It is true that diffraction sets in early on such a high res system but it is not his problem here or any time. Diffraction will limit potential acuity but it will still be able to resolve more detail than anything else out there so he is not losing anything in the desirable balancing act of enough DOF to get the subject fully in-focus by stopping down and retaining the most potential resolving power.
This case here is the problem with cameras that act like computers, they do only what you instruct them to do, not what you think is should. The AF system did what it does, focuses on the best target that falls under the focusing point(which is actually quite a bit larger than the red box indicates in the VF). The AF locked onto the best target from the point of view of an AF system: foreground, distinct edges, linear lines in either vertical or horizontal planes(assuming the cross sensors in the center) and contrast. In some modes color is a major criteria. So looking at the scene from the cameras's point of view it was expecting a pat on the head and praise for picking and locking on the obvious target, the foreground branches. The bird was a poor subject because it was the opposite of the traits that AF are looking for. For a sure shot, let the camera AF do its thing by getting close and then switch to MF( of let go of the AF-On button) and fine tune focus manually for the bird. Either that or stop down the aperture so the bird would be well within the DOF centered on the plane of the branches, so getting 8 inches of acceptable focus, forward and backward 4 inches each side of the branches. If on a tripod, this is a good case for using LV. Stan St Petersburg Russia