Simplistic idea beyond belief.

Nobody has explained yet how they turn the sleds around for a return journey.

You either run a double decked tunnel (BTW larger diameter tunnel required than the car widths being quoted) or two tunnels with a loop or runoff tunnel arrangement with rail switching between. A fan exit arrangement multiplies the complexity by the number of exits, so is not really practical.

Alternatively run each one way route in isolation with a double deck tunnel, with the lower deck used for empty returning sleds, which don't need as much headroom unloaded.

Out on a tangent, I suppose you could try a rectangular or square tunnel boring shield to reduce tunneling volume for shallow tunnels, this is something currently being trialed in Singapore.

Absolutely correct.

As someone who worked on the Channel Tunnel at the dirty end, I can confirm 'simultaneous' lining and boring has been carried out for the last few decades.

However, there will always be an cyclic element to the operation as the Tunnel Boring Machine (TBM) has to push off of something to get thrust forward and that is usually done by hydraulic jacking off the leading edge of the lining behind the head. The TBM usually thrusts forward to the extent of its jacks, then grips the side of the tunnel and retracts the jacks to allow segment erection before jacking forward once again. The 'downtime' also allows for regular maintenance for replacing cutter heads, drilling forward exploratory holes, lubrication etc.

The TBM also has a 'maintenance train' running on rails behind it, of up to 200m long of connected sub modules, which the head pulls along as it jacks forward. The train contains the logistics such as power packs, transformers, spoil handling, temp storage, grout pumps, spare parts, mobile workshops etc. As the lining has to be the full bore of the tunnel, it has to be broken down into sections that can pass through the centre of the train and then be re-erected behind the TBM cutting head. This erection of sections has to be carried out as close to the head as possible to minimise the length of unsupported ground behind the head for both worker safety and technical reasons i.e. the ground could collapse and trap the TBM from the sides. Add in any water issues or soft flowing ground and the erection has to carried out in a sealed/semi sealed environment as well. As a gap will also exist between the final outer surface of the lining and ground, this has to be pumped full of weak cement mix or 'grout' to fill the voids, both providing a waterproof annular surrounding for the tunnel as well as avoiding any point loads from uneven ground on the lining.

With the TBM head (for a 8m dia tunnel) weighing around 200 tons (with a main bearing of around 30 tons) and lining segments to be handled of 1-2.5 tons ea, (maybe 8-12 segments per metre of lining) you're not playing with small numbers to shift around and construct. Appreciate that Elon's tunnels will be smaller diameter but current tunneling methods are continuously being optimized and the TBM design is matched to the predicted ground conditions, therefore there is next to no chance of increasing tunneling speeds by a factor of 10x as is quoted..

As someone who has worked in hard rock tunnels 10,000 feet underground, where the rock was under some much pressure it was continuously 'spitting' small flakes at you, to shallow 'soft rock' tunnels where you worry about the TBM taking too much ground out and causing sinkholes to appear on surface, I have to call Elon Musk out as full of shit on this topic. Natural geology is not some nice predictable tech environment where you can confidently predict the ground conditions unless you have literally drilled a pilot hole along the entire length of the planned tunnel alignment. Even then for example you stand a chance of missing a boulder 1m away from your pilot hole that will cause a 2hr delay down the line whilst some poor worker has to enter the TBM face and and manually break up the boulder before it takes out half the TBM cutter heads.