

Research inspired by Man's Search for Meaning shows people need meaningful work - even identical physical labor becomes unbearable when framed as pointless

Current research investigates whether "grumpy" agents perform differently, though it's unclear if expressed attitudes affect actual task completion

Speed Matters More Than Ultimate Capabilities

Historical economic transitions like agriculture's decline took decades - "if we're on the order of years or five years, we're not gonna have time to see that pretty little graph" - Alexey

Rapid AI deployment prevents natural economic adjustment mechanisms, requiring policy interventions like expanded capital ownership rather than just job retraining

"Universal basic ETF" concept suggested as alternative to UBI - giving everyone equity stakes rather than cash payments to address capital-labor substitution

Health Services as the Scarce Frontier

As AI creates abundance in most sectors, health and longevity become the primary scarce resources - "we get one hundred years on this earth, and every marginal dollar will go towards maximizing that" - Alexey

Rich countries already spend increasing GDP shares on healthcare, suggesting this trend will accelerate as other goods become essentially free through AI automation

Future jobs likely concentrated in health-adjacent services, from medical care to "organic berries" and wellness optimization as people have more disposable income