> general relativity has had considerable predictive success, so replacing it with something else that works the same in some situations but differently in othe

> general relativity has had considerable predictive success, so replacing it with something else that works the same in some situations but differently in others seems more complicated than [...] First: you're very correct - we shouldn't expect an easy solution. Second: precisely the same argument applies to Newtonian gravity. It was phenomenally successful. The mismatch between observations and its model were more easily explained by lack in observations - and they were! We found Neptune because of that! That role (of plugging holes in observations) is currently being played by dark matter. So chances for dark matter existing in some form are good from a historical perspective. On the other hand, from that same perspective, chances of gravity theory being slightly(1) updated for extreme cases are also good. (1) "slightly" in the sense that the difference with the current theory, GR, would be nil or negligible in most cases.