Friday, April 28, 2017

Practical Problems With Robotic Cars

Will robot taxis take away jobs?   Are they as close as we think?

Several people have written me saying that "guaranteed minimum income" will be necessary in the future as robots and "AI" take over all of our jobs.   I think this is a little over-stated to say the least.  Jobs have disappeared when technology made them obsolete, and this has been going on for decades now - centuries even.  Somehow we all manage to find something to do.

Of course, it may not be at the pay level you'd like, but then again, no one is ever satisfied with how much money they make, no matter what the pay level.   And that is no excuse to take money from other people, which is all "guaranteed minimum income" is all about.   In order for "everyone" to get $10,000 each, some folks are going to have to pay $40,000 more in taxes, maybe hundreds of thousands more.  But I digress.

The "threat" of robotics and "AI" is  overstated in the press, which is staffed by a bunch of liberal-arts know-nothings, well, at least know-nothings with regard to technology.   All this talk about "AI" crap is a case in point.   Computers are not anywhere near from becoming sentient beings.   Our most sophisticated computer in the world - or even all of them put together, doesn't have the computing power of a grasshopper, much less a mouse.

It is like the colonization of space argument - people with no technical background think you can just go to Mars and start farming potatoes and everything will work out OK - hey we all saw the movie, right?   But the reality of trying to build every damn thing you need without the huge resources in air, water, and not to mention oil, is staggering.

The "AI" of today is really just primitive programs designed to use keywords to  search for data online, not actually "think" about your question and formulate an answer.  And yes, some such programs might pass a "Turing test" and appear to be sentient, but of course, they are not.  Not just yet, anyway.  Not for a good long time.

But alarmist headlines sell newspapers or more precisely generate ratings and click-though revenue.  So the media, staffed by said aforementioned clueless "communications" or liberal arts majors, who think their smart phone is smarter than they are (and half the time is) spout these "robots are taking over the world and we'll all be out of a job!" headlines.   Well, we can only hope they'll be out of a job, anyway, given the drivel that passes for "journalism" these days.

But what about robotic cars?  Trucks?  Airplanes?   Won't these things throw millions out of work?  Where will convicted felons go for work if they can't drive a truck?  If taxicabs have no drivers, what will immigrants with no language skills or sense of direction do for a living?  And who will stall the airplane if we don't have pilots?

All kidding aside, there are a number of technical hurdles - and social ones - before autonomous vehicles take to the highways.  The latter is more of the problem.   One of the first Patents I wrote as a law clerk (nearly 30 years ago) was on autonomous vehicles.  That Patent, for a research division related to Toyota (IMRA) used lidar to detect the presence of retro-reflectors on the highway to determine whether the vehicle was in its lane or not.

As part of this case, I had to research all the activity on autonomous vehicles, and I was surprised that it went back as far as the 1970's, funded by the highway administration.   The University of Pittsburgh, as I recall, was an early experimenter, using a box truck loaded with computers to slowly drive across campus.  On a good day, they hit only two or three students.   I'm just kidding of course, but the idea is not new, and the government has been pushing this technology for decades.

And the reason why is not hard to fathom.   About 40,000 people are killed on the highways every year - about the same number as die of breast cancer.  We all have "awareness" about breast cancer, and telethons and fund raisers, and pink ribbons.   But there is little awareness of the carnage on the highways, other than MADD which wants to put an end to about half of these deaths, which are caused by drunk drivers.

Some Luddites posit that if we go to autonomous vehicles, computer glitches or hardware errors could lead to spectacular accidents - where dozens of cars are involved in high-speed pileups on the freeway, killing large numbers of people.   And I suspect this will happen, just as airliners routinely fall out of the sky (often due to pilot error) and kill hundreds at a time.   But the overall carnage rate will be far less with autonomous vehicles than with human-driven ones, if you can call what humans do these days (eating, drinking, texting, having sex) "driving" in the normal sense.   And since such accidents will have "deep pockets" in the form of the technology companies that make the hardware and software that failed, it will be a lot easier to recover damages, as opposed to going after individual insurance companies.

The problem, of course, is what happens in an environment where there is a mixture of autonomous vehicles and human-driven ones?   We are seeing already that humans can cause autonomous vehicles to get into a wreck, as people behave in irrational and unpredictable ways.   Suppose people intentionally act in irrational ways?

For example, you've seen these videos on YouTube no doubt, where kids on their "crotch rocket" motorcycles gather together on the highway and do wheelies at 70 miles and hour, or go over 150 miles an hour, weaving in and out of traffic - usually with dire consequences.   How does an autonomous vehicle deal with that?

