One of the classic anti-computer strategies of the past was to put a Knight on a3/a6 and wait for the computer to take it with a Bishop. Computer rated doubled pawns as slightly higher than the difference between Ns and Bs. So the human would get an open b-file and long diagonal for free.

So when Black played 8...Bxa3, that was similar to what the computers used to do. Hence white's comment.

Which website were you playing at?  I don't see that sort of comment on ICC.

Yes, and I think your light-squared bishop is pretty important in the Scandanavian, because his queen and light-squared bishop are developed and there's a threat of his knight. I wonder if this means Na6 and accepting doubled pawns is playable sometimes in the Scanadavian. 

This was a 3-minute blitz game, and I would lose to 1. Na3 a lot against players better than me. But I never would have guessed how fast and dangerous 1. ...e5 2. d4 exd4 and losing the dark-squared bishop can be after opening like that (Rybka doesn't object to any of those moves though). 10. ...Be6 is what hurt me a lot. After 15. ...Bd7 it was probably over unless I could get into a Fried-Liver type position.     

1.Na3 e5 2.d4 exd4 3.Qxd4 Nf6 4.Bg5 Be7 5.Bxf6 Bxf6 6.Qe4+ Be7 7.Nf3 O-O 8.O-O-O Bxa3 9.bxa3 d6 10.h4 Be6 11.Ng5 g6 12.f4 Bxa2 13.Qd4 Nc6 14.Qb2 Be6 15.e3 Bd7 16.h5 f6 17.Nxh7 Kxh7 18.hxg6+ Kxg6 19.Bd3+ Kg7 20.Qb5 Ne5 21.Rh7+ Kg8 22.Qd5+ Rf7 23.fxe5 Qe7 24.Bg6 Be6 25.Bxf7+ Bxf7 26.Qe4 Qxe5 27.Qh4 Qa1+ 28.Kd2 f5 29.Rh1 Qe5 30.Qg5+ Kf8 31.Rh8+ 1-0

After the game he made a totally nonsensical comment: "your computer is stupid".... I don't know what this means, maybe he was cheating himself. As evidence for this, wouldn't a non-exceptional human on 3 minutes be a lot more likely to play 11. Nxe6 fxe6 12. Qxe6+ ?