Compression methods have changed over the years, and can be better than they once were. Some of the late 1990s stuff was horrid (especially Red Hot Chili Peppers Californication – my definitive test track for awful vocals. That track and album was just maxed out with all peaks brutally sliced off. It’s super bright and in-your-face.
In my experience post-2010 compression involves squishing each waveform to bring the instruments and voice into harmony. This isn’t a bad thing, and not unlike careful mic placement for quality recording. The classic 1976 audiophile test album Jazz at the Pawnshop has a close mic on the chimes/bells. This results in defined music and sounds ‘great,’ but it’s NOT how one experiences a live show.
Compression continues because the average music customer listens on junky equipment, in the car with background noise, and plays music when they are multi-tasking. IMO they don’t want the kind of dynamics possible with Classical – too hard to hear and the changes are too disruptive. Properly compressed pop (vocals plus acoustic or electric piano/keyboard/guitar, drums, and misc.) doesn’t need a ton of dynamic range to sound decent, just not brutal treatment and clipping.