With that assumption, he is assuming that since cases doubled every 5 days from March 12th until today April 4th, then deaths will proportionately double every 5 days beginning April 4th until April 27th (since deaths occur exactly 23 days after infection in his model). This however is clearly not the case when you examine where we are in the curve compared to what happened in other countries. If we follow similar curves to the other countries, we will peak sometime in the next week or so and then begin our downslope. Once we peak, if we were to maintain that deaths/day indefinitely, (which won't happen because it will decrease) our doubling time would begin at about 2 weeks and would get longer every day (since deaths/day would have to keep rising exponentially in order to maintain a constant doubling time). Given the healthdata.org model being used by the White House, our doubling time will be about 40 days by the time we reach April 27th that this guy is projecting, not the 5 that he is assuming.