Comment by jcranmer | Hacker News Reader

Comment by jcranmer | Hacker News Reader

Parent Full thread

Hyperloop feels a lot like supersonic, vactrain, maglev PRT--which is to say, 4 technologies that have all ended up miserably failing.

It's worth noting that the skepticism involved is not in any way displaced by the recent test. The major problems with Hyperloop boil down to:

1. High-speed switching, necessary to expand more than single-city pairs

2. Ability of the loop to maintain tolerances and structural integrity after weather and seismic considerations (particularly CA!) take their toll

3. Right-of-way feasibility, both in terms of cost-effective land acquisition and in terms of curve radii tolerances (note that even 300 km/h HSR systems in practice often have effective speed limits lower than that on much of the track)

4. The ability to get safety certifications--and the implications that would have on headways and consequently passenger throughput

5. Ability to maintain the necessary pressurization of such a long pneumatic tube

6. The cost/benefit of hyperloop relative to other technologies (like conventional HSR, maglev, even highways and airports).

The way that the company seems to be shrugging off all of these questions is not a good sign.

Reasonable points, but just because technologies have failed in the past is not predictive of what will come in the future. It's often times hard for us to observe progress in technology development, particularly when exponential change is involved, but the pace of development keeps accelerating and all of the sudden you reach viable points in technologies that historically took a very long time to see any advances or progress.

> Reasonable points, but just because technologies have failed in the past is not predictive of what will come in the future.

True.

But if the only track record is a string of failures, it would be madness to expect this time to be different, particularly as the main problems that plagued other technologies still persist and are being repeated while being brushed aside.

> particularly when exponential change is involved

Even minor changes have disastrous results. Case in point: BART. It's a conventional railway service in the same fashion that's being done for decades all over the world, and somehow the "exponential change" that has been attempted in that project was more than enough to render it a failure.

The main reason why people keep investing in tried-and-true technologies is because the risk of trying it again and succeeding is far lower than reinventing the wheel, even if it's a fancy, popular wheel that is promised to go very fast.

Maglev hasn't failed miserably. China is starting to adopt low-speed for intracity travel:

http://news.xinhuanet.com/english/2016-05/06/c_135339867.htm

https://en.wikipedia.org/wiki/Line_S1,_BCR

San Francisco/LA is only 400 miles. We need a 700 mph train because there's no affordable direct route.

The Chinese maglev projects are white elephants. All they demonstrate is that maglev can kind of work if you have a government willing to write a blank check in order to demonstrate the power of the glorious People's Republic; but this says nothing about their practicality in the rest of the world, where their cost/benefit ratio has to withstand comparison against other technologies.

The one maglev costs $656,729,196, if I did the conversation right. That's 9 miles of track. Is that expensive for a major city? The Second Avenue subway cost a lot more, for example:

"The first phase will be within budget, at $4.45 billion. The total cost of the 8.5-mile (13.7 km) line is expected to exceed $17 billion.'

This article says that low-speed maglevs are competitive:

http://news.xinhuanet.com/english/2016-05/06/c_135339867.htm

"The line costs between 150 million yuan and 250 million yuan per kilometer, making it highly economical compared to subways and light rail, which cost up to 800 million yuan and 300 million yuan per km respectively"

Do you have any facts to back up your "white elephant" claims?

> The Second Avenue subway cost a lot more.

This is an apples and oranges comparison. Big construction projects will always cost a lot more in the United States because of high labor costs, getting a plan that is acceptable to all levels of government and other stakeholders, environmental impact studies/minimization, and all those other pesky things that China doesn't have to worry about. You would have to do a comparison with a Chinese rail project to get useful results.

I think when you say glorious People's Republic you invite cross border comparisons.

I also agree that the comparison with a Chinese rail project would be more interesting, but the miserable political climate for large projects in the US would apply to a good Maglev technology just as much as it would apply to a junk Maglev technology.

> "The line costs between 150 million yuan and 250 million yuan per kilometer, making it highly economical compared to subways and light rail, which cost up to 800 million yuan and 300 million yuan per km respectively"

Putting a viaduct in the middle of nowhere, where eminent domain committed by a totalitarian state lets you get access to very cheap land usage, is not a fair comparison with having to dig holes in a high-density city.

I'm responding to the claim that it's a "white elephant" project in China. I'm not asking why we can't do it here.

Yes, China can build a 21st century infrastructure more easily than the US. They already have 10,000 miles of high-speed rail, for example, while the US has none.

>San Francisco/LA is only 400 miles. We need a 700 mph train...

Or a bed on a truck. $48, no subsidies needed, get there in ~0 waking hours, sold out, mostly 5* reviews http://www.autoblog.com/2016/05/09/sleepbus-la-sf-overnight-...

LAX/SFO is the second busiest air route in the US.

http://cahsr.blogspot.com/2009/10/la-sf-nations-second-busie...

6.3 million people fly it every year.

The thing is - if you look at Musk's other ventures they were in dire straits for a long time, critics called them impossible even, but he managed to turn them solvent ... somehow. SpaceX for instance was a money-pit for the longest time and was lambasted by every sensible investor. So while I share your skepticism, the "shrugging off" you are observing here is pretty much par for the course. True, Musk does indeed have a tendency to over-promise and over-sell but this is a trait he shares with almost every CTO.

> Neither SpaceX nor Elon Musk is affiliated with any Hyperloop companies. While we are not developing a commercial Hyperloop ourselves, we are interested in helping to accelerate development of a functional Hyperloop prototype.

http://www.spacex.com/hyperloop

Unless this has changed, Musk just had the idea; he's not doing any of the implementation.

Not to mention it might be economically cheaper to pursue automated automobiles since we already have the roads built.

Certainly in the short term, perhaps in the long term, but even automated cars come with externalities we might want to reduce such as costs/inputs of manufacturing, pollution, addition time compares to a fast point-to-point link, etc.

For some popular long distance routes, I could see the Hyperloop being useful. Between all reasonably large cities? Probably not.

You're comparing capex vs opex, once you get past a certain period of time. It's possibly cheaper to build automated cars - we have the infrastructure to do so already, and we're close to having the software - but is it actually cheaper to run them, after all externalities are considered?

> Not to mention it might be economically cheaper to pursue automated automobiles since we already have the roads built.

Automated automobiles won't help cutting trip times, which is the main reason why people pay to fly or to ride a high-speed train.

Not immediately. But in a future where all vehicles are automated it might be safe to have 200 mph highways or cars that automatically drive onto high speed trains, or cars that can also run on rails, etc.

Reply on news.ycombinator.com