Hacker News Reader

Comment by memetomancer | Hacker News Reader

This is a frustrating yet common sort of take. Yes, this is simple, as the article clearly points out. Yes it is obvious in retrospect. But did you do anything with your brilliant work besides bodge your terrible car a few more miles down the road?

There is value to developing the entire system... to ensuring the keypad mechanism is reasonably robust and tamper proof. There is value to understanding the vehicle as a system and reasoning out this defense strategy. There will be value in preliminary productization of something this for mass production, especially as regards the use of that terrible 12v power port and providing the 'fingerprint' in a safe range of voltage fluctuations to avoid catastrophic and probably non-obvious failure modes. There will likely be D.O.T. paperwork, and UL listing.

$1.2 million is probably a bit meager to truly develop something like this.

Yes, you can hobble some crap together on your Montero. Congratulation. Hardly a solid foundation to speak ill of this team doing something genuinely productive.

View on HN

It is obvious in retrospect because this concept has been around for 30 years. A common killswitch mechanism that I remember being implemented in the early 90's was a system that tied into accessory devices. On my friends car you had to put the key to ACC, then turn the cruise control on and off, and then engage and disengage the parking brake before the car would start. No other obvious lights, buttons, switches, etc. And you could install the killswitch device to tie into basically any 2 systems that used battery power.

And despite being around for 30+ years it’s not broadly adopted despite a significant rise in car thefts.

Is there actually a significant rise in car thefts? Or did we just hit an acute rise in car thefts of two particular models, caused by the discovery (Well, publication, really) that they are still using 30-year-old security?

Good question, it seems like car theft is on the rise, yes. And it’s spiking in particular metro areas (Milwaukee, Chicago) more than average, although national trends are also up.

https://counciloncj.org/wp-content/uploads/2023/01/CCJ%E2%80...

In California, if youre car is stolen and then found, the cops will give you a fat ticket and tow your car and then give you a ticket for it getting towed. Somehow getting the car towed is also a ticket.

Yep. Shameless re-victimization.

Moving illegally parked cars for the myriad reasons it is necessary is completely understandable. The fuckery, though, is beyond the pale. It should be a fine of the cost of doing business of moving your piece and not a penny more. Instead it's a racket

Another question you can ask is was there any incentive before to not report car thefts?

A salient issue has been that Hyundai/KIA didn't implement any anti-theft mechanisms on certain models, and recently the details about how to steal these cars has become popular knowledge, and now people who own the affected models can't even get insurance on them.

There's been some other exploits to infotainment systems, but AFAIK, they are all limited to proof of concepts. And the radio-repeater that almost works occasionally on some cars with wireless key access (better implementations have proximity detection which prevents this attack vector).

As it turns out, immobilizers are pretty damn effective.

If I owned an effected Hyundai/KIA, I'd do like we all did with 90s cars and put a killswitch in. It's not professional car thieves hitting the bulk of these cars, but mostly bored people showing of. So if YT can't show them what to do if the car won't start, they will go away.

In some countries, it is.

I would say because of how it must be installed, and that it is probably not common knowledge. In my country, it is not unheard of, but I hadn't heard of it until my electrician mentioned seeing one on a car he worked on recently. I asked if he can install one for me, and he said he doesn't know how, nor did he know the name of the person that installed the one on the other car.

> On my friends car you had to put the key to ACC, then turn the cruise control on and off, and then engage and disengage the parking brake before the car would start.

That’s a cute trick, but if a current day equivalent is integrated into modern day cars (i.e. CANBUS-based), then the security is already defeated.

No one challenged the security of the “cruise control cheat code” of the 1990s simply because there were no devices small enough. The other bit is that criminals weren’t sophisticated enough.

My father disconnected the distributor (correct word?) and took a piece of it with him. Definitely a killswitch. That was in the 70s. Cars got more complicated around 1980.

Yep, removing the rotor from under the distributor cap (correct), which disabled the ignition system.

