Assorted links


The "legal" issue with the driverless care is really a mundane economic issue: who pays if somewhen gets injured or run over by a driverless car?

Why not create a special "victim" compensation fund and be done with it already? That is, the law could grant immunity to the makers of the software for the driverless car in exchange of contributions into this fund.

How do you know what amount to set? Wouldn't that remove one incentive to have safe driverless cars?

Since this is clearly a repeated game, just price contributions at the accident rate in the last period - kind of like we do with current car insurance, really.

Or get rid of the discussion altogether and mandate that every car needs to have insurance according to its intended use - if that includes driverless operation, the insurance company WILL come up with a price. It may be prohibitive, but that really just means the tech is not there yet.

I agree 100%. I'd go so far as to say the issues of legal fault in that hypothetical are trivial; courts deal with thornier issues than that on a daily basis.

And in the long run, the answer will be "everyone will have to get insurance." Big freaking deal, especially since if these driverless cars are as safe as we'vr constantly being told, insurance will be dirt cheap.

"since if these driverless cars are as safe as we’vr constantly being told"

It's not relative safety that will matter to the lawyers, it's the defendants ability to pay. A 23 year old male driving 75 mph in a 45 mph zone with a tall boy Budweiser between his legs while he's texting his girlfriend on his phone is far more dangerous than a driverless car will ever be. But a lawyer can't sue him for a billion bucks.

But the whole point is that if the driverless car is so great, it isn't going to wreck in the first place, so there'll be nothing to sue over. And in those rare occassions that it does wreck, it won't matter because there will be mandatory insurance, which will be dirt cheap because the cars present such a low actuarial risk.

Of course, if it turns out that driverless card do wreck all the time, none of this will work, but that will be a failure of the technology, not of the legal system.

There's always something to sue over when you can parade a seriously injured child or grief-stricken mother in front of a jury.

Sorry, misread your post. I see your point.

"But the whole point is that if the driverless car is so great, it isn’t going to wreck in the first place, so there’ll be nothing to sue over. "

From the linked article:
"Bryant Walker Smith teaches a class on autonomous vehicles at Stanford Law School. At a workshop this summer, he put forward this thought experiment: the year is 2020, and a number of companies offer "advanced driver assistance systems" with their high-end model. Over 100,000 units have been sold. The owner's manual states that the driver must remain alert at all times, but one night a driver — we'll call him "Paul" — falls asleep while driving over a foggy bridge. The car tries to rouse him with alarms and vibrations but he's a deep sleeper, so the car turns on the hazard lights and pulls over to the side of the road where another driver (let's say Julie) rear-ends him. He's injured, angry, and prone to litigation. So is Julie."

We already have a very similar situation with airplanes. Big passenger jets rarely crash, and are actually much safer than cars in that respect, but when they do, very expensive financial settlements ensue.

Part of the reason is that we react far more strongly to many people dying in a single incident, as opposed to even far more people dying one at a time in many separate individual incidents. It's human nature. Plane crashes are also tailor-made for tragedy porn: burning wreckage, cockpit voice recorder soundbites, eyewitness accounts, drama and emotion, and everything else our "if there's no footage it didn't happen" news cycle demands.

Crashes involving driverless cars will probably kill hundreds of people at a time, and thus will fall into the same mental category as plane crashes. They will solve traffic jams by tailgating one another at high speed much more closely than human drivers ever could. But when something goes wrong, there will be the mother of all chain-reaction collisions. Lulz-seeking hackers might try to crash more than your computer. Terrorists might transfer their fixation on the airline industry (while ignoring fatter and easier targets) to car travel instead.

#4 makes repeated mentions of cuts in government spending in explaining the sluggish UK economy, but absolutely nothing is said about tax hikes (most notably the 2010 increase in the top rate from 40% to 50%. Why am I not surprised?

There were cuts in government spending in the UK? Last I checked they were just proposed cuts or cuts in projected increases, not actual cuts.

George Osbourne was on Charlie Rose crowing about his various spending cuts. Tackling the hard problems and positioning Britain blah blah blah. The Chancellor of the Exchequer running the government that introduced the cuts says there were cuts, the opponents on their version of the Fed board who was against the cuts say there were cuts, but our "Austerity works but hasn't been tried" crowd insist they never happened.

I wonder who is right.

