The Idea of high speed tunnels under cities and indeed crossing continents is really old. I am old. When I was a kid reading science fiction it was mentioned quite a few times. There was even a 60s era TV movie on that. So it aint new.
Mr Musk is a productive dreamer. He has ideas and manages to get people to lend him money to hire really good engineers to make them happen. If he hired me I would tell him every thing that was wrong with the idea. Then I would probably make it work as well as it could be done. As an engineer I have never failed to solve a tricky problem. This one would be very tricky indeed.
So he wants to bore a tunnel from his office to another part of LA. It will solve congestion by putting private cars on a kind of sled and zooming them at 120 mph along the tunnel. Actually this is right up one of my areas of expertise. In doing such things the limit is not how fast the sleds are, it is how fast you can load and unload the sleds at each end.
Numbers time. Assume it takes 30 seconds to get you to put your car on a sled and then place it in the tunnel to go. This is very similar to loading a roller coaster ride. It may take longer to unload it. Now in reality 30 seconds is pretty quick. There are staging tricks that could double that speed, maybe triple it, but if anything happens to one loading position it holds up everything.
At 120 MPH a sled would travel one mile. At 30 seconds dispatch that is two sleds in that mile. If you double the dispatch rate there are 4 in that mile. That is a good safe separation, but it is a very low flow rate. In an hour that is 240 vehicles. ( 60 sec in a minute / 15 sec dispatch * 60 min in an hour ) A 2 lane freeway moving at an average of 40 MPH with cars 5 seconds apart ( about 300 feet ) in two lanes is 1500 vehicles per hour. Bumper to bumper at 10 MPH is 2 seconds apart so 3600 vehicles per hour.
Conclusion is that such a tunnel in a single lane would have too little capacity to put a dent in a single 2 lane freeway. Hell normal city streets are quicker.
If you make the tunnel complex 4 lanes each way then the capacity ramps up to 960 vehicles per hour. That is a dent, but really not enough. In reality you may get it working with 6 lanes with four running one way with two back in the morning and the opposite in the evening. At minimum you need six tunnels and each one would probably share a service tunnel with another. That makes 9 bored tunnels for function, service ventilation and emergency. The resulting capacity is still one third of bumper to bumper on a two lane freeway. LA has freeways with 6 and 8 lanes.
There will be emergencies. Europe has many long tunnels and there have been horrible accidents. Death often from the inability of emergency services to get into the tunnel from either end fast enough. Being on automatic computer controlled sled is not a guarantee. The Channel Tunnel between the UK and France has had two major fires and all passengers and cars are carried on highly automated trains.
The issues are similar with the hyper loop thing. Ultra high speed in a vacuum tunnel is to compete with air travel. The US does not even have proper high speed surface trains. The cost of building surface infrastructure for high speed rail is significant but they have done so all over Europe such that short haul air travel is not needed and long haul within the continent by air is only a bit faster.
The hyper loop concept may be two or three times faster than a modern train, but if a surface train carries three times more people then it is just as fast from a people per hour perspective. A high speed surface train would be cheaper and have greater capacity.
A vehicle in a vacuum is dangerous. It is in reality a space capsule needing life support with an artificial atmosphere. We know how to do this, but to be fair astronauts are still a bit of the daredevil. A tunnel miles long is going to be difficult to maintain and to access for rescue. If the capsule leaks a load of dead people will arrive at the station as even with masks dropping down from the ceiling their blood will boil in a vacuum.
Actually forgetting the vacuum would be a good idea. You can still go pretty fast and it is much much safer. It is better than the surface as there will be no idiots racing the level crossing with trucks and buses.
To contribute to the travel solution the routes would have to be long, at least state spanning. The Western corridor from southern California to Washington is a good candidate. The similar Eastern corridor from Washington DC to Boston is better as it would serve 3 times the number of people. Although it does have a rail system that is good by 19th century standards.
You would still have to fly across the great plains because there are not that many people to support the infrastructure. And you would have to fly over the oceans. Tunneling those is simply impossible due to the mid ocean ridge volcanoes.
The most energy efficient crossing of oceans is by ship. Just saying.
As an engineer it is my job to predict how things could fail then design in prevention. But sometimes the list of what could go wrong is too long.
