Assuming that transmission does sometimes fail, the closure advocate must claim that S inevitably ends up with a warrant for Q by other means whenever it does fail. I call this claim “warrant preservation” (WP). The most prominent (and, perhaps, only) argument for WP is Crispin Wright’s: since S’s evidence E for P is expectable given either P or ~Q, S needs a prior warrant for Q in order for E to deliver a warrant for P. So, assuming that P is so warranted, closure is preserved. This is the “front-loading” strategy in defense of WP. Wright also claims that front-loading explains transmission failure. I first argue that front-loading implies basis infallibilism: we must have infallible bases for what we know if we know anything at all. But we don’t; the result is skepticism. I call this the “buck-passing argument”. I show that this argument applies to any view that implies WP. Safety theory might seem to offer a way out of this argument; but I show that this is ultimately more a liability than an advantage of that account. I then show that front-loading doesn’t explain transmission failure, and that NIFN, which does, provides no support for closure.