There were some cuts in individual programs, but overall the Tories have not reduced the size of government, only slowed its growth. And besides the increase in the top tax rate that Colin mentions, there was a VAT increase from 17.5% to 20%.

Austerity implies deep spending cuts, at the very least an actual reduction in government spending. And those who favor spending cuts also think that tax increases hurt growth. So, no, this is not a good test case for "austerity".

Be sure to write dear George then. It's the keystone of his party's policy but he apparently doesn't know the meaning of the word.

3. let's catch up when the verdict on overall national health spending comes in over the next few years. The "lesson" of letting people see a dozen different doctors in different institutions with no coordination between one another each getting paid for their own isolated specialty service is one we should be forgetting not relearning.

(3) makes repeated mentions of "private" healthcare in Sweeden. I wonder if you can really truly call it "privatization" if the government is still footing the bill for people's healthcare.

"I wonder if you can really truly call it “privatization” if the government is still footing the bill for people’s healthcare."

I understand the gist in this case is 'becoming more private'. But you are also correct that this is no more private than US Social Security is 'private' whereby the government pays, but doesn't directly provide the service.

Isn't the Canadian model just that, where the government pays for all health care, but providers are not government employees? As opposed to UK's NHS, where the doctors are actually government workers just like (in the US) teachers, DMV employees, the FBI, etc?

I think the main point with privatized health care is that it avoids the creation of public sector unions. Public transit works the same way; bus, subway etc. are run by private companies under contract to the state.

But health services are not privatized in the sense that individuals have freedom to choose their own doctors, etc. Sweden does have private health insurance and health care providers. This is increasingly popular as queues waiting for care become very long; some employers provide private insurance as a benefit.

I don't know much about Swedish healthcare, but does the government assign people to doctors? I doubt it but I suppose it's possible.

#5. Arbitrage indeed. It's been done before with stamps to some notoriety. It ended poorly.

Re: #2

Manioc is grown a lot in Asia as a staple starch crop, though I mostly hear it referred to as cassava. Heck, they're even making it into ethanol these days, not that that's really justifiable.

I recall having a baked cassava tuber in Taiwan a long time ago, but I never identified it with the "manioc" that I saw in the somewhat trippy educational videogame The Amazon Trail. Did anyone else ever play that game, or was I the only one who did what the ghostly panther told me to do?

Hmmm ... called cassava in Asia? Maybe, but I am surprised. It is called cassava in the northern stretches of South America and I am guessing in parts of the West Indies also.

I certainly hope the computer driven cars crash less often than the entertainment system on an Airbus or an Android phone.

If there are bugs or security risks, they might push updates. Have a nice day.

You didn't read the fine print, did you? It says: This device is provided on an "as-is" basis. No warranties, expressed or implied, including but not limited to fitness for a particular purpose, are made with respect to this device or software therein.

Ha ha, but the Airbus is also being flown by a computer, you know. And a computer is already controlling your car engine.

Computers have been handling mission-critical tasks for decades. They work because the standard of testing is much higher in those applications. I have spent months as part of a team testing helicopter fuel-control software that would have been a few dozen lines of code at most, if not for all the included safety apparatus.

"They work because the standard of testing is much higher in those applications."
That is not both a complete and true statement. I agree that test is important and rigorous in those fields. But they work because of rigor in requirements, design, and development. Rigor in testing is important, but only as assurance. Testing is an imperfect filter, expensive, and practical only after development has been completed. As Harlan Mills once observed, the best way to assure there are no bugs remaining in the system after test is assure you never find the first.

The reliance upon test is more typical in applications like cell phones (android in particular) where the consequences of failure are less than critical. A "Moneyball" style analysis will show that this is a more expensive way to develop, yet most of the industry is still in the "test and fix" mode. I pick on Android only because it is
1) notoriously buggy
2) related to the airbus entertainment system software
3) the same company behind the driver-less car

Aircraft and medical devices are heavily regulated, yet still have an less than stellar quality control (outside the very best) Software faults in medical devices have killed people and taken down airplanes. I certainly do not want to be on the road with millions of vehicles as buggy as my cell phone. If we are to scale to millions, the standards in avionics and medical devices are inadequate too.

#4: The journalist is trying very hard, and writes nicely. Posen, the IMF, the journalist, and everybody else in sight, according to the journalist, mix monetary and fiscal policies unalytically and indefensibly.

Headline: "Automotive regulations, existential threat for SARTRE"

Hell is other drivers.

Comments for this post are closed