Maybe dumb/simple question: every time I see #Covid19 model projections (maybe often IHME?) they have maximal uncertainty for the next few days and uncertainty diminishes over time.

Isn't this...backwards? Is it a reflection of uncertainty of the current number of cases/deaths?

I'm not paying much attention to modeling -- Jeff H points out that the point of IHME is to forecast *peak* hospital utilization and it's not meant to forecast the downslope accurately or capture the total number of hospitalizations/deaths