NORAD and the LHC are examples of successful underground projects that have addressed the seismic issues.
I agree with you that to reach the level of safety consumers will demand, more thought needs to go into it. But it's not impossible, which was kind of in line with what the article was saying.
Note that Japan's Chuo Shinkansen is to be approximately 60% underground (much of that "deep underground").
... and it's real-world, actually-being-built, long-distance (about 300km for the first phase), safety-critical (transports passengers at high-speed), market-financed project in a highly seismically active area.
That's only about half the speed that the hyperloop is being advertised at, though. That route is claiming ~500km/h, which would put you from SF to LA in about an hour, if this site is accurate: http://www.mapcrow.info/Distance_between_Los_Angeles_US_and_...
Which is not to imply that's not an incredible feat - the Shinkansens are awesome, and work. Just that things get harder practically exponentially as you increase speed, so doubling it is a very, very big jump.
Sure, I was just responding to the subthread about the feability of long-distance precise/safety-critical tunnel-construction in seismically active areas.
Nobody knows what hyperloop is anyway (and it's not even clear that Musk actually has a real plan), so it's hard to say much about the details! :]
Are there earthquakes around NORAD (it has long underground tubes?) and the LHC? I really have no idea, and Google is failing me (and it's slow going, very bad internet connection at the moment). And I feel I should point out that the LHC doesn't cross major fault-lines like any trans-USA transportation system would have to do (or even SF to LA, which is explicitly used as an example, and presumably that's much easier than LA to NY (which also crosses the San Andreas fault line)).
Anyway, I'm questioning underground tubes in places where the earth moves at "about 33 to 37 millimeters (1.3 to 1.5 in) a year across California."[1]. While that's not much per year, it adds up, and presumably such a large project would be built to last more than a couple decades (or plan for fairly routine major construction work to keep things straight).
I don't have a structural engineering background so I'm having a tough time imagining how someone could make a train-track or hardened tunnel "flexible" but I'm sure there are many ways.
It doesn't seem too hard to predict the movement of the plates and build in automatic adjustments for the track/tunnel. Have a way to automatically extend track/tunnel and move it laterally to adjust for any minor plate movement.
It seems like it would be cheaper and just as effective to build higher speed rail beside of existing rail lines. Or maybe some sort of independently powered train cars that are private so that there are no unnecessary stops, more automated, and can travel faster. Train tracks likely already have to compensate for the faults somehow, though their probably slightly simpler.
Regenerative breaking would be nifty technology to build into the trains/cars/vehicles and musk has teams with significant experience in these fields.
I imagine that, with all the of top-tier engineers that Musk has on his payroll at tesla, spacex, and solarcity, he can get plenty of suggestions for how to overcome or prevent several feet of movement in the track, humans have done more complicated stuff than that. Then make sure to make redundancy and backup of backups to prevent loss of life and help prevent gridlock in the event of a major disaster (vehicles could go backwards maybe?).
I agree with you that to reach the level of safety consumers will demand, more thought needs to go into it. But it's not impossible, which was kind of in line with what the article was saying.