Mine removed the starter relay at night. (The car didn't have a distributor AFAIK, or if it did, it wasn't as accessible).

This wasn't to foil thieves, it was to frustrate the repo man.

My hack was to purchase a 2017 vehicle with a 5 speed manual transmission. I reckon 95% of would be thieves can’t easily drive it away!

my trick too, esp in some areas. getting harder and harder to find outside of sports cars :-(

> My father disconnected the distributor (correct word?) and took a piece of it with him.

Yes, my dad was used to removing the rotor from the distributor (small piece, easy to pop off and unless the thief just happens to have the correct model handy, the car can't run) back in the 60s (maybe he did it earlier).

I'm pretty sure some form of this has been popular for just about as long cars have had an electrical system.

Yeah, but he did he get a 1.2 million dollar grant? Yeah, I didn’t think so. This guys kid is gonna have a better story.

(/s)

Dropbox wasn’t a new idea either. rsync had existed for many years.

But Dropbox made rsync more user friendly and available to people who weren't techies.

The concept of a starter interrupter has been around almost as long as the automobile itself. Ways to engage and disengage that interrupter have evolved and advanced over the years. Older folks will remember cars with a keyswitch on the front fender, and then a keypad inside, and then hidden switches like I described in my OP, and then IR and RF remotes, and so forth.

The basic concept in the linked article is not very novel, IMO. The specific implementation is cute, and somewhat current in the sense of evolution of these systems. But the whole thing is as noteworthy as the next arm64 advancement.

"i'm not a 'car guy' - where can i get a simple interface for a killswitch that only needs to be installed once and can be controlled from an app?"

^ this is where the value is, which is what the $1.2m is intended to explore.

30 seconds on aliexpress found me this: https://a.aliexpress.com/_mPSrPR0

No Bluetooth, but it does have a dedicated RF remote.

Edit: figured out the keywords I needed for the exact product you want “bluetooth immobilizer”

https://a.aliexpress.com/_mNgigFk

I can get you like, 10,000 AliExpress links for 1.2 mil, but the paper would probably suck.

I love rsync. Used it to batch some few million image files for Sephora makeup company to a couple different servers.

So does this “project”.

There are dozens of already existing products that are designed to do exactly this for pretty cheap.

https://a.aliexpress.com/_mPSrPR0

No, really, good for them.

But don’t be talking down on my car. That’s just not cool.

That beast is the workhorse of the farm and it gets the job done.

None of the windows roll down though and it’s hot as hell inside, so it discourages unnecessary use, saving the planet.

It rarely sees pavement but it drags what needs to be dragged and it pulls the utility rigs out after they deliver to us.

And that's all well and good, but maybe introspect here for a second? You're upset that your accomplishments aren't being respected, immediately after discounting the accomplishments of others.

The point is that your car's modifications and the university's are similar, but different, particularly in scale and broad robustness, which adds difficulty in ways you may not be appreciating.

$1.2 million may sound like a lot to you, but to pay a team of people to work on, and provide materials for them to work with (especially cars, which generally aren't cheap, especially used cars right now!)... Well, it likely doesn't go as far as you think it does.

The professor did gloss over briefly the difficulty in making the system work for a large number of vehicles, before arriving at a viable "signature" idea, as the article describes. Sounds like an area with a lot of false starts (heh) and time consumption, and dead ends.

The professor should have seen that he could send a signature over the airwaves to his relay since that is even more universally compatible… plus, you can buy that exact device for about $20 at the online retailer of your choice.

Vectorising the power profiles makes this a no brainier. I’ve done it, and I have no brain.

$1.2 million will fund 4 years of research for 2 professors, and 2 PhD students. It’s not exactly a career making grant.

That seems extravagant. By my calculations it should fund approximately 6 or so years.

