In order not to further irritate Marcus, let us move this off-topic discussion from the Rovelli thread. Whether the foliation is bent or not has nothing to do with my argument. By talking about a foliation, you are making a hidden assumption about classical observers. However, since the observer's trajectory is a quantum variable, it can only be specified if we leave the observer's momentum completely undetermined. Or vice versa. This means that the foliation itself is subject to quantum fluctuations, which is the essential physical novelty. In typical accelerator experiments, the detectors weigh several tonnes, and measurements are related to the detector's rest frame. The assumption that such a detector does not recoil at all is an excellent approximation, which is implicit in the assumption that the detector follows a sharp, classical trajectory. However, it is an approximation, which amounts to replacing several tonnes by infinity. When gravity is turned on, this becomes a problem, because an infinite mass will immediately collapse into a black hole. This is an essential difference between gravity and other interactions.