Michael St Jules 🔸 comments on Why does (any particular) AI safety work reduce s-risks more than it increases them?