This might be a good opportunity to make some distinctions. @kakape is a professional journalist, and his work product is the articles he publishes in Science. The Twitter feed is not the journalism, but is participating in a conversation. I find following Twitter feeds of journalists particularly useful, because they tend to have their ear to the ground about what is interesting, but of course it's no substitute for vetted, edited articles.
That particular tweet resonates poetically for me, and speaks to major themes in my essay. The belief that you can have rising cases and mortality will stay low is, as we speak, a major narrative in what I refer to as "idiot Twitter." What I get from the Tweet is that a basic theoretical understanding of the underlying processes should be enough to dispel the notion. Just as a child might not know that thunder follows lightning, the people espousing this theory do not understand basic epidemiology. It's not a scientific proof that thunder follows lightning, or that there might not be edge cases such as the flash being so far away that the sound dissipates.
Increased testing is absolutely a factor in explaining why we're seeing this puzzling pattern, but doesn't by itself explain what we're seeing. Part of the answer is lead time bias, which from what I can tell is a reasonably obscure epidemiological concept but relevant in this context. Dr. Ellie Murray explains it in this linked "tweetorial":
https://twitter.com/EpiEllie/status/1280305393516904448
Simpson's Paradox is another major factor: to really gain an understanding of what's going on, you need to look at individual states or even finer regional divisions, and aggregating them can obscure important patterns. The effect is especially important these days because different states are very different.