Hacker News Reader

Comment by ryandvm | Hacker News Reader

I'm not sure how the rest of it will pan out, but we are right on the cusp of a pretty significant job takeover by robots. Truck drivers.

Consider that we already have vehicles (in prototype) that can outperform a human behind the wheel. Hell, Stanford is sending an Audi TT up Pike's Peak at 100+ mph (http://bit.ly/9akGU1).

And consider that truck drivers are basically meat-based robots anyway inasmuch as they are all practically interchangeable.

Once the first trucking company gets the idea to start replacing drivers, that's going to put them at such an insane competitive advantage that the rest of the industry will have no choice but to follow suit. No downtime, no accidents, far cheaper, no irresponsible behavior.

Once the process is started, the entire delivery driver industry is going to be obsoleted within 12 years.

That will be the first time a massive number of people perk up their eyebrows and start wondering if their job is far behind...

View on HN

This assumes the Teamsters have no power, which they do. The power of the Longshoremen to prevent substantial automation (compared to other major ports) in the Port of Los Angeles and the resulting productivity deficit vs the Port of Rotterdam for example appears to be realtively durable.

Once the entire logistical chain is fully automated the teamsters will have much less power. After all, if you don't need workers it doesn't matter if they strike.

A strike seems like the perfect time to introduce automation.

The only thing you're not including in this equation is votes. Union leaders simply need to convince lawmakers that it's in their best interests to pass favorable legislation limiting the adoption of new technologies. It's certainly been done before.

This is where America's organization as a collection of states is helpful. At least one or two states will be willing to let automation take over once it's developed, and it's effectiveness there can become leverage to spread it to other states.

That seems to assume a (almost) simultaneous switch from manual to automatic. In reality, it would probably be done as a small test case to see how well it works and be expanded slowly. A strike during this process can still be damaging.

In my country buses have a driver and a charger, who collects the fares.

Some companies started replacing the chargers with automatic fare collection equipment. But it turned out they weren't allowed to fire the chargers. So there's still a charger in each bus who will sell you the ticket so you can put it into the machine. Oh, you bought a ticket before boarding? Well, the charger still has a job to do: to ensure you do use the ticket (the driver can't do it because the gate is in the middle section of the bus).

In contrast, I purchased a ticket for and boarded a recently-developed commuter train in my area and not once on my journey did I encounter anyone who worked for the transportation company. Not even on the crowded return trip was there a ticket check.

CalTrain (San Jose to San Francisco) kind of works like this. Sometimes tickets are checked, sometimes they aren't.

You might find reading _The Box_ ( http://www.amazon.com/exec/obidos/ASIN/0691136408/conmanlabo... ) interesting. When shipping containers were first used in 1956, the various longshoreman unions were really unhappy and fought quite hard to keep the status quo.

It gets more exciting when you have companies creating robotic forklifts (http://www.inro.co.nz/automated-forklift/), so the end-to-end manufacturing, loading, shipping and unloading could potentially all be automated.

I wonder if a robo-shipping vessel would be next? With radar, anti-pirating water cannons and no accessible decks, galley, etc it would be a huge boon for shipping. Then again, I don't know how uneventful those journeys are.

Hopefully the govt can step up with some way of re-educating employees obsoleted industries into something that interests those citizens.

To be honest I imagine that a robo-shipping vessel would be easier to create than robo-trucks. for several reasons:

* The ships are larger and more expensive and hence the add on expense is smaller, comparatively

* There's less traffic around and the traffic that is around is likely to have on board beacons, etc

* The ships already have more sophisticated instrumentation than trucks (radar, etc)

Also, I would imagine docking would be the hard part. You could have a local human crew take over that part, perhaps.

That's already done. Most large ports require a licensed local pilot to board each ship and guide it in and out.

But there are also vulnerable to piracy.

Big cargo ships are already vulnerable to piracy and the handful of crew on board does little more than give the pirates some hostages.

If the ships were fully automated, and remote controlled, you could simply refuse to comply with the pirates and just keep sailing wherever you planned.

Under the assumption that the pirates could not compromise the command and control system, or simply cut the rudder and steer the ship manually.

But yes, it was these points that led me to suggest robot ships.

Why not robosubs?

Because sealing the thing would cost a lot more money. Containerised shipping is all about reducing overall expense.

