Why Hundreds of Mazdas Tuned to 94.9 Broke Simultaneously
ฝัง
- เผยแพร่เมื่อ 1 มี.ค. 2022
- Start learning complex topics simple for 20% off by being one of the first 200 to sign up at Brilliant.org/HAI
Get a Half as Interesting t-shirt: standard.tv/collections/half-...
Suggest a video and get a free t-shirt if we use it: halfasinteresting.com/suggest
Follow Sam from Half as Interesting on Instagram: / sam.from.wendover
Follow Half as Interesting on Twitter: / halfinteresting
Discuss this video on Reddit: / halfasinteresting
Video written by Ben Doyle
Check out my other channel: / wendoverproductions
i live in Seattle, and would like to point out that while all of this was happening they kept joking about holding the Mazda drivers hostage. "thanks for tuning into KUOW, unless you're in a mazda, because you have no choice". good times
Thanks for sharing, thats actually hilarious.
LOL! 😂😂😂
It's fun when they can have a sense of Humor about it. It would be funnier if Mazda like funded a show on the station as well.
Well, It's better than most of your listeners being in a Subaru. I read a joke a guy had, Whenever a Subaru came into his shop, He'd bet on the radio being set to NPR.
Edit: I got a job at a Subaru Dealership, and 50% (Including me) of our customers have it on in their cars.
I mean that's funny from where I am sitting bit from the drivers perspective I'd imagine they would be prettu fucking pissed atbthe 1500 dollars in damages these jackasses did to there cars. Especially when it's gonna take a legal case and some luck to get them to pay for it. And one the west coast I'm not holding my breath those people will get there conpensation.
Cellist here: Sometimes we still get hired to play in dashboards, but the number of folks willing to pay such a high rate gets smaller every year. Support your local arts.
Also, that's a double bass being played with a German bow, not a cello with the standard French bow. I expect to see this in your next corrections video.
WARNING I am the unprettiest human YTer worldwide. Take the hint, dear omega
I think everyone is deeply disturbed by Sam mistaken the cello. I'm kind of surprised to see #CancelSam isn't yet trending.
Why should I listen to someone who swears allegiance to Perfect Cell
DEUTSCHLAND🇩🇪🇩🇪🇩🇪 > “frankreich”
You have no idea how hard I've tried to get 2 Cellos to play Thunderstruck in my dash!!
This issue was 100% preventable, just by following LONG established programming practices and having someone review the design and code.
I'm not a programmer, but I'm thinking just feeding the program some sample library of characters and symbols or some randomly generated strings should have identified the error and that doesn't seem too complicated or time consuming.
@@creativedesignation7880 That's one legitimate testing method. Probably would not have caught this specific issue, but what you describe is known as "fuzzing". Common hacking technique
Car programmer here. We do have code reviews. Too bad that the codebase is too huge and complex for any reviewer to be aware of every possible way a change will impact the code. And as for testing... Sure, we do that. Too bad that our Jenkins is so unstable that we can never know if a test failed because of a bug or because someone sneezed in the office. Well, better to get false fails than false passes I guess...
@@ishashka So these cars costs tens of thousands of €/$/£, and you get the cheapest possible build server? Not even GitLab? And if one change affects so much other code, you're writing it the wrong way. Compartmentalisation is. key.
Your comment with the percentage sign just bricked all Mazdas. 😃
This reminds me of a case in Finland years back when some TV towers, due to configuration mistake, broadcasted completely corrupted data, essentially random data, for some time (I think it was a few hours). I've forgot details, but basically there was one specific model of set-top-box that apparently did a poor job with checksums or something, and at some point the random data happened to make up a sequence that the set-top-box took as a command to get a firmware update that would follow. That didn't end well.
Updating wirelessly from a tower sounds like a bad idea nonetheless
@@tubbunnyWhy? Firmware is easy to send when there are spare bits because the video is freeze frame for some reason, or otherwise low-bitrate like rolling credits.
And using USB sticks not only requires having USB on the box, but also is far more difficult for the user than to just "keep the box on for 24 continuos hours to receive your update", with a "update available. Install now?" notification appearing once the update has been received.
I should clarify, that it's a bad idea while moving. If your car is stationary, and not installing the update ONLY after the download is successful (with cert checking or something to confirm nothing is incorrect), everything should work perfectly fine.
I'm saying that it's not foolproof. Updating via USB isn't foolproof either. No solution is perfect.
@@namibjDerEchte
@@tubbunny As has been shown countless times in studies and experiments: End users almost never voluntarily choose to update their devices unless absolutely necessary. They simply do not want to interrupt their service, even if for a moment. It's why you may notice that software running Windows or iOS have gotten increasingly aggressive in forcing you to choose a time to schedule an update. This would be fine in the past, as updates usually weren't that intrusive, but IoT devices have constantly been used in botnet DDoS attacks, so keeping your security up to date is a top priority.
