Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> The real feature should just be a button that shifts insurance liability to Tesla for the period of the autopilot engagement and charges you a floating premium for it based on traffic conditions per mile/km while the autopilot is on.

No amount of money makes up for human lives. You can pay a premium of $1mil per mile, if the car end up smashing into a family car at 80mph because the system was fooled by a shadow or a reflection it's still a net negative



> No amount of money makes up for human lives. You can pay a premium of $1mil per mile, if the car end up smashing into a family car at 80mph because the system was fooled by a shadow or a reflection it's still a net negative

This is a metric human drivers fail to hit every single day.

There is some acceptable level of risk in shifting driving to a robot. How you draw that comparison and draw the line is tough. It's complicated greatly because the kinds of failures autopilot makes are significantly different from the kinds of failures humans make.


> This is a metric human drivers fail to hit every single day.

Yes exactly, and when they fuck up they suffer the consequences. Which if you were reckless will most likely be years in prison. You shouldn't be able to get away by paying money which is what the comment I replied to asks for.


You're not "getting away" with anything. You're paying someone else to do a job, and that other person writes software to do the job. They're the ones responsible for the software at that stage.


You can't "pay away" your responsibility for human lives like that.

You can point towards the anomymous software engineer you supposedly paid off your responsibility to.

But you're still the one that literally killed a family.

You'll realise there are some things you can't transfer responsibility from, when you're sitting behind the brains dripping off your window.

But at least all the people who have literally zero idea what you just went through, will tell you it's really Tesla's fault who wrote the software and they're responsible really. And there's no amount of money you can pay, to make you believe them.


30,000 plus people are killed on American roads every year. Drivers killed more people in September 2001 than terrorists.

How many end up in jail?


> No amount of money makes up for human lives.

In that case we should not allow cars driven by humans either?


> No amount of money makes up for human lives

Sure, but no human life is worth spending infinite resources to protect.


There is no need to spend infinite resources. Look at the driver facing camera when the accident occurs. Either the driver was paying attention or he was doing something else. It's a legal issue, not a technical issue.


We should delineate between resources (which has a fiat value) and fiat money. I agree that fiat money can't make up for a human life, since if money is destroyed, the remaining money supply increases in value through deflation and no actual resources are destroyed.

However I disagree that a human life can't be equated with resources. Would you destroy $10 billion of value/resources in the economy to save 1 person? I don't think that's worth it, because that $10 billion of value could've been food and medicine which can save thousands of people. This is really just the trolley problem. But the bigger you make the sum (if you don't agree at $10 billion, what about $1 trillion?), at some extreme point you're going to have to say "ok, one life isn't worth that much."


> Would you destroy $10 billion of value/resources in the economy to save 1 person?

You can spin any arguments by framing them in such ridiculous scenarios. My comment is obviously in the context of people able to pay money to get rid of their personal responsibilities.


I was just demonstrating the absurdity of your comment. Taking things to the limit to demonstrate the absurdity of absolutist statements like the one you made is a perfectly acceptable thing to do and is not "spin". You said "no amount of money" and I took that at face value. If that's not what you meant then that's fine, there's no disagreement between us.

From reading the many other replies to your comment, everyone else was left with the same impression by what you said.


It's quite simple. If you drive into a family, killing them, you will realise, as the brains are dripping down your window, that the insurance premium you paid, actually only affects the opinion of people who have literally no idea what you just did.


> No amount of money makes up for human lives

Trivially untrue since you can save one life for under $1000 but almost everyone with $1000 will keep it rather than save the life - revealing that $1000 is worth more than an arbitrary human life.


That is not an example of $1000 "making up" for human life, though.


Just out of curiosity, how do I save one life for $1000?


Givewell has a model for you: https://docs.google.com/spreadsheets/d/1BmFwVYeGMkpys6hG0tnf...

They have more detailed information elsewhere. They are generally trustworthy. Of course I boosted the cost (based on memory) a little, and literalized the value (the truth is that each additional mosquito net has a marginal reduction in the probability of a child contracting and then dying of malaria).

And it looks like I made two errors:

* I had a memory of an older model - the newer one makes it more expensive

* I misremembered the optimal strategy (it isn't anti-malaria, it's Deworm the World at $1003 / child under age 5 saved).


> No amount of money makes up for human lives.

You clearly don't actually believe that, or there are a ton of activities you would stop doing. You wouldn't drive to work, because there is a chance you might kill someone while driving, and you only go to work for money, so why do it?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: