I would suggest that more & more future crashes/incidents are going to be caused by lack of understanding of avionics and glass panels
Understanding autopilot to be included in avionics, I very much agree. But, as stated so well in the "Children of the Magenta" video, it's layers of complexity. With the reality that when things get busy/stressful/task saturated, relying on avionics may be adding workload for the pilot, rather than reducing it as would be intuitive. And, when the avionics does no perform the intended function ('could be finger trouble), the workload may go much higher, or the opportunity for an un noticed error goes way up.
There have been times for me when things have started to cascade going wrong, and I have paused and wondered to myself, is this the moment of pilot high workload/distraction/task saturation that others will read about later in the report of my accident? If the answer is in any way a yes, then I need to shed workload etc. So far, it has worked. A number of times, I have turned off, or ignored the tech.
Older example for me; I'm VFR over the Baltic sea off the coast of Finland in a 182 amphibian, full Garmin glass, and steamgauge, all databases completely up to date, and a set of Jeppeson books as backup. I'm at 700 feet in the skuz, turning around to go 30 miles back to Kemi, my alternate. Kemi has just closed, and Oulu has accepted my request for special VFR, so we're going on in the skuz to Oulu, which is reporting marginal VFR in heavy rain. I'm at 700 feet over the water. I enter Oulu into the GTN 750 as my new destination, instant magenta line somewhat off to my left, and I set a 50 mile scale on the glass PFD for orientation. Every time I see a magenta line, I think of the Children of the magenta video, so I stop and think: Here I have a line to follow right to the airport VFR, but it will take me on shore... I'm at 700 feet, and I'm looking at windmills. I will not be crossing onshore toward windmills at 700 feet, and I'm not climbing into the clagg. The PFD is not displaying the obstacles at all, clear screen to Oulu. But I see windmills. I was not familiar enough with the GTN 750 to program in an instrument approach. But, I was familiar enough to have the owner, right seat, to open up the Jepp to Oulu, find me the localizer frequency, and fly over the water to intercept the localizer, upon which I was confident to not encounter windmills. intercept the glideslope and land, no problem. My lack of familiarity with the tech aboard: When you select the map scale from 25 to 50NM, it declutters obstacles (and does not tell you that it did). I learned that the next day when I read a few hundred pages into the Garmin FMS. My fault for not knowing, Garmin's fault for not indicating an automated declutter, nor suggesting a MOCA for the magenta line it painted. I reduced my automation dependence, and went to a more simple system I understood, which was less likely to induce an error.
More recent example, while training a new pilot in the new to him 182 amphibian (a different one), but again, all retrofit glass, I'm training him forced approaches. In the 182 amphibian, they're rather steep, It's a situation of get trained, and get used to it - they work fine, if you fly them right! However, all the way down: "DESCENT RATE PULL UP!" repeats. And I'm saying to him, "do not pull up, do not pull up!". The descent rate value that triggers this warning is probably appropriate for the landplane 182, but not a floatplane. In this case, while teaching, I had no choice but to tell him to ignore it, I could not find any cancel for it. When flying by myself in it, I could turn it off - with the avionics master switch!
I very much dislike having to train pilots to ignore, or not use the systems available to ease their flying duties, but sometimes it just ends up being the better way to handle things. Make everything more simple, back to what you understand well, and focus on doing what you know. There may be more actual doing, but what's happening is
only what you're doing!