I have some thoughts on this. They don't really have a central theme but I love to think about this stuff:
- Pneumatic / maglev hybrid seems like the optimal choice. With magnetic levitation and no air resistance, extreme speeds are possible.
- Initially I thought that the ideal mass transit system would allow people to carry their cars with them. The idea is that people often have to choose between the benefits of mass transit and the loss of freedom that comes with not having their car once they reach their destination. But recent innovations lead me to think this won't be necessary. I can easily see a time when all cars are autonomous and you can call one up on demand (as others in this thread have pointed out as well). Therefore, it's easy to conceive of a time when you can hop on the HyperLoop, pull out your smartphone, pick the destination, and tell the system you want a (semi-)autonomous vehicle waiting for you when you arrive.
- To make this feasible, we HAVE to get the cost down. It shouldn't take a trillion dollars or more to build this network across the US. The rest of the points will be about that.
- The system should be built, as much as possible, out of a series of standardized components that "snap together". I'm envisioning a "series of tubes" sections that are straightforwardly snapped end to end with standard mounts to hang them or put them on pylons. There would be different lengths, curve radii, etc depending on how you needed to route the track.
- These "tubelets" themselves should be constructible on-site or close to it, similar to how concrete for roads is made on-site. The idea is to minimize shipping costs as much as possible.
- Ultimately, it would be nice if we could employ molecular nano-assembly and have the tubes build themselves, but as a compromise, many parts of the construction could be handled by autonomous or semi-autonomous moving construction rigs that can set themselves up, perform most of the construction work, then tear themselves down and drive off. Again, the more standardized and "Lego like" you can make the components, the easier this would be.
- Each section should have solar panels, a way of storing energy, and a way of giving or receiving energy to nearby sections. Ideally, this would all be connected with a "master bus" that would, when necessary, send power to sections of the loop that haven't had decent sunlight in a while, etc. In general, the idea would be to keep the system as self-sufficient as possible, but have full failover when needed.
- We'd also need to solve the cultural-political problems involved in getting all the landowners to agree to run elevated track over their property.
Thinking about the last two problems, an interesting solution presents itself:
- Why not partner with the power grid organizations and use the existing transmission line right-of-way to run much of the "long haul" routes? They already have the land, they have made wide "lanes" through brush and other unforgiving terrain to run the lines, and there's the power when you need it!
- You might even consider slowly replacing many of the transmission lines with just the tunnels. These tunnels would transmit the power and also help generate it with solar panels, taking only what the need to run the line and giving back everything else. This SHOULD be a massive net positive gain. The only downside I can see would be people that consider two sections of elevated tubes to be a massively worse eye-sore than the existing wires. But surely that would be solvable?