Live data from Hacker News

Farmers look for ways to circumvent tractor software locks

npr.org

121–130 of 359 posts

Re: Farmers look for ways to circumvent tractor software locks

#121
Imagine not so long in the future if self-driving cars were forced to reveal their code because of the right-to-repair bill, 1. who without a vast depth of knowledge in C++, etc. would be able to go anywhere near it? 2. even if they did is it in their best interest or anyone else's if they tinker with the code and made the car take undesirable actions?

Maybe buying a tractor should be replaced with leasing tractors, if they never want you to fully own everything in it. I think very soon there will be more and more of a need for a new way to determine what products are allowed to be sold partially with a secret OEM key.

Re: Farmers look for ways to circumvent tractor software locks

#122

Hey all, I worked on software for John Deere. This is a throwaway account for obvious reasons. Opinions expressed here are MY OWN. I no longer work for John Deere or am associated with them in any way. I was part of one of the many teams that work on this software. Specifically I was part of John Deere's ISG division also known as the Intelligent Solutions Group. The ISG division (was at the time) responsible for tyi…

edit: I have hopefully fixed all of the places where I incorrectly parsed "worked on software for" as indicating employment.

> They have a real potential to end peoples lives in the event of failure

I hope the lessons[1] learned from the Therac-25 are part of the design. Anything that dangerous needs to have defense in depth. If there are hardware limiters, interlocks, etc adding redundancy to the software check, the risk should be no worse than repairing any other dangerous machinery. Alternatively, if John Deere is relying on the software alone for safety checks - which requires that they (and the OEMs!) write bug-free software - then I don't want to be anywhere near the affected products.

> these tractors do (in testing) fail in spectacular ways.

So it's software only. I'm amazed at the hubris of thinking it's even possible can guarantee a software project large enough to requires a couple hundred engineers that maintain "the largest, most complex codebase [you] have ever interacted with". The very existence of tests failing in "spectacular ways" is evidence that bugs do exist in the software.

> These are not riding lawnmowers.

I'm not aware of anybody that suggested they were.

> They are 30-ton combines, and 20 ton tractors tilling fields

Yes, just like the older combines and tractors that didn't require complicated software.

> Joe Farmer isn't going to have a fucking clue how to repair their software correctly.

This, really, is the key point: the farmers in question *are not trying to repair the software. The only reason it is involved at all is because John Deere the company) chose to add the dependency.

[1] http://sunnyday.mit.edu/papers/therac.pdf

Re: Farmers look for ways to circumvent tractor software locks

#123
post #89
post #71

Earlier quoted context omitted.

"This has pretty frightening implications for the state of consumer goods, if industrial goods are already going this direction." This scenario has played out thousands of times in the manufacturing/industrial sector over the past 30 years. The end result is that the vendor trying to lock their customers in to a service contract loses out to whoever makes a remotely comparable product that isn't locked in within a de…

Except the market can stay irrational longer than you can stay solvent. The issue is that any open version of a product has to compete in such a way that matches the buy cycle of the product. I don't imagine farmers buy new combines every year! These things has to work for decades! Ditto with a lot of idustrial machinery.

> Except the market can stay irrational longer than you can stay solvent.

This applies solely to stock market speculation, due to a perfect amalgamation of human psychological biases and fallacies.

In consumer economics we use utility value (iirc, it's been far too long since I've opened a textbook). Consumers will definitely drop their JD's if a "no-bullshit," almost as good option comes along, because the utility value gained will be higher than the asset loss.

Re: Farmers look for ways to circumvent tractor software locks

#124

Hey all, I worked on software for John Deere. This is a throwaway account for obvious reasons. Opinions expressed here are MY OWN. I no longer work for John Deere or am associated with them in any way. I was part of one of the many teams that work on this software. Specifically I was part of John Deere's ISG division also known as the Intelligent Solutions Group. The ISG division (was at the time) responsible for tyi…

How many lines of code does it take to register as the largest you've dealt with? The largest I dealt with took 45 minutes to grab the repo. I don't state that as a point of honor, more like a point in needless over-effort.

I would say the JD's tractor software (if taken in it's entirety) is on par with the modern Linux kernel in size, if not larger. Worse yet, the codebase is spread along many different repositories, teams, and build systems.

I've seen integration builds take days to finish (not including running of unit tests).

Look I get it, John Deere is evil. But this codebase is beyond any one person or team's comprehension. It's very massive, and it's very complicated.

Re: Farmers look for ways to circumvent tractor software locks

#125
post #83
post #62

Earlier quoted context omitted.

> Also nobody's being forced to buy new tractors, there's tons of old tractors still out and running and working every day. There are also fewer every day.

