Same, I don't think it's very productive to just say I hate it, so please allow me to add some nuance.
6DOF is fundamentally flawed in the current implementation where appears to just move the camera relative to the sphere or whatever geometry is used to project the video on.
In a stereoscopic video, the camera is fixed relative to the scene being recorded (typically at least, some scenes have a moving camera which results in this 6DOF implementation to make even less sense). It captures a specific perspective, and the video content is designed to be viewed from that fixed point. When you introduce 6 degrees of freedom to move the virtual camera in the VR player, it creates a discrepancy between the intended perspective of the video and the perspective the viewer sees when moving their head. This resultr in disorientation and a disconnect between what the viewer expects to see based on their movements and what the video actually shows.
For me, and I expect this changes from person to person, this contributes to motion sickness and discomfort. The brain expects a certain visual stability when watching a video, and when that expectation is not met due to artificial camera movements, it can lead to discomfort and disorientation.
Unless something like view-dependent re-projection is used, this geometric warping and other discrepancies will continue to cause discomfort. Unfortunately this might not be feasible due to compute requirements and will never be able to exactly create a re-projection that is correct, but there might be enough information in the original video to estimate realistic looking small movements away from the original perspective.
Having said that, what is INFURIATING is that this option is not just on by default, but it turns itself back on once turned off. Now I am far from an expert in the field of User Experience design, but even I can see this is a good way to enrage users!
(state of the art depth estimation like this, can be useful in reconstructing the information required for re-projection)
