I just watched Joe Scott's piece: What's Happening (And Not Happening) With Hyperloop | Answers With Joe - YouTube. He gets some of it, but misses more.
He's right that a long vacuum is relatively hard to achieve, but he misses a lot of the point.
1: by making the walls of the tube thick enough, the tube can withstand a hard vacuum. His example of a big tank truck being crushed by pulling a vacuum inside is irrelevant. That tank was designed to take a slight positive pressure and has very thin walls. There's a simple equation to determine the needed the strength to resist atmospheric pressure. For a light bulb or small can, it doesn't need to be thick; it's more for a bigger tube. Elon's paper talked about tubes with 1inch thick walls. (Crazy extreme: Triton Submersibles' 36000/2 recently protected it's pilots from a pressure of 16000 PSI with a spherical structure only 90mm thick. This is over 1000 times the pressure difference a true vactrain would need to tolerate a hard vacuum at atmospheric pressure.)
2: this is a lot of metal, of course, but it's actually quite cheap. the rule of thumb is fabricated steel is 1-3$K per ton. A 5 foot tube one inch thick weighs about 640 lbs per foot. At $1K per ton, that's about $1.7M per mile. Not free by any means, but a lane of freeway costs over $10M per mile, and the california HSR project was several hundred million per mile.
3: for the same reason, most of the threats from terrorists are implausible. an inch of steel can withstand a shell from a pretty substantial cannon. A terrorist with access to such a thing will be going after much softer targets.
4: small vehicles do not limit capacity. Why is this so hard for people? Small vehicles somehow manage to carry many times more people than buses or trains on our highways. You need to have more of them, but they're cheap...especially on something like hyperloop where the vehicle is pretty much a passive object with life support and not much else.
5: a big part of the benefit of small vehicles is that they can be unscheduled. There will be a queue of them waiting at the station, which unlike air travel can be in the city core, or even several stations placed around the city. get in, tell it which station is your destination, and the computer does all the rest. The vehicles don't move when they're not needed.
6: the long vacuum won't be cheap and there surely will be leaks, but they're not too difficult to track down. as long as an individual leak isn't too big, it's just a small loss in efficiency, not a showstopper.
7: putting part or all of it in a tunnel is obvious. since it's the steel tube that's maintaining the vacuum, digging the tunnel is the same problem as ever. Burying it protects it from the elements to some extent, and provides convenient support, but it also makes it harder to work on. This is the reason the (much harder) vacuum at giant installations like CERN have access all around.
8: Musk's team did an analysis of I-5 between L.A. and San Francisco, and for most of the route, speeds north of 700 mph are easy without pressing the passengers too much. near the cities they'd have to slow down a bit because there are hills and waterways and buildings.. bottom line though is that their design would be able to make the nearly 500 mile trip in about half an hour. Dallas to Houston is half that and a MUCH flatter route.
9: There are several designs for a track switch that will work for hyperloop-speed trains. A magnetic version of the one Ed Anderson devised for Skyweb Express will work, and there are plenty of others that would too.
10: Unless there is significant damage for some reason, there is NO risk of derailment with any sort of vactrain. Conventional railroads use a relatively tiny flange to keep the wheels on the rails--it meant that the track switch could be relatively simple to make and manipulate in the 1830s. Debris on the rail, damage to the flange or rail, brakes getting stuck on, pushing a train through a swittch the wrong way, etc, will all cause derailment. Monorail, maglev, etc., do not have these issues.