Modern merchant vessels are already highly automated, to the extent that crew compensation is a fairly small component of total cargo costs. Trying to reduce the crew further would be chasing diminishing returns. While it's possible to build a vessel that navigates itself from port to port they will at least need some engineers on board to fix equipment when it breaks. It wouldn't be safe to leave a robot ship drifting powerless in mid-ocean. The crew is also able to do a lot of routine maintenance during transit. Without a crew, the ship would have to spend longer in port just to have enough maintenance time.

Interesting point. Automation will only continue as long as we can save significant amounts of labor by automating. So we can expect to see more automation in the construction of automobiles and lawn mowers, but in all likelihood the fabrication of microchips is about as automatic as it will ever get.

I can't even imagine the legal repercussions of giant robotic trucks killing motorists. I would be cheaper to keep hiring drivers than paying tens and hundreds of millions for wrongful death suites.

After enough accidents, many municipalities and states would prohibit such trucks, that would render them useless.

In the US, regulation of robotic long-haul trucks will probably be almost entirely at the federal level (interstate commerce). So I doubt municipalities will have the authority to prohibit them.

Interestingly, even if the number of fatal accidents caused by robot trucks was less than the number of accidents caused by human trucks, it would be unacceptable to the public.

The number of fatal accidents would have to go close to zero.

Pie in the sky idea: if politicians had spines they could pass laws that required damages given in cases of robotic accidents to be no higher than damages given in accidents caused by humans, which is an obvious and sensible solution to the problem.

We (politicians and non politicians) are not rational creatures. An airplane bomb scares us more then a .5% increase in road deaths, a shooting more then an infection, etc. I am not immune to this. Getting squashed by a robot truck scares me.

Right, that's why the hope is that politicians will put aside their irrationality and do what is best for the people, even if the people don't want it. It's a pie in the sky idea because it would require politicians (and those who would replace the ones who get voted out due to popular ire) have moral backbone and courage to do unpopular but right things.

If people are scared, perhaps it makes their lives worse.

I don't want to be uncharitable so let me make sure I'm understanding you. Are you embracing the same rhetorical sequence that others often use to argue that mandatory vaccination is a bad thing? I.e. are you arguing that people's fear of a thing ought to be as significant a consideration as the real benefits it brings to them and others around them?

Yes I am. Be uncharitable.

I think it's all a matter of degrees. If the "right" policy offends people's religious ideas, nationalism, instinctive irrationality's, or whatever else, this might have consequences that outweigh the benefits of this policy. People being irrational can do damage.

An obvious example is civil unrest which is often a good reason not to implement liberal/libertarian policies in many places.

> An obvious example is civil unrest which is often a good reason not to implement liberal/libertarian policies in many places.

It's only a good reason because there are other good reasons to implement those policies too. Civil unrest would not be a good reason to begin beheading petty criminals, no matter how scared people were of being pick-pocketed. I'll give it to you that it is a matter of degrees, but the amount of fear that would be required to outweigh even the smallest iota of real benefit would have to be enormous. If robot trucks saved just one life per year on average, it would be morally unacceptable to obstruct their implementation even if they induced the same amount of fear as all the fears of planes crashing and terrorist attacks.

If Robot trucks saved 1 life per year, they would be killing people every week (assuming they replaced all trucks). Every week you would have reports about faulty software, potential hacks and such. It would be massive scares.

There are lots of places where lives are lost because a certain narrative is something that the public can swallow. You could spend billions of virtually useless aviation security dollars to save many lives. How would the public react to 50 terrorism deaths, twice a decade?

I wonder if human nature might help make robot trucks safer on the road, by the magic of risk compensation (http://en.wikipedia.org/wiki/Risk_compensation). A robot truck will most certainly be marked distinctively so the human drivers know there's a computer at the wheel. Human drivers will perceive this robot truck as dangerous (at least I would, and I doubt I'm alone in that prejudice) and thus drive a little safer in its vicinity to compensate. Possibly, driving a robot truck down a street would even reduce the human-on-human accident rate.

And the robot truck companies would be insane not to protect themselves. They would be wise to take a page out of the commercial aircraft industry's playbook, and instrument the hell out of each vehicle. Continuous engine telemetry and redundant camera views of the immediate vicinity recorded on a 30-minute loop, frozen after a crash and a core dump of the "driver" appended to the record for crash reconstruction. It'll be a tough sell emotionally, that the robot truck was following the speed limit and stop lights, but we'll get there.

It sounds to me like you need to start a long haul company that uses robot drivers.

Reply on news.ycombinator.com