Agreed... and what else?

Yet another hint: Carefully consider the first and final frames.

It's subject to interpretation, but here's my read on the second pour: It's an uneven extraction, not much different than the first pour, but much prettier, especially in the second half. That is, the first exhibited obvious channeling from the back-left quadrant. The second pour was a classic donut extraction -- it beads early along the perimeter, the center unevenly begins to fill in, then eventually the surrounding "cone" envelops the middle, ending in what appears to be an orderly extraction.

That's why I look for evenness in the first beading and the following 10 seconds, not the luxurious striping that is frequently oooh'd and aaah'd in still photos of bottomless portafilter extractions. About midway into an extraction, most of the evidence of an uneven pour can be obscured, as was the case in the third frame of the second pour above. There's still possible telltale "puckering" evidence in the last frame, though it could be lighting. I look for indentations in the otherwise symmetric cone and distinct blond perimeter flow mixing with a darker inner flow.

Machines that don't appreciate updosing will complain by pulling a donut. It's guesswork on my part, but I assume the added headspace allows the water to more evenly diffuse; in extreme updosing cases, the grouphead can act as a convex tamper that compresses the center and leaves the edges untouched. Crashing the puck into the grouphead can also break the basket/puck adhesion and shear the quadrant opposite the lock-in engagement point (to check, lock in tightly and remove... is one quadrant thrashed and the rest nearly untouched?). Rough handling and/or tamper knocking can produce side channeling too, though I'm doubtful that's the case here.

Closing caveat: Visual inspection will only catch gross extraction errors. I've pulled visually gorgeous extractions that tasted average and "average looking" extractions that tasted exceptional.