The comparisons of the simulated WFPC II and AC observations with the 8-m telescope observations show that, even with the assumption of very good conditions and an advanced state of adaptive optics technology on an 8-m class telescope, HST is still very competive for the observations of faint, extended objects. We did not use any image restoration for the simulated adaptive optics observations because the simulated cluster is too faint to benefit from such attempts. Because of the intrinsic FOV limitations of adaptive optics from the ground which are difficult to overcome by technological advances, the AC will have a unique capability of large FOV, high resolution deep imaging. Observations of high-redshift clusters are just one example of science programs which will be possible only with the HST, independent on the rate of progress in adaptive optics technology.