Monday, June 13, 2016

The rule for distinguishing the serious Hyperloop articles from the BS

Good reporting on the Hyperloop spends most of its time on infrastructure and cost (mainly infrastructure). Bad reporting spends most of its time on other aspects of the story, such as the technology of the pods, the personality of Elon Musk, and the widely discussed but never specified regulatory obstacles. There may be exceptions to the rule or caveats I should add to it, but I've been digging into this topic extensively and I haven't come across any.

You can make similar criticisms of much, perhaps most writing about supposedly world-changing technology. Issues like infrastructure and implementation costs are routinely glossed over or omitted entirely. The Hyperloop, however, is especially problematic both because of the exceptionally large role infrastructure plays in serious discussions of the proposal and because accepting Musk's cost estimates at face value requires pushing aside consensus opinion on some well established points.

Here's how I put it in an earlier comment thread.

It's useful to step back and think of this in terms of knowns, known unknowns and unknown unknowns. There are a lot of aspects of building an intercity vactrain that are so far outside of our range of experience that any cost estimate has to be highly speculative -- independent experts tend to think Musk is being highly optimistic in these parts of his proposal but they can't say conclusively that he's wrong – but when it comes to things like putting large structures up on pylons or down in tunnels, we have a lot of relevant experience.

Remember this one?

[Michael L. Anderson, an associate professor of agricultural and resource economics at the University of California, Berkeley.] said that while some of the infrastructure is novel, the elevated guideway was not unlike existing structures such as the Bay Area Rapid Transit's aerial tracks. For the Hyperloop's tracks, that alone would cost in the tens of billions. As for the pipeline for the cars, he said, oil pipelines are $5 million to $6 million per mile, and they are seven times narrower than the Hyperloop's would need to be. In addition, the Hyperloop track could not change direction abruptly the way an oil pipeline could.

"It really has to be built to much higher standards than anybody has ever built a pipeline to," Anderson said.

As a general rule, elevating structures greatly adds to the cost. Building something to exceptionally tight tolerances (such as those required to maintain a near vacuum over hundreds of miles of track) greatly adds to the cost. Doing something big for the first time usually has the same effect. The third is a known unknown. The first two are pretty much knowns. Engineers have been dealing with these questions for a long time.

Unlike pure science, I don't think it's possible to write about technology effectively without seriously addressing issues like cost and implementation and, in general, viability. Implicit in pretty much every story on new tech is the promise of impact in the near future, and if the tech isn't viable that promise is deeply misleading.


  1. Mark:

    One reason perhaps for proponents of this sort of scheme to emphasize the regulatory aspects is that this gives them an automatic out when it doesn't happen: they can blame the regulators.

    1. Andrew,

      I think you're being overly cautious with the “perhaps.” There is a substantial industry built around hyping just-around-the-corner breakthroughs and it can't survive if it limits itself to the current rate of technological development (you'd need to go back to the late 19th/early 20th Century or the height of the PostWar boom to find the necessary rate of actual advancement). The “those darned regulators” line is the caveat that allows everyone involved to keep to the narrative and maintain their reputations.

      Of course, there are cases of badly written rules and regulations hindering progress (I'll bet our friends in biomedical research would be happy to suggest a list), but in the cases you're most likely to hear about – Hyperloops, drone delivery, autonomous vehicles – obvious issues involving design or implementation costs or infrastructure are glossed over while vague and often hypothetical regulatory issues are held up as the reason you can't have _____. In that last example, regulatory concerns may have been the easiest part of the problem to crack. Not just here but in countries like Sweden and the UK, legislators have proven remarkably willing to change laws that block testing of driverless cars (at some point, the almost fifty-year old Vienna Convention on Road Traffic will probably have to be amended, but even there, Volvo has shown that a sympathetic government can buy you a lot of wiggle room).