psmith: <wwall>, <beatgiant>
Black's position is certainly difficult to defend after 8... Kxe5. I have tried analyzing this with Fritz 5.32 and it is just a lot more complicated than it looks. Here are some sample lines (I don't claim these to be definitive by any means, they just illustrate White's potential in the position).
8...Kxe5 9. d4+ Kxe4 (not forced, of course, other moves bear looking at) and now
(a) 10... g5 11. Bxf4! gxf4 12. Qh5 Bg5 (other moves also lead to mate according to Fritz) 13. Rfe1+ Kxd4 14. b4!! (cutting off c5 as an escape square) and White mates in at most six moves. (Fritz 5.32)
(b) 10... Kd5 11. c4+ Ke6 (11... Kxc4 12. Qb3+ Kxd4 13. Bxf4 leads to mate) 12. Bxf4 d6 (12...c6 13. Bc7! Qxc7 14. Qg4+ Kd6 15. Qf4#; 12... d5!? 13. Qh5!? Kd7 (again, not forced, but Fritz thinks Black is winning at this point) 14. Rae1 g5 (ditto to previous note; 14... Qg8 may be better but things seem unclear at best after 15. Qf5+ for ex: 15...Kc6 16. cxd5+ Kb6 17. Qc2 Na6 18. Rxe7 Qxd5 19. Re5 Qxa2 20. Re3 winning, or 15...Kd8 16. Qe5 Bd6 17. Qg5+ Kd7 18. Bxd6 cxd6 19. Re7+ Kc6 20. Rf8! winning again) 15. Rxe7+! Qxe7 16. Bxg5 Qe6 17. Rf7+ Kc6 18. Rf6 dxc4 19. d5+ Kxd5 20. Be3+ Kc6 21. Qc5+ Kd7 22. Rxe6 Kxe6 23. Qxc4+ with advantage to White according to Fritz.
Summary: White has a pretty big attack after 8... Kxe5 9. d4+ and taking the second knight may not be survivable. But as I say not much of the above is completely forced and a full analysis would seem to be extremely complicated. In addition perhaps Black should decline the second Knight after 9. d4+, although then White still has an attack after, say, 9... Kf5 10. Bxf4.