Steven Millhauser is correct to defend the short story as a form of "radical exclusion" that works through "austerity" but that can also through this very austerity "body forth the whole world." However, in making his case that the short story mostly settles for "a grain of sand" and leaves the rest of the observed world as the subject of novels, I think Millhauser is exaggerating the differences between the two forms, in a way that actually does an injustice to the novel.
According to Millhauser,
Large things tend to be unwieldy, clumsy, crude; smallness is the realm of elegance and grace. It’s also the realm of perfection. The novel is exhaustive by nature; but the world is inexhaustible; therefore the novel, that Faustian striver, can never attain its desire. The short story by contrast is inherently selective. By excluding almost everything, it can give perfect shape to what remains.
The most immediate overstatement here is in the association of novels with the "large" and the "exhaustive." This characterization clearly enough describes historically the practice of certain writers--Dickens, Dreiser--but not others--Hawthorne, Emily Bronte, certain novels--Moby-Dick, The Mill on the Floss--but not others--The Red Badge of Courage, The Trial. It also more accurately encompasses the Anglo-European novel than the American novel, which has always edged closer to what is generally called "romance" than to the "novel" and its inexhaustible realism. The romance, although not necessarily always "small," is nevertheless "selective," content, as Hawthorne put it, to "manage [its] atmospherical medium as to bring out or mellow the lights and deepen and enrich the shadows of the picture." Romance doesn't seek to "devour" the world but to transform a discrete portion of it into a version of the writer's own imagining.
This "tendency" in American fiction persists among contemporary writers, especially those commonly identified as "postmodern." Thus even meganovels such as Gravity's Rainbow or Infinite Jest, as "large" as they undeniably are, do not threaten to become "unwieldy, clumsy, crude." In both structure and style they bear the hallmarks of writers more interested in "hidden powers" than "things in plain view," intimate the possibility of "revelation," even if such revelation is perpetually deferred. Their "ponderous mass" belies an intensity of effect traceable in manner to Charles Brockden Brown, to Hawthorne and Melville, not to Richardson or Trollope, nor even Tolstoy. In American fiction, at least, the opposition between the hulking, indecorous novel and the delicate short story just doesn't very cogently apply.
A very good example of fiction to which this opposition decidedly does not apply can be found in the novels of Steven Millhauser himself. Edwin Mullhous, Portrait of a Romantic, and Martin Dressler require somewhat more room for their stories of obsession to be fully developed, but they are hardly recognizable as the sort of graceless beast Millhauser describes in his essay. They might be said to breathe a little more expansively, but they are otherwise as stylish and fully-shaped as any of Millhuaser's short stories (which themselves do have the kind of "completeness" Millhauser attributes to the short story.) Moreover, Millhauser has worked extensively in the novella, a form that at the very least straddles the divide between short story and novel, and as employed by Millhauser really only further undermines his own hard distinction between the two. Millhauser's novellas, collected in such books as Little Kingdoms and The King in the Tree, are just as elegant and selective as his stories (as anyone else's stories, for that matter), but they certainly do not shrink from assertions of "power," which in Millhauser's case results from the effort to encapsulate the world through fable and a twisted kind of allegory.
I think that ultimately all fiction involves a degree of "Faustian" striving, and that no fiction accomplishes "perfection." Fiction can never sufficiently "attain its desire" such that no further variations on a theme can be achieved, no additional aesthetic avenues of approach explored. And while it is possible to identify a strategy of "radical exclusion" that often does allow us to differentiate between story and novel, there is no reason why this strategy can't be practiced in those longer prose narratives we can't categorize as "short" stories and by tradition call novels. Millhauser's novels and novellas do this, as do, in different ways, the novels of Nicholson Baker, for example.If a novel has to conquer "territory" for it to be classified as a real novel, then I suppose Millhauser's taxonomy makes sense, but I don't see why this needs to be a defining feature of the novel in the first place.