There's a very good explanation here by Dana Nuccitelli about the recent paper by Santer and others which identified the problem with some (modest) overestimates of atmospheric warming in modelling. (It's to do with errors in forcings estimates, not the models themselves. Climate change denialists instead will claim the models are all wrong and cannot be relied on to make any policy decisions.)
The thing is, the latter is the whole lukewarmer argument, isn't it? Because the models might not be precisely on point for a certain period, you can never rely on them to make policy decisions. It's a classic case of making the perfect the enemy of the good.
Incidentally, there was a Science Show recently that gave voice to certain key climate change "skeptics" (you can read the transcript here), but the one key impression you got listening to it was the age of the voices of the skeptics.