Cover the two PhD students at the NIH payscales for PhD students on a standard training grant[1] ($43,894 not including benefits) and you've used up over a quarter of your budget on less than half the salary needs, completely ignoring any research costs that need to be covered on top of the much higher payscales of the professors. Plus a large number of PhD students in this kind of work make more than the states stipend above. Not extravagant.

https://osr.ucsf.edu/news/nih-update-ruth-l-kirschstein-nati....

Where are you seeing $44k? The link you gave shows payscales for postdocs, and points to another page [1] showing that predoctoral trainees get $27k.

Also, in my field and in my region, $27k is massive funding. I don't know anybody who makes that much, let alone $44k, and we also don't get tuition or benefits covered. Our TA/RA union is currently striking because it's essentially impossible to live off of funding alone.

[1] https://grants.nih.gov/grants/guide/notice-files/NOT-OD-23-0...

This was my conclusion as well. Will wait to see what the response is here.

I'll give you that I misread bullet 2, so the total is a little over 31k. But grants that fund salaries for predoctoral scholars don't just fund the salary itself, they also cover the additional funds listed on that page. You can't partially fund a trainee on a grant. In any case, this wildly misses the forest for the trees - 1.2 mil in grants does not cover 6 years of salary plus research costs for 2 trainees and 2 professors full stop.

But it’s not 4 years either.

Absolutely, 6 years with a 1.2mil grant is ridiculous. I was just hopeful that somewhere there were PhD students making enough money to live from research

No way. Half goes to overhead. 600k/4 people/4years = 38k/person/year.

Did you see the picture of what they built? I wouldn’t describe it as refined or particularly professional.

It sounds like his system is more refined than the academic one. It certainly has more features.

I’m not disparaging their work. It is probably really cool, and they probably published some great information that will be useful to many. I don’t doubt it was challenging for them, but I do doubt that the problem was fundamentally challenging from en engineering perspective.

As for my “work” it is literally insignificant tinkering by a bored old fucker with nothing better to do than chat on hacker news.. I don’t even respect my work, and anyone who thinks more of it than digging a ditch is just wrong and has obviously never dug a ditch.

But, just calling it like it is, the “signature “ thing they are working on is something that is already solved for decades and if it took anyone more than a week they may not have a clue what they are doing. I have implemented a version of it myself in a technically adjacent application.

In case anyone cares enough - and you probably shouldn’t- feel free to read my incoherent ranting that follows:

In my case I use load vector analysis it to detect and characterise loads on our microgrid. We have several buildings and houses, and we run 100 percent solar on an off grid system.

Using an esp32 and a current transformer coil on each of the three phases, with some good 16 bit ADCs, we monitor and characterise loads. Each of the refrigeration compressors has a somewhat unique starting and load profile. Each water pump in our utility system similarly has a unique startup and load profile. Same with air compressors, fans, and other equipment.

The profiles are programmed into the esp32 by putting it in calibration mode and switching the load off and on 10 times. It’s a pain in the ass because you have make sure no big changes happen in the power system in the meantime, but it works.

The MCU saves the signature as a vector and assigns it a number if it doesn’t sit too close to any existing vector signature.

It is really good actually, even being able to discriminate between identical pumps on the system because of their supply impedance and loading.

I’m not a data scientist or an actual engineer so I adapted some vector code from a DSP project, and the whole thing took me about 2 days using the Arduino IDE (please kill me)

I’m basically an idiot. Anyone who does this for a living should be able to do it in less than half the time.

There are still some rare false negatives because a grid can be quite chaotic, but in general it’s very accurate. In a simple D.C. system like a car in the off condition with predictable loads I would fully expect 4 nines discrimination.

What they did was cool, but it wasn’t hard. Not saying it wasn’t hard for them, and maybe they learned a lot, but I’m pretty sure that 1.2 million to solve the problems described in the article is two orders of magnitude off of reasonable.

From the provided description, If a single engineer with decent tools could not have this from zero to a production ready GERBER file with masks, stencils, and the works to send off for automatic assembly inside of a month they should probably look for another line of work.