It's not a terrible idea to autonomously update a system wirelessly, especially for what I'm assuming are just drivers for a TV box. I don't know much about what caused this issue, but if your device is attempting to execute a nonexistent firmware update there should probably have been a few more steps to actually verify the validity of the request and the package that is being pulled. If your TV box isn't making those kinds of decisions, and just blindly accepts any firmware update request without doing some sanity checks, then that is an awful security problem.
I think updating your devices to keep them up to date is a good idea @@BetaDude40, I just think that updating over the air might cause some issues. I'm all for security updates and even some cool new features, and I can definitely wait, but updating over a TV tower, without even verifying that the data is legitimate and correct, is a bad idea.
I've made a few coding mistakes in my time as a dev but "hardware destroyed due to unexpected input" isn't one of them.
I recall it was possible to kill the monitor of some early IBM PCs if you were programming the graphics card directly and wrote the wrong value to the wrong register.
To be fair, it's nearly impossible to damage PC hardware by a programing mistake. On a micro controller you're often able to physically destroy something by setting the pin state wrong.
Surprisingly easy to do when prototyping...
(Not that this radio issue sounds like anything more than a soft lock due to utter incompetence and lack of a factory reset button.)
@@weakspirit_ I've carefully avoided any coding job where safety is a factor!
@@TheVideoGuardian thats what im saying, it blows my mind that these systems dont have a safe boot/admin environment you can boot into and just clear cache, and do basically whatever, should have a boot mode that only starts the required processes. also, i dont see how you cant just re-flash firmware...... sure maybe not end-user but a dealership surely can...
Same. It's amazing to me that a venerable car company would be so trusting with it's digital inputs. I also think it's crazy that they don't have hard-reset function. Whenever I code things that talk to external sources, I'd ALWAYS think to add data validation and Incase I failed to vet the data correctly, some way to reset as a last resort.
That % formatting bug is a huge issue. Whatever programmer caused that absolutely should have known better.
Don't forget to sanitize your inputs, kiddos
I can't call a person who wrote that code a programmer. Sanitizing inputs is coding 101.
It's not like we just had several absolutely MAHOOSIVE security issues that affected almost everyone worldwide because of text formatting.
I posted a "how programmers sanitize inputs for people that are smarter than the average bear" on this thread.
@@TheMetaMoss alright guys, I got absolutely no idea about anything programming, but I'm curious, so what is sanitizing the code, and how would it have affected it?
Having worked in IT for almost 20 years, I can very confidently state that I *DO NOT* want my cars to have any type of connection to the outside world other than the OBD port and only the most limited computing power necessary.
While there are actual advantages to having certain systems being electronic vs purely mechanical, there are many systems that have absolutely no reason to be computerized and I feel like it's just a way to force owners to bring them in to the dealership because they can no longer diagnose or fix it themselves.
And a justification for charging thousands more than anymore needs to pay for a car.
It is also a way for manufacturers to institute control over what you can do with 'their' products.
Your car is not yours even though you own it, much like most PC software is not yours - Adobe (for example) just lets you pay to use it for a while.
Sadly this practice of turning everything into a service is only going to get worse, see for example the subscription to the heated seat or remote start 'service' you will need for some new cars.
Stop paying extra every month? Remote start stops working, in spite of still having everything you need for it.
Don't even get me started on Tesla and the battery life thing...
I wouldn't limit to the OBD port but they should all be wired connections anything can be in the ether. Ever heard of a spark transmitter?
i read an article a bit ago that said glass displays are cheaper than putting a bunch of knobs on the dashboard. think it also said it seems safer to fiddle a knob, than work a glass display. i don't know i haven't been in a car since they have been able to turn the headlights off when you leave them on.
I work in IT as well and I second this 100%
Oh you KNOW the auto industry is totally DROOLING over high tech's ability to sell expensive consumer products that turn into paper weights after about 6 years. Nothing they'd like more than a car that displays, "firmware not supported" after about half a decade or so. Planned obsolescence on steroids. And now bikes are following suit. Sad, sad, sad.
I got my start in IT then got into cars. I've been horrified watching technology creep into the automotive world--the sheer volume of computers and wiring stuffed into a modern car makes me sick to think about. That said, I can feel a shift coming--there's going to be a snapback against this kind of thing. Maybe it'll take some sort of Ford Pinto-sized disaster to do it, but all it's going to take is ONE carmaker to come along and introduce a line of "simple" vehicles that are fairly stripped-down but still reliable, well-built and pleasant to use. People might laugh at this now, but it's definitely going to happen. We've got "dumb phones" making somewhat of a comeback, why not "dumb cars" too?
Having worked in vehicle infotainment engineering, I find it amazing that worse doesn't happen more often. ROM images on these head units can take hours to compile, have software from tens of suppliers, and end-to-end testing is practically non-existent, let alone exhaustive suites of edge-cases like this.
I would have thought they used something like openbsd and open source code for most of it and wrote new features they want themselves
That explains why they're always such a clunky and miserable UX nightmare.
Doesnt make sense. U would think u have a dev environment to test first before u put it in production. Car manufacturers r really that incompetent huh?
