Regarding diverting to the surface with a toping cone. I thought this is what they already tried when the methane at that depth froze up and clogged the pipe to the surface. The problem is that there is just not enough data on working at these depths. Also they drilled another 3 miles to reach the oil, and I would be surprised if all 3 miles was just muck.
Of course there's a casing which acts as a wall for the hole. The top kill idea is to work within the casing. Just piling stuff on top of the opening would presumably mangle the casing and create a much bigger mess.
My interpretation of the current video is that it's starting to work. On the left, you can see the backblow from pumping the "mud", on the right you can see the venting of the oil, which seems to be starting to diminish. I'm thinking that the "mud" started by forming a counterflow along one side of the casing, and began to descend. As it gets deeper it starts to pinch off the outflow. Let's hope.