Bloomberg Podcasts · Situational Awareness' 439% Return Was the Warning

1. 1. John Kelly's 1956 discovery shows that beyond a certain risk limit, taking more risk only increases the probability of very bad outcomes, not good ones.
2. 2. Situational Awareness's 439% return indicates they were well over the Kelly limit, making a blow-up inevitable.
3. 3. Aaron Brown, an AI bull himself, argues that having a long-term vision for AI doesn't justify taking unlimited risk.
4. 4. Situational Awareness likely lacked a qualified risk manager, as their portfolio and public statements show no attention to risk.
5. 5. The $10 billion fund size is misleading because the Anthropic stake is still carried at $5 billion, likely overvalued, and most investors who entered near the peak are underwater.
6. 6. The private Anthropic stake is the safest part of the fund because it cannot be leveraged or used for daily margin.
7. 7. Banks like JPMorgan did proper due diligence and were able to exit before losses, unlike Archegos where they got burned.
8. 8. Citadel and AQR learned risk management through painful losses in 2008 and 2007, respectively, which is why they are successful now.
9. 9. Situational Awareness's complex positions, including long and short bets and puts, require asking whether they can survive the worst-case scenario.
10. 10. The Situational Awareness collapse is not a canary in the coal mine for the AI trade; it's just overleveraged players getting hurt.