With the advances in computerized adaptive optics, surface telescopes (like Keck) are now outstripping the Hubble Space Telescope (HST) in resolution. See, e.g., Quintuplet stars.
The only atmospheric effect that the adaptive optics cannot cope with is the absorption of certain wavelengths. For those wavelengths, however, there are other space assets up there.
The HST can still deliver pretty good images, but are the cost of a servicing mission and further operating costs justified?