Here's another of many MANY fascinating Mentour videos:
In all of these I've watched, I don't recall any (excepting Boeing 737 MAX MCAS) where autopilot, or other computer aids, failed. (There are often problems where the human pilots overlook what autopilot is doing.) There are problems with mechanical controls: on one flight the rear stabilizer control was jammed, and the stabilizer could no longer be trimmed. This made the airplane very difficult to handle.
On the flight linked above, the autothrust on one particular plane had been flakey for years. When switching from take-off thrust to climb-thrust, the system automatically reduces thrust but instead of pulling both thrust levers back a little, one jams so the other is reduced to idle to "compensate"! The thrust levers can still be set by human . . . if the human notices the problem. Every several months the system was inspected, rebuilt and/or lubricated ... but it would fail again after some months. In fact there was a failure the day before the accident flight but, with the on-going problem well documented, the pilots didn't bother to note the failure in the log.
Shortly after take-off from Bucharest airport, the tower calls the plane and advises to start the left turn earlier than planned. Captain (pilot monitoring) has to reprogram navigational computer. Distracted by this, he doesn't notice the autothrust problem: #2 engine remains at maximum thrust while #1 engine slowly moves to minimum thrust. With pilot flying banking sharply left per tower instruction, this asymmetric thrust increases the bank (and/or yaw) and becomes a big (and fatal) problem.
Shortly after setting "flaps up" the Captain makes a painful sound and does nothing else for the rest of the flight. This (heart attack?) might be considered the major cause of the crash, but in principle the remaining pilot should have been able to stabilize the aircraft. Instead he apparently never notices the asymmetric thrust. The workload during critical flight phases is huge even with TWO pilots but should it not have been almost instinctive for the pilot to reduce the bank angle and glance at the thrust levers?
As a layman, my opinion has no value but I'm happy to offer it anyway! There have been several Mentour videos in which wrong thrust lever settings caused a problem -- wrong settings that would be immediately and clearly seen if pilot glanced at the levers. Improper thrust is a major cause of flight instability.
There have also been previous videos in which programming the navigational computer distracted the pilot monitoring and led to problems (e.g. neglecting to glance at thrust levers). With so much navigational redundancy anyway, can't such reprogramming be deferred during critical (high-workload) flight phases?