Featured Toyota debuts its first SDV (software-defined vehicle)

Not necessarily "extremely long" periods of time.

Less than 100 years passed from the time Nebuchadnezzar II made Babylon a world power until it was conquered by the Medes and Persians under the rule of Nabonidus.

Persia remained a dominant power for only some 200 years.

Greece's expansion was made by the conquests of Alexander the Great who only ruled for 12 short years, from age 20 until his death at 32 years of age. After that, the empire was divided among his generals into 4 separate nations. The remaining Greek nations were then conquered by the Roman Empire some 130 years later.

Others did last quite a while. From the time the Roman Empire conquered the Greeks and Macedonians until it fell were around 650 years.

 

Yes, that is great news. But again, remember that all of this will cost a lot of money.

It's kind of like healthcare. It's better than ever, saving a lot more lives than ever. But I do know an increasing number of people who "just can't afford it." I just switched jobs, one major reason being that our premiums alone were starting to reach 1/3 of our income, and that was on the cheapest plan possible with a hefty deductible that all together wasn't that far from total yearly income, if ever needed.

Can I afford a new Toyota? Hands down no! I can't hardly afford a place to live in.

We need better traffic safety. But we also need to keep hard working citizens from facing poverty.

Personally, I don't see a human solution that fixes it all. Hopefully, over time, the SDV take-over might trickle down to the rest of Americans in the used car market, making traffic 99.9999% accident free. I just don't know if it'll happen quite like that.

 

In some ways, I see a software defined world as being an easy way to do the same kinds of things to customers.

You want to live longer?
You want better safety?

Then get ready to fork over the $$.

Hopefully I'm wrong. We also don't have the technology yet to say how much it costs compared to how well it works.

But like anything, it's driven by profits, not altruism.

 

Sure, no doubt about it. But there are other OSes used in automotive...like QNX and Green Hills. Successful and relatively small companies.

Mike

 

We are already much safer than 99.9999%. There is about 1.3 deaths per 100,000,000 miles. That is about 99.9999987% safe if my math is right.

Mike

 

99.9999987% safe compared to what? Do those deaths account for pedestrian accidents too?

The average American drives some 12,000 miles per year. If that's all he drives in 35 working years, then during those years the average American racks up 420,000 miles. If one dies every 76923077 miles, then every 183rd driver will die (assuming all drive those same amount of miles on average).

 

Despite the claims to the opposite from Apple and Microsoft, It does not require "a massive amount of long term investments" to maintain well written software. All it takes is one person who understands the code. I've maintained software for years ( around 21367 words, 5902 lines, 143814 characters) . We had a fair following among the home automation enthusiasts. After 25 years I finally released the code to the public domain because I could no longer follow the logic of the code that I'd written when I was a much younger man. Doctors blamed a stroke.

There were many reasons that I stopped maintaining the code. First was that the repository that was used for chatting was shut down.
Next was that the other person who had taken over maintaining the code had died. Then there was the advent of the Amazon Alexa.

I've had several programs that I created which lasted for decades. One of my first programs (written in the 1970s) was still in use when my company did it's Y2K audit. The program was NOT Y2K compliant. The person who was maintaining that system had to re-write the logic surrounding the truncation of the mm/dd/yy so that it was consistent. It took just a couple of afternoons to find and fix the Y2K bugs.

I don't code much any more. But it's not the cost of hiring people. It's the inability to recall the reasoning for laying out the code in the way that I used.

 

It's one thing for old code to work today. For an example, good ol' MS DOS can work natively on modern CPUs based on the x86 architecture with little or no modification.

But it's another animal when the internet and reasons to hack are involved.

An SDV will be connected to the internet and that connection will have a huge ability to control and manipulate the vehicle. There will be very good reasons (in the eyes of criminals) to hack such a connection in order to steal cars or wreak other sorts of havoc on society. This will mean there will be a need for constant security updated and patches.

Anyone want this kind of thing happening to cars all over the streets? And it wasn't even a planned attack.

 

I don't see the logic for the assertion that the internet and hacking means more maintenance costs. It might mean more opportunities for slovenly developers to write bad code, but it's not really a given that they HAVE to write bad code. Nor is it a given that they have a golden hall pass to skip over the need for vigorous examination of new and existing code.

 

Well, no offense, but you must not be familiar with IT security. This is the whole reason that phones, tablets, laptops, desktops and other computers receive security updates and why some software is also on a never-ending update cycle. This is why that if you go to any security IT class they'll tell you over and over again, "Keep your device updated!"

Most, if not all major tech companies pay professional (good) hackers to try to find vulnerabilities on a constant basis on their currently supported products, even if they're over a decade old. Then they try to patch those discovered vulnerabilities via updates. They also respond to successful malicious hacks and attempt to patch their products against those. With today's extremely complex tech, it is impossible to find every potential security flaw before the product launches, and this is one reason why so much software is updated constantly.

Add to that, that just as consumer tech keeps improving, so do the tools for hackers. Now hackers can use AI. Next thing you know they'll be able to use quantum computing. They are able to use faster computers that help them figure out how to hack some other device more quickly. And that's not mentioning all the other things hackers can get their hands on. The more they hack, the more they know, and the more they know, the more they can hack.

All of this means that your firmware, your OS or your app that you wrote today with no known vulnerability is going to have vulnerabilities that present themselves as time goes on. And eventually these vulnerabilities can and will get to the point that software updates alone won't fix them; you'll need new hardware to stay safe. Saying that a person can just write an app or operating system that will never need security patches is just wrong, unless the device has no connection to the internet. Then it still can be hacked, but is much less likely to be hacked.

You want to be a good tech company? Then pay professional hackers to continually help you find security vulnerabilities on your products and to rewrite the code to mitigate these vulnerabilities so you can push out security patches via updates. You may even need to install new hardware to keep up.

And like I said before, if there's a reason to hack, hackers will hack. Even on something as simple as a child's "smart watch" that is designed only to make phone calls to her parents wouldn't be hacked, but my niece's was, and now, if you try to call anyone on it, it goes straight to an adult hotline. There are many more reasons why a hacker would want to hack an SDV, especially a self-driving one.

 

I've been involved with localized IT security for a few years at this point in time. I've got my own view of it and experiences. They're not always spoken about in easy to understand languages and it's easy to get lost while learning about actual fraud and malfeasance. One can read about why the OS / App / Automated Tools - being used regularly, to find out how the security vulnerability was found, and how it hopes to close a vulnerability, like a zero day for instance.
Zero-day vulnerability - Wikipedia
I've seen a reference to a story about AI creating it's first zero day.
AI agents is another interesting subject in security circles, since AI is included to one degree or another now included in nearly every consumer grade tech device being sold.

Here's a video about driverless tech being tested on major hwy delivery routes at what some believe is an alarming timetable.
14 minutes - youtu.be/CQrQrOPmszE

 

link doesn't work

 

This is it:

 

As one who modified software (OS and Compilers) and ended up managing a secure operating system development group, the demands of security for a 6000 lines of code firmware item and for a Operating System complexity design code base are two entirely different things. Very few OSs are designed before the first line of code with security first as the number one aim. Because no hobbiest or purely commercial outfit would go to the lengths to prove every code path was secure, would subject their code to outside and government security agency audits. And endure the delay to market that hundreds of testing cycles for every change entailed.

Doing it right is really hard, really expensive and requires a small constantly communicating group who debate every design without ego. I was fortunate to know a few of those folks.