trevor comments on Yann LeCun on AGI and AI Safety