The problem with these studies is the mild symptom cases may never even be reported as cases, I know alot of people that had covid symptoms but never got tested, and would never get tested unless it got to the point where they needed medical care. So if you just had a "pretty bad" flu and sat at home you likely never got tested, and thus would never show up in these studies
Hell I know a couple people that were told by local medical center to not even bother with the test, they just assumed it was covid, was told to self isolate and monitor.
Both can be true - if you have a variant that is less deadly (but still deadly for some) and it is more transmissible then it will kill more people in absolute numbers. Also it can overwhelm hospitals, cuaing additional deaths and injuries (due to late treatment or treatment errors).
Delta is less deadly because more people have some immunity, either because of vaccination or because they have already been infected. Also it came late in the game, and many of the people who are the most at risk are already dead. Treatment has improved too.
So the virus itself isn't less deadly, it is just that we are better prepared.
the relative deadliness of delta... we'll just not ever really know. Though IIRC, some early surveillance with all the controls (only looking at unvaccinated, etc) suggested that it was LESS deadly for the elderly, more deadly for other age groups, net less deadly.
Virus mutations can be better for us, and they can be worse. This article https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7095397/ paints a sensibly nuanced picture.