@Out of Office having bad design that bricks the system over file extensions is NOT a ' edge case ' problem, it's a ' lazy design/engineer ' one.
@@0wl999especially with image files, as they basically all have magic at the beginning
Preventing this types of errors is pretty important, but being able to recover from them is far more important. This is why we have these hard to reach factory reset buttons.
This was the first thing on my mind! Having a reliable way to at the very least roll back to the default firmware & clear all forms of storage should be mandatory. I encountered a software bug last year which caused my phone to freeze in such a way that not even the power button worked. I had to wait a day for it to run out of battery and shut off, so I could charge it and turn it back on... Super annoying.
This isn’t 1988, you can’t just hit the reset button.
@@335m5 sure, you got to push it with needle or paperclip
That was my first thought too: they really should have included an option to wipe and reset the computer, so a bug doesn't irreparably destroy it. Of course, these days, it feels like wasteful practices of forcing consumers to buy new items instead of making it feasible to fix existing ones seems to be the norm.
Backups and failsafes can fail too. Should you.. make a failsafe for your failsafe? Shit's going to break sometimes.
I love this channel so much. The humor is *terrible*, which is actually great!
The family dog man!
So bad it's good
@@UTKETCHUP like me at life... Wait...
@@samn6498 ah yes
I feel like when someone says a joke is bad it’s actually good, because they would be not laughing and criticizing the jokes if it were actually bad.
Okay, correction to this article -- the Mazdas affected by the issue only needed to be factory reset using a USB drive with an update. They did NOT have to be totally replaced to fix what amounted to a software bug. Mazda had a solution ready within 3 weeks and dealers were deploying the fix fleet-wide that month. Most cars didn't see this bug ( obviously ) and had their infotainment systems fixed using an "OTA" ( over the air ) update for cars that had mobile connectivity -- or a USB based upgrade that could be downloaded.
Tech info:
HD Radio wasn't the first radio standard to be compatible with a digital download service.
FM radio had RDS back in it's later days.
I remember radio stations using it mainly for two purposes.
Firstly, to give the listener the station's name on their radios (applicable when the listener has a radio with a digital screen and RDS support).
Secondly, to allow the listener to jump between the station's broadcast towers if the listener is on the move and moving between ranges of the station's broadcast towers (Applicable, if the listener's radio has RDS and AF support).
Why "had"? RDS is still used widely today throughout Europe (and possibly the US, too).
Wasn't digital download, but yes. Also supports emergency broadcasts and traffic reports.
Was gonna make this exact comment. We have had RDS in Europe for decades. The audio isn’t digital but the data it sends is. RDS also supports automatically switching the radio over to a station when the news or traffic alerts come on. Also is the basis for TMC, the digital traffic system that sends out traffic updates to Satnav.
Believe the US never got any version of RDS for some reason.
Car companies: "We take cyber security very seriously. It's impossible for a hacker to hack your car... trust us"
Also car companies: "If you use a file without file extension it will literally brick the onboard computer"
It's a lot harder to hack a computer if it bricks itself first.
Also car companies: We don't want right to repair because Independent repair shops can hack your car
@@joshuacerniglia2501 often the goal of an attack is to limit availability, bricking the system definitely accomplishes that.
I mean bricking a system does kinda make it impossible to hack lmao
I guarantee you QA, if it existed, said "we notice that if the file has no extension the radio blows up" and some fuckface middle manager or dev went "you moron, nobody is ever going to do that, ship it already"
I had a colleague who worked at Mazda on the infotainment system, and I forwarded that podcast episode to him. He was thrilled since this sort of thing was his team's job and they had no idea.
Thrilled? He should have been deeply embarassed.
@@3rdalbum i believe this commenter was using the word thrilled in a sarcastic sense
@@3rdalbum I assume he didn't work on it since he had no idea, even then, I think it's really funny sometimes to see these things, unless you caused them, in which case it becomes embarrassing
So, did he come back with a solution, or are those radios truly bricked?
@@3rdalbum ESL here. Yes, he was kinda shocked, and probably embarrassed. I am sure his old teammates were too.
1:56 RDS, or Radio Data System which was developed in the 1980’s could send things like what song you’re listening to, or traffic information, or even sync your radio clock with the data received. So FM radio did have a way to send at least some type of data to your radio.
It still does and US FM stations also are using it. It is coded carrierless on the sideband around 57 kHz of the modulated audio signal (38kHz is the stereo difference signal).
“When criminal irresponsibility becomes an industry norm.” Really, we need to stop telling people that software engineering is easy and anyone can and should do it. It's as hard as any other engineering profession and should only be attempted by those who (a) actually know what they're doing and (b) understand and take seriously the many risks they're subjecting their users to. We honestly wouldn't suffer from having less, higher quality software in the world-software re-use is a virtue for this very reason.
I mean, there's a lot of money to be made in fixing other peoples' screwups.
