Featured Motortrend: Prime vs Volt (and other alt. vehicle comparisons)

Yes, I'm 60+ and spent my career in the computer industry. From experience I know that AI and especially autonomous cars are still in their infancy and I believe that neither is sufficiently mature and robust on which to trust millions of peoples lives.

 

Anything can be hacked. Given a brilliant mind and no constraints on time and money, my prior experience says you can never be 100% sure. Automated verification, 3-letter agency certification, multi-tens of millions and 10 years to develop can help but perfect?

I watched a 3-letter agency get very excited at what a single high school kid did over a single weekend. I followed the kid's career long enough to learn that some amazing hacks are possible on darn near every bit of code ever written or hardware for that matter.

 

If you guys don't think autonomous vehicles will be here in a very large scale very soon then you are smoking some old man weed.

 

I don't always drive to go someplace. Quite often, I go someplace to drive. Although it's less and less pleasant as the intelligence density gets less and less.

As for AI, there are a few others concerned. Stephen Hawking, Elon Musk, and Bill Gates Warn About Artificial Intelligence | | Observer

 

Alright...I'm done with this conversation. I feel like I went back in a time machine and I'm trying to explain a smartphone to someone from the 1930's.

 

I don't know as I'd say anything can be hacked -- maybe anything existing can be hacked..

If you give me physical access to a machine, it would be easy to hack, and I'm a bit worried about CANBUS. It's kind of an old standard that really doesn't cover security that well. (at least as far as I know.) So a bluetooth CANBUS is "worrisome" to me. For remote machines, it's getting harder to hack, firewall tech is getting better as well as knowledge of how to use them properly, and enough so that a lot of the "hacks" these days are phishing attacks or browser exploits. (they rely on the stupidity of users to infect a machine) This wannacry thing wouldn't have been an issue with existing tech if people actually updated their OS's every once in awhile and block SMB at the edge.

Now, a new advanced AI in totally autonomous vehicles, I would *expect* that part of the AI's tasks would be to detect intrusion attempts and to execute pre programmed responses of what to do with the car. Just pulling over and stopping wouldn't be good.

 

I had a "bag phone" back then too, didn't need it much, but when I did, it was the best thing ever -- up to that point in time anyway!

Anyway, I've actually found this discussion interesting. I wonder if all us Prime people have "range anxiety" when thinking about full EV's.

 

Not scared as much as burned once too often by over promised and under delivered.

My family lives in Oklahoma and a trip in September, 2016 convinced me the charging network between Huntsville AL and Stillwater OK was (is) not practical.

Bob Wilson

 

A major problem with "smart" and self driving cars is there is no security planned in.
The US government timetable for vehicle-to-vehicle communications standard and required implementation leaves no time to design security into the system. It cannot be added on later.

 

Bob Wilson

 

where can i get some of this, i am desperate.

 

Road trip: Denver.

Bob Wilson

 

i always wanted one of those watches, btw.

 

We are trying to turn the car industry into the drug industry.

Autonomous will be forced down our gullets whether they are ready or not because top men demand it.

Truth is the legal framework, standardization, and testing protocols are non existent , I have worked with "autonomous " vehicles and they are laughably sad in many situations.

On the road there are a multitude of situations they either can over react to, causing them to drive slow and irrationally which makes other drivers become impatient thereby causing accidents (But it won't be the auto cars fault it delays or stops irrationally) or we can take the Tesla approach where the car under reacts in some situations also causing accidents.

In the city in tight allies the autonomous cars are useless.

Autonomous cars also can become crippled in winter weather where a human does fine.

Infancy is a good way to put it get ready for thousands of Nancy grandma cars causing traffic chaos because we don't want to take the 20 years to properly vet and test the damn things.

Reality is we don't even know how to test them correctly yet.

So just because we can doesn't mean we should.

 

There are known good security principles & testing protocols but they are not easy, convenient or without additional cost to implement. We know how to test for security holes. There is no test to guarantee a system is secure, but that should not stop fixing the testable security flaws.
Hackers will likely have a field day.