The thirty-whatever thousand number is direct deaths that can be observed and counted up, I.e. a vast undercount of the actual number.
The Lancet determined a couple of months ago that you could at that point conservatively estimate about 186,000 dead, 7.9% of the population. Conditions haven’t been improving in the couple months since that happened and it was a conservative estimate anyway, so I said 10%. The truth is there’s a lot of uncertainty, it could be higher or lower than that, but saying it’s 1.5% is definitely wrong.