No, I think the real problem is assuming that software alone can act as a fail-safe, which is impossible. If a computer has to always be powered on to keep a safety critical component in a safe state, then no amount of coding can prevent it from being fail Dangerous in the event of a total power loss, or a major short-circuit. The bottom line is that all safety critical components must be _mechanically_ fail-safe.
@@LordSandwichII I don't quite understand. Cars generally are certainly not “mechanically fail-safe”-many faults will see them plough into the side of the road at speed, and some will even see them explode. Cars with drivers in them are catastrophically dangerous. But we pass laws against extreme incompetence, iterate on failures, and hope that it's good enough. We should do the same with software engineering-as I say, this particular example, if the details are as described, is just absolutely unacceptable and would have been spotted by the most cursory code review, even if proper tools that would have prevented it mechanically were unavailable.
There's no such thing as perfect safety, but there is such thing as not being a total ass!
@@LordSandwichII personally, I liked it better when your car wasn't controlled by a CPU.
Its just another one of those things that cars don't need, like throttle governors and catalytic converters.
@@jakehildebrand1824 Well maybe you don't care about you children dying of some kind of horrible lung disease before their 12th birthday, but I do!
A random manager at Mazda: nah, this is too weird, it'll never happen.
The Engineers who found the bug and reported it: I told you so!
More like the QA team that found it says I told you so after having argued with the dev team telling them to fix it only for a manager to say that would never happen
The percent issue is because a lot of programming-languages use % for formatted text for example "%d" means "take the next parameter, and add it to this part of the text as a decimal number. If however there are no more parameters to use, things will crash. Because format-string vulnerabilities are so common, some tools have started to ban user-controlled formatted strings entirely.
log4shell was basically a fancier version the same thing
I'm quite surprised this was even an issue in the first place. Any common compiler from the past decade will warn against using any old string as a format string if you try.
No, the issue is that the guys who made this software were really shit at their jobs.
Actually, depending on the compiler, it will actually take whatever is in the stack and interpret it as the necessary data. This is a reality bad security exploit if programers allow users to make their own format strings. They can actually see data used in pass calls if they put in a random %s.
Like with SQL injection. The take home wisdom is that you _always_ sanitise inputs, and _never_ produce format strings from externally supplied strings.
Back in "the old days" you got in your car, turned a knob to turn the radio on (click) and then turned it further clockwise to adjust the volume to the desired level. You then twiddled the other knob until you found the radio station you wanted. And you could do all this usually without taking your eyes off the road and you got simple AM music.
Then clever technicians then started to outdo each other and made mechanical push button radios to store and change the station by pushing one of about five different buttons that you had to program in.
Then someone else decided to make the tone adjustable, there by making it even more complicated and tedious. Then came loudness adjustment and then multiple speakers and balance controls, left and right, front and back.
Then came FM radio, so more complications. Next was a cassette tape player included which regularly chewed up the new tape you had only just bought.
Then it was music CDs ; more complications!
Just when we thought it couldn't get any worse we had a second FM choice, FM1 or FM2.
And then it turned into a bloody computer that had to be programed for about 20 minutes and it all erased if the car battery was disconnected. Old people had to get their grandchildren to program these SOB devices.
So, I don't listen to the radio anymore, it's all too hard!
Oh for the "good old days" with two knobs when you could just go click it on and twiddle the other knob to tune it in!
On the bright side there hasn't been any decent commercial music produced for a couple decades now. Best you can hope for is a soft rock station playing stuff from the 90s.
If you're incapable of programming an FM radio, you're probably not capable of living independently and belong to an elderly home.
@@Eidako That's why all my vehicles still have cassette & CD players!
@@sunbeam8866 I thought I was the only one (with a CD player).
@@Support_Ad_Blocker How about a cassette player? 🙂
4:53 At least *she* is holding the soldering iron *the correct way* !
Good on you, HAI, for sourcing a real-life-compatible stock video clip! 😁😁
At 0:20 in the video, the stock clip is of a radio I spent 3 years helping make for GM as a consultant to a supplier. Fun fact: That radio was awesome for its time (2006~2013)... three models capable of RDS Text, talking to XM radio and OnStar modules and it handled all the various sounds your car needed to make (chimes, turn signals, etc...) The two upper tier models played MP3 tracks from CD-Rs, the top tier had a 6-disc changer built into the unit. Could a percent sign in a message mess up that radio? Possibly... I only had a direct hand in the radio's code for a small bit, most of my work was to support the radio from a development, QA, and service perspective. Anyway, that was my "cool story"
wow! small world.
That actually touches on one thing that occurred to me here wouldn't killing the infotainment system on a model that recent likely also kill features the car is legally required to have working eg turn signal sounds etc. I guess it could be that that there is a separate audio mixer that simply mixes those sounds in with the output from the infotainment systems DAC in hardware after the fact. But if it is relying on the software to do it surely this would have the potential to bugger up the roadworthiness of the vehicle according to the regulations in many countries. I think failure of the indicator (aka turn signal as you call it) sound in a car will fail the annual inspection (MOT) in the UK for example thus making it illegal to drive until fixed.
