I too found that particular set of frames quite interesting. So the last telemetry had stage 1 at 5700 km/sec and 22.3 km in altitude (so about 13 seconds from impact). The landing burn starts and at 24:55 "something" impacts the water really hard next to the barge. The announcer calls "landing burn has started" but you don't see any change in the telemetry (altitude or velocity) so its safe to say the downlink is still not functioning at that point). Then at 25:41 (about a minute later) we see the stage sitting on the drone ship. So once again I'm really surprised they haven't figured done something to stabilize video. Even it it were a 'drone dinghy' 500 yards off to the side holding a steady lock on a satellite or an aerial drone.

The other interesting bit is at 23:42 where we see one of the grid fins in the process of burning off the side of the rocket. Then we lose downlink video from that point on with the stage. It is flying at nearly 4,000 mph at an altitude of 87,000 feet. The SR-71 goes half that fast at that altitude and experiences significant heating on its skin, since drag is proportional to the square of the velocity you're taking 4x the heating.

It has to be a very fine line between burning these things up vs landing them. And clearly the more 'abuse' they will tolerate during re-entry the more fuel you can use toward lift mass. Still, I'm always surprised that they don't use more rocket power to slow down while above 100,000' to avoid the high speed re-entry and then use streamers or drogue chutes rather than RP1 to slow down in the thicker atmosphere.