Thanks for your good reply!

This seems like a rather strange book but I guess that's what also makes it kinda interesting. 

I actually played this early Qb6 idea myself for a while especially because the engines thinks it's quite decent at first sight and not many white players are aware of this idea but with accurate play from white I think black is in trouble. The move 9. Nf5 I think is the strongest but 9. Bb5 is also critical and should be investigated. 
The main line he gives up until 12...Kxg7 I totally agree with but 13. 0-0 is probably stronger than 13. Kf2 basically because the king can end up slightly more exposed on this square to Qxc2+ and such.
After 13. 0-0 Qc3 14. Qc1! turns out to be very strong and a lot of different threats are starting to emerge and right now and white will in general have very good play on the darksquares with black's important darksquared bishop already gone. White is immediately threatening Bh6+ which would completely box in black king and rook on h8 and another very typical idea is to play Rb1-b3 followed by Qb2+, when there might be Rxb7 tactics and immediate problems on the darksquares if white manages to get in Bh6 also white has more slower plans of f4-f5 while black has problems developing especially because b7 is always hanging. Dealing with all of this is very hard for black and it certainly was enough to scare me away from playing this variation. The engines usually sees this variation as equal for quite some time but the advantage rises for white as the game goes on. None of the games in my database are very impressive though and there's a lot of improvements to be made for both sides, for example two games in the database went as follows:
14... h5 15. Rb1 f6 17. Rb3 Qa5 18. Qb2 Qc7 19. f4 h4 now 20 f5! TN?! seems very natural and very strong which forces open the g-file instead of 20. h3 Rh5 which happened in both of the games in my database when it's harder for white to open lines with f4-f5 not being as easy to get in.