While the overall safety stats has improved, crew errors are becoming an increasing factor in the accidents that do occur.
The preponderance of the evidence, it's true, is that the modern, automated and multiply-redundant airplane breaks less often than its tab-and-cable ancestor. Also, since the 1950s, we've stopped expanding the commercial airplane's flight envelope and focused on making it more reliable and more efficient. We have not re-engineered the human being in the same way.
Nonetheless, there's no evidence that the rate of catastrophic "crew errors" (see below) relative to the number of flight cycles has increased. That suggests that the "magenta line" theory of declining skills may be flawed.
Moreover, what people call a "crew error" usually isn't. It's an indication that crew selection, training and maintenance of competency fell short of what was required to handle an off-nominal situation, and it's consequently related to the degree of abnormality that faced the mishap crew.
Now, I am not sure that even the FTFA-fundies here would argue that the unwarned-against activation of MCAS - a running-in-background gadget that commandeered the most powerful effector on the airplane - wasn't quite severely abnormal. And did it occur on a highly automated airplane? Negatory, sir. It afflicted a Topsy-developed hybrid of a simple 1960s servo-mechanical jet with a 21st-century digital overlay.