Or suppose some kid decides to cut off a line of autonomous vehicles and then slam on his brakes, just for "fun" to see the robot cars all pile into each other?   You might think this is farfetched, but I've seen this happen on YouTube - motorcyclists cutting off trucks and slamming on their brakes in some sort of road-rage incident.   It even happened to me, once, when some dweeb in a BMW bike rode my blind spot for ten miles (smart move) and when I waved him on to pass (with less than all five fingers) he pulled in front of me and slammed on his brakes - a totally dumb move that could have left him dead and me buffing a small scratch off my bumper. 

 People do idiotic things, we should assume that.  And the problem for autonomous vehicles won't be the technical ones, but the social ones - just as the colonization of space will present enormous technical hurdles, but even worse social hurdles.   How would an autonomous Uber taxi work out?  You call the car, it drives itself to you, and you find out the previous user threw up in the back seat - or left all his McDonald's wrappers in there.   It already has happened in New York City with "Zipcar" (remember that?  It was supposed to be the wave of the future as well!).   People show up to rent their Zipcar and find the interior trashed by the previous user.  It then comes down to a he-said, she-said argument as to who left the trash in the car, usually with the trashy person winning in the end.

And of course, there will be people who will stage accidents with robotic cars, claiming injury, perhaps with the aid of a friend who cuts off the robotic car and zooms away.   That much is predictable.

And the cost of such vehicles isn't going to be cheap.   And I think about half the actual cost will be insurance, just as half the cost of a new general aviation aircraft these days is liability insurance for the manufacturer.  Owning a car may indeed become rarer in the future as few can afford to buy one.   And whether non-autonomous vehicles will be allowed on future expressways is up for debate - if they are indeed allowed on public streets at all, after a certain point.   At that point, the public will be at the mercy of the autonomous taxi companies.   Having all your eggs in the Uber basket could be a risky move, given how aggressive and underhanded that company has been.

But are these things "just around the corner" as journalists like to say they are?  I am not quite convinced.   It sells eyeballs to say these things, but I suspect there will be significant delays to the introduction of the autonomous car.  Like Elio's three-wheeler (what ever happened to that?) it will be introduced "next year" for years to come.  And "flying cars"?   Again, that is part of a government-sponsored research program, but so far, it looks as though most flying cars will be a novelty, not a reality for most Americans.   And autonomous helicopter Uber taxis, well, that may take some doing.  There is a reason they closed the heliport on the Pan Am building years ago.

Maybe I am jaded, but the future has been late in arriving for several decades now.   We're still waiting for our clean and cheap atomic energy, our undersea tunnel to London, the rotating space station with the Hilton Hotel aboard - and of course our Moon and Mars colonies.   It seems the only real advances in technology in the last few decades have been more inward-looking.  We have better computers, better data collections, better ways of cataloging and monitoring our behaviors (while at the same time, they become more and more irrational).

But even if these technological wonders come to pass, I am not worried about legions of unemployed people being created as a result.  At one time, every phone company had legions of telephone "operators" who connected virtually every phone call.   Every major corporation had an office building in Manhattan with legions of accountants using ledgers and adding machines - and later on, primitive punch cards.  At one time, every attorney had their own secretary who typed all his (and it was a his) letters by hand.   At one time, even middle-class families had a "maid's quarters" or a live-in cook, in an era before dishwashers and microwaves.  At one time, a typical farm had dozens, if not hundreds of employees, which today have been replaced by a handful.

All those jobs went away - or most of them did.   But somehow, we managed to find new jobs, new things to do, new vocations.   Unemployment is at all-time record lows, even as we still import huge amounts of goods from overseas.  The "threat" of robotics to jobs is about the same as the "threat" of imports.   For some reason, it seems to have taken a long time to put us all on the breadline.

So what's the point of all of this?   Well, maybe that for starters, we need to take the media with a grain of salt and realize they are not informing us, but selling us to advertisers, and thus want us to click on stuff we think is cool.  So they sell the idea that robots are going to take over or whatever.   But it ain't about to happen just yet.   And when it does happen, it will be a gradual thing that will take decades if not a century or longer - and in fact is a process that started decades ago.  Yet people are still working.

So the idea that we have to "redistribute the wealth" to compensate for a robotic "AI" future that has yet to come to pass is sort of idiotic.   It is just another gambit on the part of the Left to take money that doesn't belong to them on the premise that they have so little and others have so much - the argument the slackers and layabouts and communists have been making since time began.

Might I suggest that they just get a job instead?