Would you like to ride to work in a jet-powered roller-coaster gun? Elon Musk thinks you would.

Musk, of SpaceX and Tesla TSLA fame, recently tweeted that he had received “verbal” approval from the U.S. government to build a “Hyperloop,” a train system that would enable travel between Washington, D.C. and New York in just 29 minutes, some time in the near future.

This is, without a doubt, an extremely ambitious project. Is Musk promising more than he can deliver?

A Hyperloop is essentially a jet-powered, high-speed train that travels close to 800 mph through a tube and is accelerated by magnets. The idea is certainly very futuristic, but hardly new, as Hyperloop is a type of vac-train. First proposed as early as 1799, vac-trains are trains that travel through a tube that has little air inside and can travel as fast as thousands of miles an hour. Why? No air means very little friction. It’s why spaceships travel so fast in space—with no air, there is little to slow them down.

But having a total vacuum in a pipe is hard. Pipes are leaky—just think how much your faucet drips. Hyperloop gets around this major issue by only having a partial vacuum in the tubes (meaning they can leak a little). With a partial vacuum, some air will begin to build up in front of the train as it travels, which would normally slow it down. But Hyperloop uses a jet engine inside the train to suck what little air builds up in front of the train out of the back, allowing the train to move with hardly any friction. This is what makes Hypeloop different from most vac-train concepts.

In this sense, the Hyperloop is not a breakthrough of science, but of engineering—using existing technology in a novel way. Indeed, recent proof-of-concept tests seem to indicate that the Hyperloop team is making progress toward a working demonstrator.

True to form, Musk has given few details on exactly how this new Hyperloop route might look. At face value, it appears Musk wants to build a (mostly?) underground Hyperloop route between New York and D.C., which are about 200 miles apart. But constructing a 200-mile tunnel is no small feat. Currently, the longest railway tunnels are all under 50 miles long.

To build this mammoth tunnel, Musk will use another fresh venture of his: The Boring Company, which specializes in underground tunneling. The Boring Company will also use novel engineering to tunnel faster, cheaper, and better than anyone else; a feat which seems reasonable to achieve when you have billions of dollars and excellent engineers at your disposal—which, realistically, solves most engineering problems.

But is it even possible to build a Hyperloop train that moves people quickly between New York and D.C.?

Probably. The technology is there; they just need to engineer it.

Is it possible to build a Hyperloop train into a 200-mile underground tunnel on a reasonable timeline that moves people in 29 minutes and isn’t prohibitively expensive?

Probably not. Musk would need to quickly and concurrently mature two, completely undeveloped ideas in parallel. There’s only so much that can be achieved, even with endless money and engineers. And let us not forget that Musk has failed to deliver on big promises before: the Falcon Heavy is years behind schedule.

But perhaps before we ask can we build the Hyperloop, we should be more concerned with whether we should. The cost of this Hyperloop project is likely to be astronomically high. A different Hyperloop route between Los Angeles and San Francisco is conservatively forecast around $6 billion and doesn’t require a 200-mile long tunnel be drilled.

Consider this: There are about 140 round-trip New York-D.C. and D.C.-New York flights each day with 100 passengers (max) on each 90-minute flight, and there are 120 New York-D.C. and D.C.-New York Amtrak trains with about 300 passengers on each three-hour train ride.

A 29-minute New York-D.C. Hyperloop could save, at most, a combined time of about 6 million minutes (12 years) across all commuters each day. But, imagine you could build a better public transit system inside New York that saves every metro passenger 10 minutes every trip. There are nearly 9 million passengers that use public transportation in New York each day, so you would save a massive 88 million minutes (167 years) across all commuters each day.

In my mind, Musk’s idea has two components: building tunnels quickly and cheaply (The Boring Company), and a futuristic high-speed train (Hyperloop). The Boring Company is, unironically, the most interesting part of this idea. The Boring Company could facilitate a whole new subterranean transport network within a city. Available to all, it might allow entire populations to move through a city faster. Conversely, the Hyperloop is a bit like a rocket: very expensive, and moves few people very quickly to places most people don’t need to be.

Currently, most cities don’t even have a modern public transport system that is fit for purpose. The Boring Company could potentially change that by making tunneling quicker and cheaper. Why then, should we pay to dig large and expensive tunnels between cities when we could build smaller systems inside a city that serve more people? Even as a rocket scientist, I realize that we often need more roads, schools, and hospitals than extra rockets. But if we’re going to spend billions of dollars improving transport, do we really need a jet-powered, roller-coaster gun? Perhaps we should be focusing on something a little more boring.

Leon Vanstone is a rocket scientist and aerodynamic engineer.