@@seraphina985 I can see sounds being an issue. The radio I helped with wasn't really otherwise critical to the operation of the vehicle, and ran on a "lower communication bus" of the vehicle - however, there was one morning I arrived at Warren Tech Center to see the entire Instrument Panel Cluster all messed up, and they were blaming our radio for the problem (spoiler: it wasn't our fault) - it was a network issue, however. In general, most of the controllers on these vehicle networks use standard libraries, and are tested and fairly robust.
I don't think audible signals are a requirement here in the US (or the Canadian market), but when present, I believe there were some specific standards for how they were supposed to sound.
There's nothing awesome about OnStar... or GM. That's not to say there arent awesome engineers at GM, just not sure why they'd work there.
And yet the Nav data is forever stuck at 2015, if you get the last disc....
"These sort of software bugs are going to have higher and higher stakes as cars become more and more computerised"
As a software developer, I'm keenly aware of this. The fact that some cars today are entirely incapable of retaining their basic functions of "go, turn and stop" if the one or more of the onboard computers aren't functioning correctly is an immediate no-buy from me. Limp modes are okay, they protect your car if something goes wrong but still let you keep control of it. But having no mechanical backup system that makes the electronics redundant is a huge issue of newer cars IMHO. One of the reasons my daily driver was made in '91 and is still my daily driver is that everything is mechanical. If something fails in the electronics, the worst that'll happen is the check engine light coming on and maybe you'll lose some of the electronic functionality. The car will still drive to your destination, albeit likely with reduced performance and efficiency. Cars today simply don't seem to be able to claim that same feat.
i think cars should use a lan like system for their systems.
one for safety and drive, one for infotainment, one for comfort, and one as a central hub/onboard black box so the systems can talk to each other (and not connect to each other directly)
and so there is a place were info is kept for every drive, and fault codes can be read.
that way a software bug in the infotainment system shouldn't effect driving (and cars shouldn't have their speeds displayed in the infotainment screen), all of these could be placed on one board.
all of these systems could easily be built into a single board (possibly with a few daughter boards for say electric motor speed control, or displaying an image on the speedometer.
@@Reverend_Salem agreed. Just keep everything separate instead of relying on each other to work
@@Reverend_Salem they basically do. Modern cars use the CAN bus, which stands for controller area network, and allows all of the systems to communicate together.
Thats why I dont like the idea of steer by wire and brake by wire
ur paranoid bro
Make cars more complex, add computers to them, wire all the systems together so that if a superfluous part fails it can damage the main function, then fight to end "right to repair". Modern cars are trash, get an old one. I can't wait for those remote kill switches to go live. Those will be abused.
Thanks so much for covering this! I heard about the news whilst conversing with some of radio station managers and I was wondering what the public's take would be. Truly a car that bricked because of a file extension is like maximum fail to me. And I like the HD IBOC.
I especially appreciated the five-second aside to shit on NFTs. It's like a national pastime at this point, I love it.
I assume you must mean the "laundering money thru pictures of monkeys." I didnt understand that reference so this helps.
If we could get NFT and EV shitposting merged together, we could probably organize a competitive sport
As a matter of principal there should be nothing you could transmit that would interfere with anything on your car. The fact that it was possible means that Mazda systems are probably faul of various certifications/compliances.
This radio station might have done it by accident, but anyone can transmit a signal; some might transmit something malicious on purpose.
No fault with the radio station, it all lays with Mazda.
Totally agree on that. It's some basic rule in computer science : never trust user input. Any fucking noob programmer learn it really, really quickly even if self-taught. The fact that it happened is the proof that either mazda hired someone totally incompetent, didn't bother to make security verifications or even that someone qualified this job to the "not my job" case. In any case, that's entirely mazda fault for not putting a proper way to "factory reset" their comp not checking user inputs correctly, and not commiting to safety measures, or giving the job to an apprentice and not checking it after.
That's even litterally the first security question "did you parse user input to prevent malicious exploitation, database dumping and shit". It is acceptable not to do it on developpment side and delaying it later to do it properly, but putting this in production like that, especially with cars, and no way to reset it is almost criminal. Hope their fired the ones who did this. This is not accpetable
@@feuby8480 And especially do not trust user input that's come via an unreliable channel like a radio transmission! Even if it's safe when it leaves the transmitter, it might not be after soaking up interference. Error correction codes exist for a reason, and even they fail regularly. Just look at all the blocky glitches in a digital TV broadcast.
This, exactly. I thought car companies had pretty high standards for code security and robustness, considering that it runs on things that weigh over a ton and move at high speeds through populated areas. Yet every two-bit web developer knows that user provided or external data should be treated as untrusted and potentially hostile (and if it comes in over noisy channels, it's also constant low-grade fuzzing). Failing to decode something that might have come in over radio is not an exceptional event, and should just be silently ignored, not panic the system and crash or brick anything.
