@Volume
This is wrong as well. Please have a look at the available FDR data first. What triggered the trim movement was several short nose up inputs during a time during which the main inputs were left-right. The 4 minutes of stick back (a shorter, but still extremely long time even full back) were applied after the trim had already reached the full nose up stop, when basically all was done already, and the nose was even dropping below the horizon, which can somehow explain why full nose up inputs were given at that time. This does not mean, the systems brought down this plane, of course it was the pilot doing unbelievable errors in understanding the situation and steering the plane ignoring all procedures and hand flying basics. But he was not acting as stupid as it sometimes is stated in an enormously simplified version of the event.
For the time being I can see no link between both cases, except that it happened over water with severe thunderstorm activity in the area. But this time it was early dusk, not pitch dark night. I find it highly unlikely that similar attitude deviations remain unnoticed if you have some outside reference. I find it highly unlikely that a climb was not noticed, when such climb was requested but explicitly disapproved by ATC.
I recall that the trim works as an averaging function of the pilot inputs. PF on 447 was "mixing the mayonnaise" using the side stick with strong bias to nose up function. If he had mixed mayonnaise totally randomly then trim averaging function would have been zero, no trim.
I find it hard to fathom that it could be the same as AF447. However how can we ignore the coincidences: possible ITCZ CB penetration, possible climb towards the top of the altitude envelope in warm air, possible max trim. At the least I think you will agree there is a high probability of pilot overload as a common factor.