Interesting. I stand corrected, though I'm now curious how it divides up tasks in video decoding.
I did a bit of experimentation on my MBP (2.16 Core Duo, so pretty old) at home, and found that on your video 720p resulted in Safari using about 160% CPU and it playing more or less smoothly. 1080p had both cores pretty much pegged (and it played very badly).
Not sure why it's not maxing both cores on your computer with 1080p, but regardless it wouldn't likely play properly anyway.
Also, ARF900 is right that with the resolution on your screen 1080p shouldn't make any difference--you have a 1280x800 pixel screen, and widescreen 720p is 1280x720, so apart from the black bars it's exactly the right size for your monitor at fullscreen. It is, however, going to be noticeably better than 480p, particularly with how close you are to the screen.
Now, while the resolution of 1080p does you zilch good, were your computer capable of playing it, in reality the 1080p YouTube video might look a bit better than 720p because the higher bitrate is essentially resulting in less compression artifacts when downscaled to 720p. Even so, it's almost certainly not worth worrying about--just stick to 720p.