Administrating the test for it is both expensive, and time consuming, so the gross number of known infections is likely to be substantially underreported because of selection bias (why run expensive RNA analysis on a blood sample if the person is having mild cold symptoms?). There are probably a bunch of other confounding factors, such as most people getting it are probably asymptomatic or just think they have a typical cold/flu. It also appears to be non-lethal to children, as opposed to seasonal influenza.
In simple gross numbers, more people die in car accidents every day than the current worldwide totality for this virus. Comparing against US CDC stats, it isn't even remotely close to statistical estimates for flu deaths so far this year in the US alone(~16k).
My suspicion is if you concede that gross numbers of COVID19 infections are substantially underreported, the death rates fall into line with typical flu, maybe marginally worse.
The reasonable response to this is what people should already be doing for influenza: practice good hygiene, stay home if you're sick, take caution when interacting with at risk populations to minimize their exposure, etc.