Behe's book has been pretty thoroughly rebutted. In fact, all of his 'irreducibly complex' examples have been shown not to be irreducibly complex by his own definition. Ken Miller, on the plaintiff's list, has done a nice job on this.
Also, based on some things he's written recently, I suspect Behe may be wavering. In a recent interview, he referred to some of Lenski's work on in vitro speciation in microbes as a 'compelling' argument for standard evolution, if it turns out to be true. Lenski's result is embargoed prior to publication, but I understand it's big.
After reading the complete article, I would hardly characterize Behe as being anything other than sceptical of Lenski's work.