Sunday, February 21, 2016

The Possibilities For Mayhem With Self-Driving Cars

Comments to Steve Sailer:

 * One thing is for sure, self-driving cars will be a boon for terrorists. No need to kill yourself nor risk running out of courage on the drive to the target. Just pack a self-driving car with explosives, place a couple of dummies in the car and give it a target. Sounds like a winning combination.

 * Beyond the automated bomb carrier for crazies and a way for drug lords to move drugs around without mules and patsies, these cars will be very tempting targets for computer hackers. I’m not talking about script kiddies and other wannabes but the hard cases who crack companies like Target and Home Depot like they were eggs. Oh let’s not forget hostile foreign nations either who have it in for Americans. The first thing the bad guys will do is steal or buy one and dissect it, the same way safe crackers buy safes to learn how to crack them. Once they figure it out, they’ll run a couple rich dudes off the road in some spectacular fashion or simply turn the the cars on when they are in the garage and gas pipe the people living in the house then have the car go rogue like some deranged criminal. Then watch the stock price of Google and the car manufacturer tank. The possibilities for mayhem with these robot cars will be endless.

 * Anyone notice the commonality between self-driving cars and sex robots? Both of these nerd projects apparently have the goal of de-skilling teenage boys. In 20 years we’ll have a generation of young men who won’t know how to drive, and who won’t have to acquire social skills for dating girls. I have to wonder if some of the same people play a role in developing both of these technologies and in propagandizing us about their alleged advantages.

 * Keep this in mind: autonomous vehicles do not need to be autonomous everywhere in order to be viable products. Look at Tesla autopilot and, soon, Cadillac’s supercruise. Good on highways, not on surface roads. People will buy that. Drive to the freeway and then let the freeway part of the commute get done by computer. Makes the commute easier. The range of conditions in which autonomous vehicles can operate will expand with time. As soon as an autonomous vehicle can handle some local roads a blind person who lives at a local road address an autonomous vehicle can handle can start taking trips to every place that the autonomous vehicle can go to. That’s a step-up in mobility. As for avoiding pedestrians in parking lots: There’s already a pedestrian collision avoidance option on some luxury cars. The tech for doing this is going to get better every year. Specific addresses in Santa Monica in 2025? Hard to say. But certainly some starting home addresses and destination restaurant addresses will be viable for an autonomous vehicle in 2025.

 * The other alternative would be to have a homogenous, high-trust society like Japan’s, where the little kids just take public transportation home from school.

* I would guess that the very situations which we ourselves imagine would create the most difficult scenario for a self driving car — where a crash is about to take place — are the situations in which the self driving cars will do far better than a human driver.

It takes us a good deal of time to process something dangerous, but a computer with the proper sensors should be able to figure that out almost instantaneously, and know what the best response to it is to avoid a crash or reduce its harm. The physics here, of the objects in question and the car itself, and of the brakes and/or accelerator and/or steering, are pretty basic and easy to compute. (To me, there seems to be analogy here to how many people used to think that the hardest thing to get a computer to do was play chess well, when, in fact, it was among the easier things to accomplish).

The things that seem to elude the systems at this stage seem instead to be odd things one wouldn’t really anticipate. Evidently, the Tesla self-driving car, as I recollect, did some weird thing when it came to certain exits on a freeway, going off when they shouldn’t. But those are presumably things that are ultimately pretty tractable, once known.

What I personally would worry about more would be situations which are more predictably difficult, such as bad weather or night driving, where the information from the sensors may be greatly impaired.

* It’s easy to imagine, but why be terrestrial-bound? The unique selling point to raise the billions would surely be to float safely above the potholes, gridlock, mangy dogs, aggro humans and revenue-raising highway cops. This thing can’t run until it leapfrogs.

* Detroit has a light rail system that nobody uses and that loses money. It’s also terribly named. It’s called the “Detroit People Mover.”

* Pedestrians are easy for a self-driving car.

They have complete omniscience of what’s around them in a 360 degree radius with sensors. The only issue is pedestrians coming out of something like obscured corners the same way they are big issues for humans, if cars are connected and share information, probably not an issue at all.

The real issue is can we get sensors that work in all conditions? Eventually I hope so, we’ll have to. Decision making is the easy part with these cars.

Also if self-driving cars become a thing our whole country will be reshaped to fit them.

There will be very clear divisions between traveling zones and drop off zones. It won’t be possible for pedestrians to cross into traveling zones, same for cars in pedestrian zones. We won’t need parking lots, cars will just drop off and go to centralized area until someone else calls for them (like automatic taxis). We won’t own our own cars, we’ll subscribe to a service that sends a car whenever we need one.

Cars will congregate in a local standy by area where they detect higher density, people can schedule cars to be somewhere, e.g. before they leave for work.

If we can truly get self-driving cars working, the whole world will change. Cars are cheap, private, and it’s easy to build a ‘track’ for them. People will always prefer cars over public transportation, bikes, etc..

What I look forward to most is all of the jerks who are sitting on and overvaluing the land in prime city areas and see their values crumble. Who cares if I live 30 miles away from Manhattan, I can just take a nap/work while I commute there.

Density won’t be a concern with self-driving cars. We’ll be able to travel the same distance in half the time, and we can do whatever we want while we wait.

* You know how we’re all concerned about the power the SJW/lefty authoritarians have over our activities online? Now imagine them having that same power over most of our offline activities as well. That’s the potential of self-driving cars.

Say you want to make a 2am run to Taco Bell. You summon a self-driving car and enter your destination. The car measures your weight and body fat, and its Bloomberg Chip rejects your destination.

Or, say you summon a self-driving car to go the pub, after an argument with your spouse. She uses her smart phone to falsely accuse you of domestic violence, and that triggers the car to deliver you to your local police station instead for a debriefing.

The flip side of autonomous cars is less autonomy for you.

* Target/Home Depot hacks were nothing.

You don’t understand how laughably outdated most security is in retailers. There just isn’t much emphasis on fixing it because banks/credit cards will cover the costs, and they can cancel card #s in an instant.

There was nothing technically impressive about those hacks.

How often do you hear of hackers hacking mobile signals (e.g. LTE) and getting internet for free?

Never.

How often do banks lose money to hackers, excluding social engineering (getting password from someone?)

Never.

Everyone has this grand image of hackers in their head, but you can’t actually hack anything properly encrypted. 99.9% of hacking is social engineering.

* If gasoline is $2 per gallon and I can work in my car, why wouldn’t I want a big cheap house in the exurbs instead of an small expensive apartment in the city?

Similarly, much of the movement toward the city is driven by drunk driving crackdowns, but if I can drink and have my Google car drive me home, why pay to live in the city?