"If it bleeds, it leads"
Deaths is the only relevant measure.
You can also corroborate this by comparing it to new hospitalizations, which are also up.
“Deaths is the only relevant measure” - not sure where to start with this except to say that this is not at all what epidemiologists seem to think and I won’t address it further without some very dramatic reasoning and evidence.
No, it suggests that you’re doing more testing, and finding more cases. Which, exactly as the OP said, is a metric that can be manipulated by doing more testing.
The whole reason we emphasize positivity rate is to try to compensate for the inherent bias in reporting raw case counts.
There have been far more cases than we have ever formally detected with testing. There’s plenty of room to increase that number by testing more people, but it will not affect hospitalizations or deaths - which is exactly what we’re seeing.
Positive test rates measure what was happening about a week ago. Hospital admissions reflect activity a week or two before that, and deaths often get reported a month or more after the events that caused people to get infected.
You have to look at all the data to get an idea of what is happening.
https://www.worldometers.info/coronavirus/country/us/