The air on the outside is still going to *aggressively* want to rush in through any little crack.
Air is not magical. You can't put a pinprick in a partially evacuated tube and have it just suddenly equalize. Viscosity on the order of the size of small cracks highly limits the rate at which air can migrate in. A little crack or a leaky seal is simply not enough to overcome an air compressor.
To put it another way: the pressure differential here is approximately one atmosphere. Large trunk natural gas pipelines have a pressure differential of about 13 atmospheres. By your logic, a natural gas distribution infrastructure is utterly impossible because "the natural gas on the inside is still going to *aggressively* want to rush out through any little crack".
Let me see if I've got this straight: we can't build regular maglev trains because they're super-expensive (the engineering, construction and maintenance would be incredibly difficult), so... we'll just make it that much harder by wrapping a (partial) vacuum tube around it???
First off, let's make this clear. Hyperloop is not Maglev. In fact, the design document notes that they could use Maglev, but dismisses it as too expensive: "A viable technical solution is magnetic levitation; however the cost associated with material and construction is prohibitive." Hyperloop uses air bearings - skis operating in ground effect with the pipe.
Maglev trains are expensive for many reasons. The cost of having the track be able to provide forward propulsion however usually represents only the tiniest fraction thereof. First off, you have the reasons that rail is expensive, period (right of way costs, environmental reviews, and all of the other overhead). Then you have to have the entire route be able to lift up a multi-dozen to multi-hundred-tonne train. Not just propel, but actually hold it stably in the air, which is a far more difficult challenge for many reasons than propulsion - you either have to have an extremely precise computer-controlled fluctuating magnetic field in a train with hanging magnets, or you have to have the entire track be magnetized or be able to magnetize, in a manner that resists dynamic instability.
Hyperloop only involves propulsion, and the accelerators represent just a few percent of the length of the track. It's a tried and tested technology, use around the world, and their budget for it is in-line with industry norms. There are all sorts of trains today that use linear accelerators, almost all of which represent way more length of accelerator than Hyperloop needs. Examples include
Airport Express in Beijing (opened 2008)
AirTrain JFK in New York (opened 2003)
Detroit People Mover in Detroit (using ICTS) opened 1987
EverLine Rapid Transit System in Yongin (opened 2013)
Kelana Jaya Line in Kuala Lumpur (opened 1998)
Scarborough RT in Toronto (using UTDC's (predecessor) ICTS technology - opened 1985)
UTDC ICTS test track in Millhaven, Ontario
SkyTrain in Vancouver (Expo Line (using ITCS) opened 1985 and Millennium Line opened in 2002)
Limtrain in Saitama (short-lived demonstration track, 1988)
Nagahori Tsurumi-ryokuchi Line in Osaka (opened 1990)
Toei edo Line in Tokyo (opened 2000)
Kaigan Line in Kobe (opened 2001)
Nanakuma Line in Fukuoka (opened 2005)
Imazatosuji Line in Osaka (opened 2006)
Green Line in Yokohama (opened 2008)
Sendai Subway Tzai Line in Sendai, Japan (under construction, opening 2015)
Line 4 of Guangzhou Metro in Guangzhou, China (opened 2005).[5]
Line 5 of Guangzhou Metro in Guangzhou, China (opened 2009).
Line 6 of Guangzhou Metro in Guangzhou, China (opened 2013).
If maglev was a good idea, it would have already been built.
1. This isn't maglev. It's a ground effect air vehicle inside a tube that simulates high altitude. Levitation works via air pressure.
2. Maglev has been built: Shanghai's Transrapid, Japan's Linimo, and there are two under construction, one in Beijing and one in Seoul.
That's not to say that there's no engineering challenges - there are. I think the biggest thing for the test track to prove will be the air bearings, to show that there's no high speed instability modes with them in a realistic tube. Most of the other stuff is pretty mature tech, however.