There is a rule in computer security that no security system is uncrackable. There is a similar rule in software engineering that no software system is fully bug free. Sorry y'all, the trade-off for all the benefits computer software brings is that sometimes weird and bad stuff will happen. Some of you might remember the log4j security bug from late last year and how many problems that caused, or the various times CloudFlare/AWS went down and took half the internet down. As a seasoned software engineer... well, shit happens!
@@whong09 I agree with you : no security system is uncrackable. Bugs occurs, that's... An evidence. We are not talking about BSOD here, where you just have to restart the computer.
The fact that there is always a way to hack is the reason why we have standards, procedures, rules and certifications. And as I stated, the very fucking first rule is always "do not trust external inputs". You know why this is mazda problem ? Because ONLY mazda had problems with that. All others complied with basic rules.
I don't know if you are a programmer, but honnestly, your comparison is like saying "there will always be car accident, shit happens", look last week, 5000 people died in an accident. When we are especially saying that everyone should fucking put a seatbelt. That will not prevent everything. But sure, that will REALLY helps.
And moreover did you really watch the vid ? This was not a hack. This was not malicious. It was user (radio) error. How can you design such a bad computer system that an user error could completly destroy it, requiring hardware replacement ?
If mazda cars were hacked by some criminal discovering that and deciding to break shit, we would probably be here saying that basic rules were not followed, but like you stated, nothing is uncrackable, and shit happens. But that's not the case.
Had a similar experience with a model of BMW about 10 years ago with satellite radio. Sirius sent a signal patch out that created an error and stopped the tuner. It only effected people who were using the radio at the time it went out. And the only way to remedy it was to wait for the next patch update. I work at a dealer and there was nothing we could do to fix it.
This video makes me appreciate my '99 Buick Lesabre so much more than I already do...
"automatically downloaded and attempted to open those files"
jesus christ that is a terrifying sentence
“And much like the family dog, the only way to fix these terminally ill car computers was by shelling out for a new one”. That really got me more than any of his other jokes have
Made me angry.
@@veramae4098 it's not accurate. At least one dealership (likely all of them) were fixing the issue at no cost to the customer. The fix took about two weeks to manifest.
It did require replacing a $1500 "connectivity master unit", but again the story I saw said the dealership was replacing it for free.
@@Kinkajou1015 Sounds like a dealership that really values their customers and/or a dealer that's very sure Mazda will pony up and cover the costs.
They probably have a way to do a factory reload, but the time and coat to ship them back to Japan to do thar would be prohibitive.
Great Channel, interesting stories told in 5 minutes, not dragged out for a half hour.
"This was partially a result of how these cars were built, but it's also a result of how this radio station broadcast it's radio signals"
Dude, this was entirely on Mazda. "Never trust user input" is one of the fundamentals of coding.
If this obvious but accidental example was missed then how many deeper rooted issues do you think are just waiting to be exploited?
Honestly, HD radio sounds like a bad idea in the first place. What stops anybody transmitting data that your car automatically downloads and tries to open?
Wait, does that mean that instead of sending a .jpeg, .png or .gif, (or a file with a missing extension), someone could instead send an executable file and get the car to run arbitrary code?
it is probably more complicated than that. The OS of the infotainment system has to be able to get the file and execute it, which is (hopefully) not what it does as it is not expecting an executable in that stream of data. But bugs happen and I'm sure that's something that is gonna happen in the future, maybe.
Well, you can always use the OBD2 port to hack any car anyway and lots of other cars just entering the "smart" iot age had open wifi networks to control the car
Who knows? On the one hand unlikely as this was some graphic lib trying to interpret data which is not the same as executing that data as instructions, on the other hand this is really shoddy work so I wouldn't be surprised.
Here's how this can be fixed:
If the file isn't jpg, gif, png, or another image file type, or if the file extension is missing, immediately switch the extension to png.
@@_lunartemis or just do it the proper way and read the file headers :P
I'm surprised this didn't happen sooner. With 10,000 FM radio stations in the US (not sure how many are digital) a benign mistake like the station's was bound to happen.
10k is pretty old data, currently there's about 25k of them.
It's kind of sloppy that the data protocol is even dependent on file extensions.
@@romulusnr Hubble Quality Control: bows to Mazda
Clearly Mazda software devs were responsible this time. The worst result the "mistake" should have brought was that the metadata was not shown.
Not to mention what fringe areas do to data because of low signal strength and dropouts.
i live in Seattle and listen to 94.9 very often. I did a double take when i read the title and thumbnail lol. ya learn something new every day
I'm old school, I repair things that break. Instead of just throwing it away and spending another $1500 for a hard to find an original, try having someone trained in electronics ( like me, 40 plus years) actually finding the reset pin on the microprocessor and physically resetting the radio. DOH! We are such a throw away society.
Most users (billions of us) know that it's easy to simply turn off-on power or press a reset button as you wrote. The code creators however, don't seem to know this; for any appliance you have at home, the owners manual or on-line how-to guide show us to go through a complex series of instructions to reset your computer or TV or whatever. I think they are just arrogant with a attitude, "software can fix anything" so ingrained into their heads that they forget that simply resetting the damn thing is the easiest thing to do!
im pretty sure the issue is that the cars do not come with factory reset buttons. good luck figuring out how to manually delete those cached files by yourself.
