I went through something like this when I read Jaynes's PT:LOS. The object of my idolatry was the Bayesian approach to data analysis. I don't think it would be accurate to say I ever had antipathy towards it, but I know now it's not the panacea I originally took it for. To a lesser extent, I idolized Jaynes. Again, I never hated him or his ideas, but I now have a better idea of where he went wrong.