Hello, I'm a cinematography student from Argentina, and there seems to be some serious misconceptions here among many of my teachers as well as many industry professionals. This is pretty much what they say:
"Since h264 videos are made of GOPs, whenever I import an H264 video in premiere pro (or any editing software, for that matter), and trim/cut/crop it at some point in the timeline, if the cut was made at some point between 2 I frames of the original video, some of the information for interpreting the last frames may be lost, therefore resulting in a loss of quality on the export."
From what I understand, whether I cut the clip or not, Premiere is still reading the information to correctly "rebuild" each frame, from the original file. And when rendering (if exporting to H264), It won't use the same GOP structure as the original file (for example, an I frame in the original video may be compressed as a B-frame in the final export). When rendering an specific frame, premiere doesn't take in consideration only the particular data associated with that frame in the original video, but also the data present in the contiguous frames required to "rebuild" the whole frame. So when it renders out the video, it won't take that frame as a B-frame, but as a complete frame.
Your English is excellent and I'd fully agree with you. Of course, any time you transcode any compressed video to another compressed format, there is "some" loss to be expected - adding a generation - but what the instructor said did not make sense. If that were the case, we'd see issues just playing/previewing the footage in the timeline. Plus, H.264 is not the only video to use GOPs - think of HDV, XDCAM, etc. that also use GOPs with MPEG-2 for instance.