Comment by drzaiusapelord | Hacker News Reader

Comment by drzaiusapelord | Hacker News Reader

Parent Full thread

>universal no-fault s probably the only easily managed way

Easy? Maybe. Best? Absolutely not.

I don't like the idea of being punished financially because some jackass on the road won't drive correctly. Insurance companies love it because now they can raise both our rates and not bother with any kind of investigation. This doesn't mean its good for the consumer. Jackass driver loves it because he never has to reform his jackass ways. Your own link details these criticisms:

Critics of no-fault argue that dangerous drivers not paying for the damage they cause encourages excessive risky behavior, with only raised premiums and a higher risk rating as the potential consequence, and no jury awards or legal settlements. Detractors of no-fault also point out that legitimate victims with subtle handicaps find it difficult to seek recovery under no-fault. Another criticism is that some no-fault jurisdictions have among the highest automobile-insurance premiums in the country, but this may be more a matter of effect than cause (e.g., the financial savings from no-fault may simply make it more popular in areas with higher automobile-accident risk, or high insurance rates may cause more drivers to go uninsured, increasing the attraction of a no-fault system).

How can one party be a "jackass" if both cars are autonomous?

My criticism is of no-fault used today.

From a futurist perspective, I imagine there might be scenarios of people taking manual control and causing an accident or manufacturers writing bad software. Imagine if Toyota's newest update causes 10x as many crashes. Why should I subsidize their incompetence via my personal insurance premiums?

If autonomy happens, ideally, the car companies should be paying insurance. The same way they pay for warranty. Its all in their control and its an incentive to keep them honest and care about making cars that don't crash.

That seems to be the solution. I hate no fault because I trust a stranger only as long as he will suffer consequences of his actions, and under no fault there are none.

Even if we have a no fault insurance scheme for only unmodded autonomous cars that means the manufactures have no incentives to make their cars safer.

Do people really drive differently depending on the insurance environment? I don't care what kind of insurance I have, I don't want my neck broken.

Moral hazards are real things.

http://en.wikipedia.org/wiki/Moral_hazard

I imagine there are many personality types who only behave because it would cost them too much to misbehave. My gut feeling is that this is probably 80% of the population.

"Why should I subsidize their incompetence via my personal insurance premiums?"

If you think about that, that hardly even makes sense. "Incompetence" has been a perfectly normal part of the insurance package for a long time. I assure you measurable amounts of what you've paid in insurance have already gone to "incompetence".

What if jackass in question installs a mod that makes the car drive faster (and therefore more accident-prone)?

As with current insurance, if you mod your car without telling them they will invalidate your insurance.

Once you do tell them you've altered the programming on the cars computer they are more than likely to tell you to find someone else crazy enough to insure you.

The self drive feature of a car is about to get very proprietary, like a flight system there is going to be assurances that the software the manufacturer installed is the same one that's running. Crash investigators are going to have a tonne of information from these cars (Assuming they don't blow up in a Michael Bay style crash).

I have no intention of ever traveling in an autonomous road vehicle. So the GP's question stands: why should I have to pay for Tesla's firmware bugs?

You're under the assumption you'll still be able to afford insurance for a non-autonomous vehicle, something I find unlikely if self driving cars are as safe as they've already proven to be in the million+ miles they've driven.

Use of self driving cars hockeysticks-># of human drivers plummets->premiums skyrocket for remaining human drivers.

Right now, in a lot of states, you can just post a bond in place of having insurance.

It also might work out that self driving vehicles end up reducing the rates for human drivers. I think a lot of road risk actually comes from a small pool of drivers, who will tend to be the most eager to stop insuring themselves. So even if if the pool of human drivers shrinks a lot, the risk involved could lead to lower rates than we see today.

You do realise that 1 million miles is an infinitesimal sample size, right? According to the US DOT, in the past 12 months vehicles have traveled just over 3 trillion miles on US roads[0]. It's also worth mentioning that substantially all of those autonomous miles were on freeways, which also happen to be the safest roads for human drivers. It's simply not possible to draw any meaningful conclusion from such a data point.

[0] http://www.advisorperspectives.com/dshort/updates/DOT-Miles-...

We don't have to quibble on data points; autonomous vehicles will continue to get better, at a faster rate than humans can get better at driving (or actually, you know, concentrate on it instead of eating, using their cellphone, texting, and so on).

Humans aren't obsolete yet, but we're well on our way.

We'll have to agree to disagree. You can put your life in the hands of firmware and I can stay home, and we'll see who lives longest.

I put my life in the hands of firmware every time I fly, at 500mph in commercial aircraft. Hundreds of thousands of miles, seems to have worked fine.

Reply on news.ycombinator.com