As a CS student i think that the requirement to buy a new inbord computer is a scam. You should be able to reset the memory of the computer by cutting its power/deleting its memory.
in theory you should be able to, at worst you open it up and let a small voltage onto the memory reset line without powering the rest of the circuit, file gets garbled or totally erased and the error handler dumps it because its no longer a valid file instead of being valid but wrong
Even you need to change the computer it should be free because it's isn't user fault at the first place
That's what I thought, this sounded more like a software issue so flashing an updated firmware would fix it.
Yeah. If you can get the bad files out of the cache, then the problem will be resolved even if the vulnerability would remain. There has to be a way to wipe the electronics and load a fresh version of the firmware. Your typical mechanic probably can't figure it out, but a Mazada recall program sure could- and that's what they owe the victims
Sadly sometimes it is not that easy. The file is probably somewhere in a flash memory together with all the OS files, so it is nonvolatile, completely erasing the flash is not really an option due to the OS also being there, and selectively deleting the file is probably not possible due to the OS not having facilities to allow external access to the file system. If it was possible to put the infotainment system in upgrade mode a carefully crafted update package might have worked, but otherwise the only option is to probably unsolder the flash chip, reflash it with a clean OS, and solder it again. Not something you would do in a usual shop, so replacing is the only option (then the old one can get refurbished and installed in a new one etc)
I remember back in 2004 a local radio station broadcast a signal that would kill the batteries in Volkswagen Touregs. I worked at a vw dealer and we had dozens show up with dead batteries. Turns out the radio signals were making the radios stay on and kill the batteries.
This happened to my 2010 Chrysler minivan just last year. Radio wouldn't quit after the key was turned off. It was silent and dark but something was still ticking that drained the battery. Ended up having to replace the hardware for like $400. It's been good since then.
One of my first jobs in engineering back in the 1980s was in a system test facility where we got to hammer away on new software and/or firmware releases for a satellite communications system and its network control software. In addition to operating the system per the normal instructions, myself and some of the other folks would enter random things into text fields and press weird keyboard sequences and so on. We wrote more bug reports as a result than all the other groups combined :)
Reminds me of a joke.
A software tester walks into a bar.
Runs into a bar.
Crawls into a bar.
Skips into a bar.
Dances into a bar.
Climbs into a bar.
Flies into a bar.
He orders:
1 beer.
2 beers.
0 beers.
99,999 beers.
-1 beers.
qwertyuiop beers.
A lizard in a beer glass.
Testing complete.
An ordinary customer walks into a bar and asks where the bathroom is. The bar bursts into flames.
No matter what you do as a software tester, you will never out-test the incomprehensible array of random actions an ordinary person can come up with by accident.
I'm from Germany and during my uni time here I learned about how cars need to have certain fail-safes and risk assessments and need to fullfil certain coding guidelines and have fallbacks in place.
How on earth is this possible? Well I guess in Mazdas case they determined: "What's the worse that could happen if someone had no infotainment? Not much, they'd be bored."
Wow, I never thought too deep about HD Radio because, when it first came out, most of the time the radio was still not connected to the car in any means other than to draw power and output to the speakers. But now that infotainment centers are so intertwined with the rest of the vehicle, I really see how HD Radio becomes a vulnerability. Someone could broadcast malicious code on those frequencies, and it would be automatically downloaded, without question (assuming it meets requirements the computer enforces to accept the signal for processing). I'm not sure all of what's possible with that, but this may just be the tip of the iceberg. Let's hope I'm wrong.
You "could" send code via RDS and UTF-8 Chars. But i guess most recivers does some form of sanitazing of text, only a guess tho as im volonteering at a local FM station so not 100% sure about how much code that could be sent and nothing i would like to try to do On-Air. However some testing would have been done in this area.
@@pmr1049 yeah, I’d assume most radios are coded in such a way to accept the text only as a string, and nothing more. But since Mazda had this vulnerability, it makes me wonder what other undiscovered vulnerabilities may be out there.
@@brickson98m yes but also there is chars limits you cant send that much text on a given line as its rotating per string. But there might be areas like in this instance where something could be sent. As trasnsmitter really doesnt care ”whats going out the pipe”. Just a bunch of 1 and 0 over an Multiplex
@@pmr1049 and @brickson98m, How cute. With one of you guessing the station sanitizes the text being sent, and the other assuming the radios process only text, you've laid out exactly how what happened in the video was allowed to happen.
@@ElectroDFW Awww what cutie XD
When I first saw "tuned" in the title, I thought of car modification/tuning. The amount of modified Mazdas is insane
Yeah, not sure why they are so popular with ricers.
