Pretty sure your death numbers are off. 7-day moving averages are more meaningful as they take into account reporting lag over the weekend. The 7-day moving average for deaths on Feb. 16 last year was 2,234. This year it is 1,985 (so deaths are lower now than last year). The case load (taking the 7-day moving average) was 83,109 last year and is 129,470 this year. So, we are having fewer deaths in spite of higher caseloads. This has less to do with Omicron being "milder" and more to do with decreased severity of infection due to previous exposure to COVID via vaccination or a previous infection. Comparing deaths and cases on a single day are somewhat unreliable, as deaths are an indicator of what took place in the weeks preceding, not simply that day.
A better comparison between winter 2021 and 2022 is to compare cases on Jan. 15 (this is when they peak out most years because of increased social interactions around the holidays). 7-day moving average case on 01/15 were 237,602 in 2021 and 815,460 in 2022. Deaths peak out about 2-weeks later, so taking the 7-day moving average on 01/31, we get 3,273 in 2021 and 2,612 in 2022. So, we are getting fewer deaths from significantly more infections. Overall, though, the CFR dropped by a factor of 4.3. Omicron shot past vaccines, so a bunch of people got infected, but fewer died. Previous infection tends to outperform vaccination, so hopefully, next year the CFR will be even lower.