Odd Lots James van Geelen on His Viral AI Doom Scenario
310 snips
Feb 28, 2026 
James van Geelen, founder of Citrini Research and author of a viral AI scenario, discusses a fast-moving AI capability curve and why he framed a severe 2028 disruption as a scenario to watch. He covers potential rapid white-collar displacement, vulnerabilities in private credit and insurers, agentic AI undermining platforms, and where AI might actually generate profitable returns.
AI Snips
Chapters
Transcript
Episode notes
AI Capability Curve Could Compress Transitions
- Rapid AI capability improvements could compress decades of labor-market change into a few years, creating intense transition risks.
- James van Geelen highlights an exponential capability curve since ChatGPT where tasks go from minutes to multi-day autonomy within years.
Model A Bear Case For Rapid AI Disruption
- Investors should explicitly model a bear case for rapid AI-driven disruption rather than only justifying past moves.
- Citrini wrote the scenario to force investors to consider non-obvious downside outcomes and monitoring metrics.
Falling Inference Costs Rapidly Change Economics
- Falling inference costs and algorithmic gains can make previously uneconomic tasks viable very quickly.
- James notes cost per cognitive task fell ~10–30x in a year, shifting economics for many jobs and services.