> They appear in the Kerr metric Only inside the inner horizon, which virtually all physicists working in the field consider to be unphysical because of the "i

> They appear in the Kerr metric Only inside the inner horizon, which virtually all physicists working in the field consider to be unphysical because of the "infinite blueshift" effect at the inner horizon and the fact that it is a Cauchy horizon: you actually can't predict what will be inside it based on any data from outside it. (Note that the outer horizon, like the horizon of a Schwarzschild black hole, does not have these properties.) The prevailing opinion seems to be that the interior of a rotating black hole will end up being qualitatively similar to the interior of a non-rotating black hole, with the inner horizon and region inside it of the idealized Kerr solution being replaced by something that looks more like a Schwarzschild interior. (And of course if we figure out quantum gravity that might change things even further.) > We certainly use the Kerr black hole solution to model things in the actual universe. We use its exterior to model things in the actual universe, yes. But none of those models depend on the interior having the properties that the idealized Kerr solution has. They certainly don't depend on their being CTCs somewhere inside. > the fact that CTCs appear in solutions to GR but seem physically implausible suggest that the basic physics of time are not fully understood No, it suggests that, as with any physical theory, you have to look at actual data to see which of the infinite number of mathematically possible solutions to the theory's equations describe things that actually exist. All of the solutions that describe things that actually exist are globally hyperbolic and do not contain CTCs or any other weird "time" properties. That suggests that, as far as things that actually exist go, we have pretty good models of "the basic physics of time".