Everyone Hates AI (EP. 453)

Episode Summary

Executive Summary: The episode centers on a viral “AI doom” research piece that imagines massive white-collar job loss, falling markets, and economic disruption by 2028. The hosts argue the scenario is compelling but overstates how quickly humans, businesses, regulators, and markets adapt. They debate AI’s labor impact, market reactions in software/financials/private credit, and whether the selloff is a buying opportunity.

Main Topics: AI doom scenario and market reaction (Priority: 5/5): The hosts unpack a viral research note forecasting 2028 unemployment of 10% and a 40% market decline from AI-driven white-collar displacement, and discuss why it briefly shook markets. Human behavior vs. technological determinism (Priority: 5/5): They push back on the idea that AI eliminates the need for human relationships, network effects, or service providers, arguing real-world behavior is messier than textbook supply-demand logic. Investment strategy amid AI-driven volatility (Priority: 5/5): Michael describes buying beaten-down names like Microsoft, Salesforce, ServiceNow, Blackstone, and CrowdStrike, framing the selloff as a possible overreaction and a potential buying opportunity. Labor market and productivity effects of AI (Priority: 4/5): They debate whether AI will destroy jobs or mainly change workflows and create more demand elsewhere, with emphasis on the transition pain for white-collar workers. Private credit and Blue Owl concerns (Priority: 4/5): A separate deep dive covers Blue Owl and the vulnerability of private credit/BDC structures, redemption pressure, and the market’s skepticism about management explanations. Broader market breadth, sentiment, and volatility (Priority: 4/5): The conversation notes strong global breadth, rising 1% up/down days, a risk-off rotation into staples, and the possibility that AI anxiety could alter risk appetite across markets. Pop culture and recommendations (Priority: 2/5): The episode closes with lighter discussion of movies and shows, including Terminator 2, Crazy Stupid Love, Industry, House of the Dragon, and other recommendations.

Key Arguments: The AI research piece is well-crafted and worth thinking through, but it is still scenario analysis rather than a prediction of inevitability. Human relationships, network effects, and service industries will not disappear just because AI can automate information gathering or workflow steps. Even if AI reduces some white-collar roles, firms and consumers will respond dynamically, creating more lawsuits, more demand for services, and more work in adjacent areas. The transition could be painful for displaced workers even if it does not produce 10% unemployment. Markets are overreacting to AI fears in names like American Express, Capital One, Salesforce, ServiceNow, and private-credit-related stocks. Buying the market or index-based strategies is safer than trying to time the exact winners and losers, but the hosts still buy individual names when they feel panic is excessive. Private credit issues look more like a confidence and liquidity problem than a broad credit crisis so far, but communication from some managers has been poor. Global equity breadth and resilient macro data argue against an immediate collapse, even if sentiment is fragile. AI may be deflationary, but a political and fiscal response could support assets like housing rather than destroy them. Most people are not paying close attention to AI yet; when adoption becomes embedded, the narrative may feel less dramatic than it does today.

Data Points: 2028 unemployment forecast in viral piece: 10% - The article debated on the show projected unemployment reaching 10% by 2028 due to AI-induced white-collar job loss. Market decline forecast in viral piece: 40% - The same scenario piece projected the market falling 40% by 2028. Current U.S. unemployment rate: 4.3% - Used by the hosts to argue the AI-disaster scenario has not yet materialized. Tech employment share of payrolls: 2.3% - Goldman-related chart cited to show tech is a small share of overall employment. Software publishing share of payrolls: 0.4% - Another labor-market data point used to argue software layoffs alone may not move the whole labor market. Ned Davis breadth statistic: 66% - Share of S&P constituents outperforming the index, cited as evidence of broad participation beneath headline volatility. Global markets near highs: Highest level since 2004 - A chart showed the number of overseas markets more than 2% above their one-year peak at a 20-year high. 1% up/down days pace: On pace to break the 2000s record - A chart showed the 2020s are tracking toward an unusually high number of 1% daily moves. 2000s 1% up/down days: 840 - Reference point for the most volatile decade due to the dot-com bubble and GFC. Mag 7 year-to-date drawdown: Down 21% - Used to illustrate how much pressure the largest tech names were under. Mag 7 drawdown from highs: Down 11% - The hosts discussed the psychological impact of the group’s decline on the broader market. Microsoft drawdown from highs: Down 30%+ - Michael cited Microsoft as one of the names he was buying into weakness. Blue Owl stock decline: Down 60% - Example of how severely private-credit-related stocks were punished. Blackstone fund inflows: $600 million in January vs. $1.1 billion in November - Wall Street Journal data cited to show inflows were slowing but still positive. OpenAI projected burn through 2030: $218 billion - Used in a comparison to historical cash burn from Netflix, Tesla, and Uber. Netflix cumulative free cash flow burn: $11 billion - Historical example showing how much cash a winner may burn before succeeding. Tesla cumulative free cash flow burn: $9 billion - Another example of high upfront burn before business success. Uber cumulative free cash flow burn: $18 billion - Used to compare with OpenAI’s projected capital needs. Average movie runtime in 1980s/1990s: About 100 minutes - The hosts argued films used to be shorter and often should be again. Current average movie runtime: 115-120 minutes - Used to argue movies have become too long.

Pivotal Quotes: "We had overestimated the value of human relationships." — Transcript speaker / cited from article: The hosts highlighted this line as the most objectionable and unrealistic claim in the AI doom piece. "Turns out that a lot of what people called relationships was simply friction with a friendly face." — Transcript speaker / cited from article: Presented as the article’s harshest thesis and rejected by the hosts as misunderstanding how the real world works. "If you can eliminate $150,000 jobs and replace them with $20,000 line item expenses, like, are we sure we want to be doing that as business leaders?" — Jamie Dimon: Referenced in the discussion of the social and political backlash that could follow aggressive AI adoption.

Implications: Listeners should expect more AI-driven volatility, especially in software, fintech, and private credit, but not assume linear doom. The bigger near-term risk may be sentiment shock, policy backlash, and a painful labor transition rather than instant economic collapse.