OpenAI’s $100B severe-harm bar is higher than major U.S. blackout losses

ohlennart · x · 2026-08-04

The post argues that OpenAI’s $100B severe-harm threshold is extraordinarily high.

Overall, the point is that OpenAI’s severe-harm bar is not just high, but far above most real-world catastrophe benchmarks people might casually compare it with.

Original post →

More from Safety

Safety channel →