You'd be surprised how long old, robust vehicles live for. I don't have any clear data for tractors, but older trucks that can be repaired with minimum tooling get used and re-used in places in Africa, almost to no end.

This is not so relevant for modern agriculture, but my family still has the Valmet 361D tractor my father bought in 1963. It runs as good as ever. The diesel engine needed to have the cylinder tubes replaced some 30 years ago, and other mechanical parts do wear out and are replaced, but this thing isn't going to be out of use because of software issues.

It's too small to be very useful in modern farming, though.

Re: Farmers look for ways to circumvent tractor software locks

#126
post #121

Imagine not so long in the future if self-driving cars were forced to reveal their code because of the right-to-repair bill, 1. who without a vast depth of knowledge in C++, etc. would be able to go anywhere near it? 2. even if they did is it in their best interest or anyone else's if they tinker with the code and made the car take undesirable actions? Maybe buying a tractor should be replaced with leasing tractors,…

Well, modification of certain software systems could be easily regulated. Even now you can't just modify all the mechanical parts of your car. For example, if you remove or bypass your catalytic converter, you can be fined $2,500-200,000 and charged with a felony. There's also a provision for any citizen to sue you into compliance if they know you've done it.

Doubt that happens much, but if you ever became someone the gov't doesn't like, they can probably just go see if your muffler is up to code. Or any of the million other felonies you might be unknowingly committing due to the expansion of federal regulations.

Re: Farmers look for ways to circumvent tractor software locks

#128
post #121

Imagine not so long in the future if self-driving cars were forced to reveal their code because of the right-to-repair bill, 1. who without a vast depth of knowledge in C++, etc. would be able to go anywhere near it? 2. even if they did is it in their best interest or anyone else's if they tinker with the code and made the car take undesirable actions? Maybe buying a tractor should be replaced with leasing tractors,…

They lock down the software in their well understood interests:

- vendor lock in (this is not good for customers)

- safety (testing this stuff is expesive and lengthy. Every change forces a re-test. They are liable for this stuff)

- defending the brand value (if there is an accident because of modified software non-hackers will blame the manufacturer for making it possible)

This all weill get worse with neural networks:

Imagine people poking around bits in the matrix representing the neural network of the sel-driving AI.

They will not have the slightest clue what they are doing, but they don't want their System to be locked down. They download "steering optimizers" and similar stuff from sites and overwrite the neural network.

Who will be to blame for the accident because of this?

With neural networks even the manufacturer cannot completely understand every nauron's effect on the complete system, as opposed to traditional software. Manufacturers have huge responsibility, and curently only locking software down is their only way to defend themselves.

Re: Farmers look for ways to circumvent tractor software locks

#129

Hey all, I worked on software for John Deere. This is a throwaway account for obvious reasons. Opinions expressed here are MY OWN. I no longer work for John Deere or am associated with them in any way. I was part of one of the many teams that work on this software. Specifically I was part of John Deere's ISG division also known as the Intelligent Solutions Group. The ISG division (was at the time) responsible for tyi…

> If a team of hundred of engineers struggle with their codebase internally, Joe Farmer isn't going to have a fucking clue how to repair their software correctly. > But trust me on this, locking this down is a very good idea. Joe Farmer may not know, but Joe farmer has "the internet", and ability to pay an independent software engineer. And people would absolutely find and fix bugs that john deer is missing. Honestly…

What happens if this independent fix results in a death? Do you really think JD will escape liability?

Re: Farmers look for ways to circumvent tractor software locks

#130
post #122

Hey all, I worked on software for John Deere. This is a throwaway account for obvious reasons. Opinions expressed here are MY OWN. I no longer work for John Deere or am associated with them in any way. I was part of one of the many teams that work on this software. Specifically I was part of John Deere's ISG division also known as the Intelligent Solutions Group. The ISG division (was at the time) responsible for tyi…

edit: I have hopefully fixed all of the places where I incorrectly parsed "worked on software for" as indicating employment. > They have a real potential to end peoples lives in the event of failure I hope the lessons[1] learned from the Therac-25 are part of the design. Anything that dangerous needs to have defense in depth. If there are hardware limiters, interlocks, etc adding redundancy to the software check, the…

There is definitely defense in depth but, I never said the software was bug-free. :)

Also you're doing the unfortunate thing of associating me as part of John Deere. I'm not. I had a small hand is writing a small part of this software, a long long time ago. I don't give a shit if they sell tractors or not. I don't care if farmers modify the software or not. I don't really have a dog in this fight. I'm just a guy with an opinion who's seen how the sausage is made.

Post reply on HN