Employee Jacob Coxon resigned from Anthropic on Tuesday, posting a warning on X that AI labs are racing toward self-improving superintelligence and taking what he thinks are unacceptable risks with human lives.
Investors can't model existential risk, but they can model regulation, which is the real threat to AI infrastructure bets. Every dollar spent on the current AI build-out assumes rapid, uninterrupted growth.
None of this argues for selling anything. It argues for knowing what you own. Consider asking yourself two questions:
- How much of your portfolio relies on AI spending compounding for years?
- Would you comfortably hold those positions if AI spending merely continues rather than accelerates?
If both answers pass the test, stay the course. Long-term investing works because it doesn't require reacting to every headline, and the cheapest time to examine a concentration is while the market is still calm about it.