Sir William AAR plate “B” clearances (the structure gauge of most of the Northeast Corridor) are too tight to accommodate one hyperloop tube per existing railroad track (10'8" wide, 15'2" tall, per track, vs. a suggested 14' diameter hyperloop tube), thus negating any cost savings beyond using existing right-of-way since new structures must be built to carry the tubes over roads and rivers; plus the curves aren’t broad enough to sustain high speed. In effect, a 4-track railway with a maximum capacity of 25,000 passengers per hour per track would end up being replaced by one pair of hyperloop tubes (perhaps also a rush-hour direction center tube) with a capacity of 3300 passengers per hour per tube — a capacity reduction from 100,000 passengers/hour to 6600 passengers/hour (or 9900 passengers/hour with an express center tunnel) — most likely with no significant reduction in travel time.
(Furthermore, the existing North River tunnels would have to be widened to accommodate the hyperloop tubes, thus negating any potential cost savings.)
This is to say nothing of the hundreds of industrial customers that rely on local rail freight on the NEC, nor the heavy freight trains which use it on the off hours, or the numerous marshaling yards which can only be accessed through the NEC. The several drawbridges along the NEC would also be a significant engineering and capacity constraint on displacing trains with hyperloops directly.
(Again, my primary argument, though, is not that it is infeasible, but that its low capacity will make it impractical.)