Of course, if they work like I do which is to say they don’t, very much, and they mostly drink coffee and fuck off all day, then I’d give them a month and a half knowing full well they did all of the actual work in a week of panicked thrashing, creating months of technical debt in every line of code to build the glass house that somehow works without passing any of the tests but that’s fine you just rewrite the tests.

Of course certification and things like that are a whole different beast, but this was a CORE research grant.

Given that it's an idea that has been in production vehicles for 40 years, I doubt you'd need to spend $1.2M to "develop" it.

1 line barely acknowledging the criticism, 4 lines defending the car whose feelings I can assume have been mortally wounded. The defensiveness around the car is ironic given how casually you threw out your needlessly negative hot-take.

Nah, you can’t hurt pure evil. It just sits there, awaiting its next victim.

I hate that beast, but it’s my beast to hate.

You can’t just talk shit about it from your comfy chair, or sitting on the toilet with no circulation to your feet, or whatever — that’s something you earn.

You earn it with mild first degree burns on your right leg and tinnitus like the rest of us.

If I seem abrasive and unnecessarily combative, it’s probably just the incessant itching of my leg and the trauma from driving that thing.

He sounds like a monster of a vehicle. Loyal and strong. The goodest of cars.

There are dozens of similar mechanisms for sale on Amazon/aliexpress. A car alarm with an immobilizer is more advanced than this “innovation”.

They are claiming that the novel part is using voltage fluctuations to unarm the immobilizer and claiming that it requires less installation since the signaler device can plug directly into the cigarette outlet. A wireless relay requires the same cuttoff relay installation as their “new” idea, but is even more convenient because you don’t have to install a bodged together keypad on the cigarette lighter, and short your electrical system to cause voltage fluctuations.

They have blown through 1.2mm in grant money and their product is a bunch of prototype parts from a $50 arduino starter kit. It isn’t polished, it isn’t ready for consumers, it is a single prototype.

I guess the idea of causing voltage fluctuations is novel, but they sort of reinvented a $30 wheel for 1.2 million.

I work for a major OEM in automotive. Getting ANYTHING “simple” into real cars, especially anything related to physical access and starting the vehicle, is a huge undertaking. $1.2 mm is cheap for this sort of feature, assuming that money goes to the actual implementation, standardization, homologation, and integration on the assembly line.

Car OEM are far as example of efficient work…

It’s not efficient. That’s the point.

The car is a complicated product. It’s not a website. It’s not an app. My employer has 120k+ employees and factories in every continent except Antarctica. Regulatory bodies interject with anything related to access and security, and those bodies are different in every country/region. The product itself is massive physical good that many countries consider domestic production of which to be a matter of national security. Every single physical change to the product is analyzed by bean counters. Shipping the product requires at least some level of expertise in mechanical engineering, chemical engineering, hardware, software, and manufacturing. You need factories, regulatory approval, supplier networks, programmers, drivetrain engineers, management, people to lobby the government, accountants, and much more. You need it all.

You’d be shocked at how difficult adding a single physical button to any given car can be. Scoffing at $1.2mm for a new ECU that relates to security is naive. “I could do this in one day in my garage” is not how shipping a change to automotive products works.

> $1.2 mm is cheap for this sort of feature, assuming that money goes to the actual implementation, standardization, homologation, and integration on the assembly line.

Maybe I'm reading it wrong, but it sounded like the $1.2mm went to some prototypes and a research paper.

Well if that’s the case then it is indeed a rip off for the taxpayer.

They already spent $1.2mm. They have a prototype hand wired together. This isn’t even close to production ready, and it never will go into production because almost every new vehicle has an immobilizer built in that is authenticated via an nfc chip in the key that does exactly what this does, but transparently without driver input.

