The investigation of electricity at the atomic scale requires atomically sharp electrical contacts. These atomic point contacts (APCs) provide the exquisite spatial resolution in scanning tunneling microscopes and are the "alligator clips" used to electrically probe individual molecules. APCs have also been predicted to be ideal quantum-limited displacement detectors, sensing the distance between the point and another nearby object at the limit imposed by the Heisenberg uncertainty principle. The primary difficulty with an APC displacement detector specifically, and with APC measurements in general, has been that they are too slow to resolve dynamics on the nanosecond time scales relevant to many interesting systems. In this talk, I will describe how we increase 1000-fold the measurement speed typically achieved with an APC. We evaluate the speed and quantum ideality of an APC-based displacement detector by sensing, with nanosecond temporal resolution, the femtometer-scale random thermal motion of a nanomechanical beam resonator. By resolving this random motion, we find both the displacement sensitivity and the random momentum backaction of the APC. We are able to sense the motion of the beam with a sensitivity about 30 times worse than the standard quantum limit and find a sensitivity-backaction product about 900 times larger than the required by the Heisenberg uncertainty principle. The non-ideal behavior arises not from electrical noise added by subsequent amplification but from a source of excess backaction intrinsic to the APC itself. I will discuss a possible origin of this excess backaction and prospects circumventing it in the future.