gwern comments on Toward “timeless” continuous-time causal models