There is zero novel research here, and the entire purpose of the 1.2 million dollar grant was research. All the value you are mentioning is related to bringing a product to market, which is something that the grant did not require and universities don't usually follow through. Most of the time transition to industry happens is when there are motivated companies who do all the work to bring the device to market, but need university patent licenses and expertise to do so. This would be a great senior project, but it is a complete waste of money for a cyber security grant.

I disagree: the device monitors battery fluctuations to 'authenticate' the driver. the fluctuations need to be a specific pattern - delivered either by a device plugged into the 12v accessory port, or by some specific pattern of driver behavior, such as quickly flashing lights, activating wipers, etc. This is indeed a novel approach.

And it is a fair sight more involved than a simple kill switch, by the look of things. The research aspect comes from exploring the practicality of such an approach. This exploration requires prototypes, test beds and investigators.

Who's really to say what the results of the research will be, at this point? In my opinion, I think smart phones and NFC are probably the way to go... but I'm not going to hop on the internet and make scornful remarks until I know more. I'm not sure why you have done so?

That's the problem with the whole concept. Anybody can build a shockingly simple kill switch for $5 and a 5 video on YouTube. What are they trying to bring to market exactly? Cars have been around for 100 years and there have been hundreds if not thousands of these things brought to production during that time. They're all junk, they all fail and flop.

And great if you thing that those voltage fluctuations are gonna be consistent. Eventually some switch will corrode and then the person's wiper switch won't fluctuate the voltage properly. Nobody will want to reset their clocks using this every time they get in the car. Your break-in alarm won't work with the battery disconnectred. Car manufacturers will be pissed that you're disconnecting the battery because they can't get your telemetry and the car can't update while you're not int it. And then when you have problems, this will be the first thing ripped out of the car by your mechanic. This whole concept is flawed, and anybody with basic car or electronics knowledge will stay away from this thing because they can do it themselves.

And here's the kicker... anybody who doesn't have basic knowledge won't be hooking this thing up to their battery. They are terrified of even touching the battery. Congratulations on your marketing BS, but it's clearly not thought out from a common sense perspective at all.

If you read TFA then you'd realize they've solved almost every issue you throw down. They allow enough current through to power electronics (like your break-in alarm) but not enough to turn the engine over.

The target market for this is not "anyone with basic car or electronics knowledge who can do it themselves"... it's, "people who want an extra level of defense against car thieves".

This is the same kind of take as going to a nice restaurant and loudly exclaiming "$50 for a steak?! I could pay $8 at the butcher and make the same thing at home!"

It's more like a researcher getting a $1M grant to study whether putting salt on a steak makes it taste better, and a chef saying "Wtf, we've been doing this forever"

To be fair, most universities are great at interesting research but are also terrible at even preliminary productization. I highly doubt this $1.2M will go towards DOT paperwork and UL listing. This will go to a research prototype, then either get dropped off at the tech transfer IP office (good luck there), or spin out a startup. In the latter case, I'd have much rather seen this grant go directly to the startup, than pay the high Uni overhead.

$1.2 million sounds like a lot but there is a team of people working on it for a whole year. There is some insurance OP doesn't have in case it's proven one of these devices did cause a crash. If this was some Kickstarter I feel like it would cost more and be 3 years behind already.

> There is value to developing the entire system... to ensuring the keypad mechanism is reasonably robust and tamper proof. There is value to understanding the vehicle as a system and reasoning out this defense strategy. There will be value in preliminary productization of something this for mass production,

https://a.allegroimg.com/original/03e206/1de3f26447d79428246...

Optional extra on Series 1 Citroën XMs, an immobiliser keypad programmed into the engine ECU. It cost about 100 quid in 1990 money, on a 40 grand luxury car. Most V6es and 2.5 diesels had them, few 4-cyl petrols or 2.1 diesels had them.

There's no need to spend $1.2M developing something that's already existed for a long time. This was actually a development of a similar keypad fitted to most Citroën CX Turbos, from the mid-1980s. The idea is nearly 40 years old.

It's inexpensive proven technology, and it works well.

Reply on news.ycombinator.com