@badastro well, the Rayleigh criterion says that it should be bigger than 1.22 lambda / theta, where lambda is the wavelength of the light and theta is the angular resolution of the observed object (in this case a dinosaur of order of 10m)
This gives:
D = 1.22 * 0.5x10^-6 / inv sin(3.4x10^-14)
I can't get any inverse sin calculator to return a result for such a small value that wasn't zero, so I'm going to estimate it as being approximately 10^-14.
So my guess for the telescope diameter is anywhere between a million kilometers and infinity.
Edit: I read the article and I think they got the 66ly in metres wrong (they said it was 6.6x10^23m) -- but they use the angular resolution as 10^-21 radians which increases my calculation by 10^7.