There are some rebuttals by Peter Bender (his abstract for the 2013 April APS meeting may point in useful directions). For the space-based experiment, barring c

There are some rebuttals by Peter Bender (his abstract for the 2013 April APS meeting may point in useful directions). For the space-based experiment, barring concerns about the technique itself, the right question may be, "Even if the atom interferometer might work, is it a better bet than the existing plans for LISA?" I haven't studied the atom interferometer GW stuff in great detail, but the measurements look challenging from a feasibility and systematic perspective. Any one objection can be addressed, but there are a lot of them, making it challenging in the aggregate. Continued work on the ground in the atom interferometer field over longer baselines will explore both. If that work can show that it works as proposed, then pessimism will turn to optimism; experiment is the arbiter of truth. As for the second point, regarding clocks, my initial response of "no" came from exactly your argument. I hesitated when considering a global array of clocks. For wavelengths comparable to the size of Earth and smaller, there's enough phase difference that perhaps a sufficiently precise clock, read out sufficiently fast, might pick off a signal. Additionally, if there are signals in the micro-to-nanohertz regimes, it might appear as an unexplained fit residual to an overall model of the gravitational potential at any one clock. My GR-fu is insufficient to make reliable estimates of the sizes of these effects (I bet they're small!).