As a matter of fact, he's one of the premier experts on cosmology in the world (and a colleague of mine here at Penn). When it comes to testing cosmological models against hard observational data, there's none better.
The Big Bang theory stipulates that at the beginning the volume of the universe was zero. After that, it expanded (extremely rapidly) and continues to do so, such that the total volume of the universe grew and continues to grow at a high but finite rate.
That refers to our horizon, but not to the whole. In the standard Friedmann cosmology, the volume of the universe can be represented as the "surface" of a 4-dimensional hypersphere. We have a finite horizon, or "Hubble volume"--the volume of the universe that is receding from us at a velocity less than the velocity of light--but the whole is (in principle) finite, too: it eventually curves back upon itself.
[Geek alert: read carefully what I said. In any cosmological model, there will be parts of the universe that are receding faster than light. This does not violate special relativity, because these regions are out of causal contact with us.]
The question remains: what is the radius of that hypersphere? It could be anything, according to Friedmann; it's just something you have to measure. On Earth, for example, you can measure the radius of the Earth simply by noting that an equilateral triangle, 10 million meters on each side, will have interior angles of 90 degrees each, rather than the 60 degrees of a small equilateral triangle.
We now (as of 4 months ago) have a similar measure for the largest possible triangles that can be drawn in our Hubble volume. No matter how big, the interior angles sum to 180 degrees. This corresponds to a hypersphere of infinite radius; in other words, the universe is flat.
This agrees with the theoretical prediction of Inflationary Cosmology.
[Geek alert: the curvature could also have been negative, in which case the universe would have a hyperbolic shape, rather than a hyperspherical shape.]