Even assuming a virus has a 'goal', in human terms, it would be limited to maximal spread and very little else. Once a host has spread it enough (presumably by a dampened-down initial fatality rate, whilst still evading any antigens) there's no particular selective pressure to not ramp up to full Ebola/Rabies/necrification effects (just practical limits based upon what tricks its evolutionary envelope allows it to employ).
Sure, it can technically settle down to zero (obvious) symptoms, but allow an individual to shed itself upon virtually all daily contacts for the next 30 years without giving them cause to suspect they are all becoming Typhoid Marys. That would accomplish matters, but it'd be as unlikely as the Mr Hyde version to get zero fatalities as part of the mutation that accomplishes this. So you get people dying of rare over-aggressive manifestations of it (or complicatedly co-morbid mixing of illnesses), who would not have done so if we had still worked to not allow even a stealth-Covid the trivial person-jumping potential we know we can largely avert.
There are no easy answers, brethren, sistren and othren. So protective, preventative and reactive measures should be considered as vital (even if less effective, or less obviously effective) for Omicron until... Well, at least until we see Pi, Rho and Sigma (or see them applied to currently pre-VOIs, and maybe thennbecome VOCs) and see if they're also game-changey in a worse way. Or at least justify our better continuing practices continuing.