Hosts debate the Catrini report's SaaS apocalypse thesis and what the Stripe data actually says about AI's economic impact
Feb 24, 2026
Key Points
- The Catrini report's SaaS apocalypse scenario went viral and landed on the Wall Street Journal's cover, but market reaction showed skepticism rather than panic, with most indices staying green.
- The real problem is that serious AI impact discussions have become literary narratives rather than analytical, driven by genuine uncertainty about macroeconomic effects and poor real-time data availability.
- The timeline matters more than the doomsday binary: major AI capabilities historically take years to mature, pushing most transformative economic outcomes beyond the 2028 frame that consensus doom narratives imply.
Summary
The Catrini report's SaaS apocalypse thesis went viral and landed on the Wall Street Journal's cover as a doom narrative. The report models a scenario where AI commoditizes software and collapses the industry. Market reaction has been muted. A few names dipped a few percent, many have already recovered, and broad indices are green.
Catrini positioned the scenario as one of several possibilities with low probability, but spent 100 hours modeling only that outcome. The signal contradicts the disclaimer. If you spend that much effort on a single scenario, you suggest it deserves the most consideration regardless of stated probability.
The real problem is not whether the scenario is plausible, but that serious AI conversations have become literary instead of analytical. Genuine uncertainty exists about AI's macroeconomic effects. The quality and supply of real-time data on that impact is poor. The vacuum is filling with competing narratives: boosterism, fear, and storytelling. Nobody actually knows what will happen in the next year or two. Frontier labs cannot yet fully describe what they are building. Economists lack models for economy-wide effects of a technology whose properties remain uncertain.
Timeline matters more than binary outcomes. Internet predictions failed across the board, but many came true over 20 years. The Catrini frame is 2028, which is radically different from a 20-year adjustment window. Hollywood workers could absorb digital filmmaking, CGI, and AI tooling over two decades. The claim that one-shot AI films eliminate employment prospects in two years is a different threshold entirely.
Major AI advances have taken years. GPT-3 to usable language models required several years. Early Dall-E to coherent image generation took time. The gap from mostly sloppy to dialed to 99.9999% perfect has real time cost. That schedule pushes most transformative outcomes further out than consensus doom narratives suggest.