Hacker News Reader

Comment by jjoonathan | Hacker News Reader

> With no human operable "helm" there is no way to divert the ship, short of posessing an engine more powerful than the ship's engine

Because engines are impossible to disable, GPS is impossible to spoof, and control systems are impossible to hack even with unlimited physical access.

> no crew to hold hostage

There's a fancy robotic ship and a bunch of cargo to hold hostage.

> the practical amount of looting is a fraction of one container

How do you reckon? I suppose if the value-density of the cargo were low enough then nobody would have incentive to steal it, so they would have to threaten to destroy it instead.

> The real objections should be about failure modes of ships and what a crew can to to mitigate them.

You're absolutely right. But piracy is one of those failure modes.

View on HN

There are low hundreds of attempts and a few dozen successful ship hijackings per year. It is, overall, a very small part of freight theft, the vast majority of which happens on land. This should be obvious: on land, you don't need a boat big enough to float a container to steal a container. You just need a access to a semi-tractor and maybe a trailer.

If you make the technical demands on a successful hijacking higher, I expect these numbers to fall.

I also expect self-driving trucks to sharply cut theft, rather than enable it.

Can you elaborate why you believe self-driving trucks would reduce theft?

They feel like the easiest thing to rob on any low-traffic road - you need two people in masks, one walks in the middle of the road in front of the truck so it stops in order to avoid "traffic accident", the second one takes whatever he wants from the back, and both leave through a side path too narrow for the truck to follow. Even if you get a cop car there within 10 minutes, all you've got is a video without faces.

How is this different than what would happen to a manned truck?

Less predictability = more risk to the robbers, as they have to deal with the driver.

1. It would be harder and less safe to force him to stop the truck - drivers expecting a robbery sometimes don't stop in dangerous situations but attempt ramming the robber or running away at some risk; an automated car would be much more passive since the liabilities for running over a pedestrian are of a much higher priority.

2. For a manned truck to be robbed that way it would generally need to be an armed robbery, and at least in Europe most thieves don't have any easy access to guns;

3. If you get caught, you have a chance to get away from charges - it isn't a crime as such to walk in front of a truck, and the prosecution may fail to prove your connection to whoever actually takes the stuff;

4. This may be done also by young teens/preteens who are almost immune to criminal charges even if they get caught; currently they are limited mostly to shoplifting, but this would be within their abilities.

Consider that a self-driving truck would spend less time parked.

What proportion of loss do you think comes from robberies as opposed to theft when the barrier is reduced to finding an unguarded stationary vehicle? My guess is that robberies makes up a tiny proportion, and that they'd continue to be rare because of the perceived barrier of having to stop the vehicle.

Presumably all you have to do to make a self-driving truck stop is stand in front of it. There should be zero risk of it running you over.

Reply on news.ycombinator.com