Carson Block Warns About AI, Talks ETFs and Credit Spreads

Summary

• Core thesis: AI poses a profound, underestimated systemic risk to the U.S. job market and economy, likely leading to significant job displacement within 3-5 years.
• In leading tech firms, current AI models already enable one worker to replace seven colleagues, and exponential model improvements will accelerate this displacement across the knowledge economy (e.g., law).
• Believes markets will price in this dislocation before actual job losses occur, potentially causing a crisis with speed and severity akin to the 2008 Global Financial Crisis but on a faster timeline.
• Believes passive investing (per Mike Green's theories) has warped markets, creating massive flow-driven multiples (e.g., $100 market cap increase per $1 inflow for NVDA) that will violently reverse during outflows, leading to a "falling knife" scenario with insufficient active buyers.
• Primary investment expression is through long-dated, out-of-the-money put spreads on high-yield (HYG) and investment-grade (LQD) corporate bond ETFs to bet on widening credit spreads.
• Rationale for using credit ETFs: volatility is relatively cheap, and a crisis could create a liquidity mismatch where ETF market makers step back, making it difficult to price/transact the underlying bonds, amplifying downside for those short the ETF.
• Advises against shorting the "Mag 7" or specific software companies directly due to timing uncertainty and risk management; prefers convex, capped-risk option structures.
• On major private AI/space companies (OpenAI, Anthropic, SpaceX): agnostic on IPO valuations; would not initiate long or short positions but might buy through a systematic momentum strategy if they enter the S&P 500 and score well.
• Views the Iran conflict as a market distraction; expects focus to return to the AI/job market thesis once geopolitical tension subsides.
• Key uncertainty: Timing of the market dislocation. Believes it is inevitable but the exact catalyst and horizon (e.g., 15% job losses in 3-5 years) are unpredictable.