> the cost of learning a seemingly arbitrary layout is probably less than the cost of using an inefficient layout, in the longrun It might be—but that’s no con

> the cost of learning a seemingly arbitrary layout is probably less than the cost of using an inefficient layout, in the longrun It might be—but that’s no consolation if you’re less likely to get to the long run. Or if you get to the long run, find out their models were off, and realize it was extra time wasted.