The distance to the earth might have been coincidentally the same as the distance to the other satellite. LEO isn't that far up, the one that took the photo was only ~600km AGL.

But then the resolution can't have been that high, or the other satellite would have been as big as the ISS. It must have been a lot closer than earth.

The range to the imaged satellite is overlaid on the images - it's between 90km and 130km over the sequence.

They've got 0.3m on World View 3, more than enough for what was shown in the gif

Edit: you're probably right, actually.. the silver square on the lower section of landsat 8 is ~1.5m high (guessing from the image on wikipedia), so there should only be ~5-6 pixels for it, but there are clearly more.

The sample distance (SSD) is noted on the GIF, and ranges between 6.5 and 4.6 cm for the images shown here.

Agreed. The focal range is something like 90KM to 1400KM based on this image and other off nadir images that have been shared by DigitalGlobe/Maxar, and it really makes you wonder what the focusing mechanism is fully capable of.