The argument generally goes something like this, as articulated by a couple commenters on XGH's blog:
"We all proceed under assumptions that cannot be properly scientifically tested. You assume other people exist. I assume God exist. Neither of us is bothered by the fact that absolute truthiness of either of those propositions can ever really be known. We simply proceed with our lives, you and me both."
"The simple point is that there are some certainties in life that come prior to reason [such as the existence of the external world/other people]. Faith is one of them. Just because you can't prove it doesn't mean that it's not so."
The problem with the argument is that it's just a bad analogy. The real comparison would be, "You assume your direct empirical experience with other people is what it seems to be, rather than an elaborate hoax or dream. I assume God exists."
In other words, we have a great reason to assume the external world or other people exist--namely, our sensory experience. It's just that we can doubt these things with an extreme form of doubt, and we assume they are what they seem anyway. We're not making up an external world from a blank. Similarly we can have good reason to accept the results of a scientific experiment if it works--it's just that we can always doubt it based on the induction problem. In contrast, there's no empirical experience that provides a basis for faith in God.
To make this point clear: for the analogy to work, the believer would need to have direct empirical experience with or evidence for God, and then say, "Well, I could doubt this experience/evidence as the work of an evil demon, but I will assume it is real just like I assume the external world is real." This is not the case, unfortunately, which means that there is just no connection between the faith case and the other cases.
Without this connection, one would need some reason to put faith in the category of "things accepted even though they can be doubted." Otherwise, the argument basically goes, "You have what I am calling an unprovable belief [because it could theoretically be doubted], so I can believe anything." Literally. There's no discrimination between one belief and another--faith is automatically approved because something else can't be saved from Cartesian doubt, with no justification for putting faith in the same category. (Notice how they always say, "And faith is the same," without explaining why.) If that's true, rational discussion has been thrown out the window since all beliefs are fair game without justification, and it's pointless even to bother thinking about it or trying to justify faith (or anything else).