Your AI bet can be right and still run out of money. Grab the two-clock prompt: what has to be true, how long it really takes, who controls your runway, what pays today.
Aschenbrenner read AI right and still had to sell. Apple has looked slow for two years and it doesn't matter. The forecast was never the problem.
TL;DR
- Situational Awareness, an AI-focused hedge fund led by Leopold Aschenbrenner, sold a large part of its public stock portfolio to Citadel due to lender pressure after steep losses.
- The sale was a liquidity-driven event for the leveraged portfolio, not necessarily a verdict on Aschenbrenner's overall AI investment thesis, as private company positions were retained.
- The article contrasts Aschenbrenner's situation with Apple's, emphasizing that having time and resources (like Apple's cash engine and hardware) is crucial for long-term AI development, even if progress appears slow.
- Two critical 'clocks' for AI bets are identified: the technical clock (when capability arrives) and the financial clock (whether the investor can afford to wait for it).
- Apple's strategy of developing its own silicon, controlling the operating system, and leveraging its massive installed base provides it with multiple attempts and a long financial runway, funded by its current business.
- The article proposes five questions to evaluate AI bets, focusing on what needs to be true, the timeline, who controls the runway, what generates value currently, and what options improve over time.