They measure different things, so it depends on what you're trying to understand.
If we tested everyone in the country/state all on the same day, and had the results back immediately, the case count would go way up but the positivity would go way down. It would be both encouraging and likely disappointing at the same time.
I like to think of the case count as how much spread we know about. The raw impact of the virus that we know about, and a likely indicator of the deaths and impacts to follow. While the positivity tells us if what we know about is everything or just a partial picture. When the positivity was high, those days over 10%, we had no idea where spread was occurring. The case counts on those days were incomplete and missing information. We didn't know the entire picture or where the virus was spreading. As we identify contacts and exposures and test them, we'll drive the positivity down. We'll begin to know where the virus is, who it's spreading between, and we can effectively isolate infectious people until they're no longer infectious. If we could identify and test fast enough to drive the positivity low enough, we might know almost everyone who is infectious.
That's where watching the positivity is useful, the lower the number, the closer we are to getting community spread under control. If we get it under enough control, we'll be able to reduce cases and the case count will fall too then.