They’ve developed a new machine, Prufroc, that’s supposed to be much faster and capable of “porpoising” (ie angling back up and curving straight) so you can avoid having to excavate a deep station through other means.
If they can get the cost equivalent to a typical surface street in urban areas, they can now compete with surface streets. Much easier to autonomously drive in purpose-built tunnels. Or put a train or a bike trail or moving walkway or whatever.
So they are actually trying to solve the right problem. And using cars is a good idea because cars are really cheap per passenger in terms of upfront capital cost, and they meet the speed, autonomy, safety, and electrification requirements while also allowing use on just regular paved surfaces. If the one-dimensional autonomy works well, they might be able to reduce follow distance at speed enough to exceed a highway lane throughout. And with high occupancy, also a typical light rail train or whatever.
Would be kind of neat to have autonomous driving at 120mph immune to weather and with low enough capital cost that we could afford to crisscross the country with them, without much impact to existing land uses. A clear win if it works. Actually, a clearer win than full autonomous driving on existing roads, as that might quickly clog up surface streets. Plus this is probably technically easier than perfected level 5 autonomy which is still a sort of open ended problem.
Do you mean $10m per mile?
Congratulations, you’ve rediscovered the subway.
If Musk wants to make the argument that he’s found a way to drastically reduce subway construction cost, I’m all ears. But he should jump straight to that, rather than this ludicrous idea of zipping personal cars around under the streets; a proposal that doesn’t come anywhere close to the transit density needed for an actual mass transit system.
Well, he does keep showing renderings of public transit stations full of Teslas and not, you know, people. It's not really a big leap to say that he doesn't understand when he keeps marketing as if he doesn't understand.
Turns out Teslas are cheaper per passenger seat than a subway car plus have less requirement for specialized infrastructure (rails, etc). So if TBC is gonna be doing the whole project, they intend to use Teslas. It’s not a bad engineering choice. With automation (not hard in a tunnel), can achieve similar throughput as a Subway.
Minimum train space with pretty standard signaling can be as low as 90s. With modern moving block systems that can be pushed even lower. The tube trains have a seating capacity of around 250 people and 800 people can be standing. This gives a capacity of between 3 and 12 people per second.
How are you going to inject 1 car per second (to reach the lower level) into the tunnel in a safe way? And even if you could do it - Consider how huge the stations must be. They have to contain 100+ cars plus space to enter/exit them (couple seconds to come to a stop, 10s of seconds to leave the vehicle - with some people quicker than others -, 10s of seconds for other people to enter the vehicle, couple seconds to accelerate and synchronize to the flow..).
Folks who dismiss getting subway-like throughputs haven’t actually done the math.
Sub-2 second spacing is quite unlikely to be used, since stopping distance (with 0 reaction time, which is not too unreasonable with LIDAR/RADAR) dictates around 2s spacing between vehicles (1.9s at 100km/h for teslas I believe). And while humans might be fine with ignoring safety aspects computers can't and shouldn't.
Nevermind that building a subway-light vehicles (more similar to a people mover) is not really the same thing as using slightly-modified Teslas as subway replacement (which is what it seems they will be using for now)..
You know and own everything about the other process so you can match them far better.
This is what Elon does. Hacks processes, not technology.
How is this something that can be solved with the yet to be invented new TBM?
From your link -
"The "ad infernum" theory positing property ownership "to the center of the Earth" has also been eroded. A review of modern American jurisprudence demonstrates that the theory is more poetic hyperbole than binding law, and that broadly speaking, the deeper the disputed region, the less likely courts are to recognize that the surface owner holds subsurface title. [21] Appraisal studies of subsurface projects such as subways, deep storm drainage tunnels, and particle colliders consistently conclude that such projects, built well below the area that the vast majority of surface property owners ever put to use, do not deprive the surface owners of any value."
That is it's design. That is exactly what Elon is doing.
You might argue it can't do what it's setting out to achieve, but this has never been a technological solution.
He's designed a technology to fit with current societal processes rather than the old, get society to accept a technology.