AI safety researcher David Krueger argues halting AI development is the only practical way to address concentrated power risks · Digg