@@Bacopa68 If you want to see ricers, keep your eye open for all the "pimped" non-Evo Lancers driving around. Interesting fact: real tenth gen Evos have square gas cap doors, all other pedestrian Lancers of that era have round gas cap doors. So if you see a Lancer with a huge wing, loud exhaust, tackly LED lighting and a round gas cap door, the owner just threw cheap mods at a 148hp turd.
Mazda always had a sporty track focused side. Throughout the 1980/1990s Mazda was very successful in sports car racing. Now days people just know Mazdas as being expensive latte cup holders and baby or puppy haulers. Didn't always use to be this way, brand use to have character
Chrysler had a problem similar to this back in the 90's when OBDII came out but it had nothing to do with radios. On those early Chrysler cars if you scanned the computer and unplugged the scanner without turning the key off first it would glitch the computer and set the charging voltage to 0 volts and the alt would never turn on again unless you flashed the computer with new specs. Shops were replacing alternators left & right before they found out it was the computer.
This reminds me an awful lot of what happened to me and a fellow apprentice. Our programming teacher wrote a little bat to delete all those unnecessary files compilers like to generate. Files that the completed projects no longer need and that the compiler can easily generate again if necessary. We used Borland C++ builder, which also generates files without an ending. In the batchfile you'd simply write "del *." to get rid of them, and in school, on my USB-drive, and even the local harddrive of my work-PC this worked fine. However, on the server it got interpeted slightly differently: first of all it got treated like a "del *. *", deleting ALL files, and it also worked on the whole drive instead of my folder and its subfolders. Basically I ended up deleting all the projects I had access to which due to me hopping between departments were quite a lot. Thankfully a friend in IT and the recovery function dealt with that rather quickly, but my friend showed the batchfile to his boss (telling him that I'd asked whether this is safe to use) and I got a nice letter to give to my teacher explaining how dangerous this can be and that if we really had to bother with saving a few kilobytes we should do it manually
Love my analog car! And by that, I mean it has a few dozen various ecu’s that control everything from the engine timing to knowing when to set off the airbags. What I like is my car has a manual transmission, standard dials for heating and cooling, and no touchscreens. It’s about as analogue as you can get these days, in the used market that you can daily drive.
*laughs in 1994 Volvo* I'd like to see them hack my cassette deck. About the worst they could do is turn Dolby NR on and off.
@@moconnell663 by 1994 you probably have the same ECUs as OP, but of course they’re not connected to the antenna at all.
@@kaitlyn__L what's funny is when i was going through the electrical schematics looking for a wiring issue with the lights I discovered that the ECU has absolutely no idea how fast the car is going. Absolutely none. The rear speed sensor is hooked up to the speedometer in the dash and nothing else. The front sensors are hooked up to the ABS module, which has no connection to the ECU at all. There is nothing which could be described as a BCM. It's an automatic transmission, but is totally self-contained, no coms except for a switch to inhibit overdrive.
@@moconnell663 that’s not too unusual, the ECU looks at shaft and spark timing and so on, and maybe measure fuel consumption, so I can understand not really needing to know how fast you’re going, especially since RPM are decoupled from MPH anyway.
and probably gets a mile per gallon
Wait so both Mazda and NPR took zero responsibility and didn't pay for anything? I don't expect it from NPR, but Mazda definitely should--seems like a safety hazard that might be prevalent in other systems and definitely worthy of a voluntary recall.
The Question is, if this is still the Case in new Mazdas. I have a 2021 Mazda 3 .
@@BenBensen293 If so, I'd bet on it that there will be a software update incoming, maybe even over the air, if Mazda is supporting this. But they will for sure not let this happen again.
NPR did nothing wrong. Transmitting a file without an extension is fine.
@@BenBensen293 4th gen Mazda 3 uses a different infotainment system than the one in the video. There was a bit of a revision in the 2017 or 2018 model year to introduce Android Auto and Apple Car play, which coincides with the video saying it only effects 2014-2017 Mazdas, so I'm assuming based of that information that the issue has been fixed for years.
If Mazda doesn’t, that’s a class action waiting to happen. According to the bbc tho, Mazda is offering an application to people for a “goodwill” replacement, it just isn’t clear how difficult it is to get and when those replacements will actually ship.
Been a Seattle resident for my entire life (3+ decades) and used to listen to kuow a lot, I hadn't heard of this. Wild stuff.
I've heard of a similar incident affecting TVs. Some station sent bad closed caption data and it overflowed some memory buffer in the receiver. Depending on the TV it would suddenly change settings, mute itself, or stop working entirely. (Or, in most cases, not do anything.)
Who says newer is better? Smh.
Seems like new technology is sometimes worse.
That %i isn't even about using it somewhere, it's about trying to do something stupid. It's almost like someone decided it'd be a good idea to eval station names instead of just, I dunno, doing sanity checking and outputting it. That way lie exploits.
Honestly everything about this just strikes me as terrible design decisions.
preeeecisely
@@chyza2012 Why learn and use `fputs` safely when I can use `fprintf` and not have to learn a second function name?
@@killerbee.13 The function call sprintf(dest, "%s", someString) will act like strcpy but report the number of copied bytes rather than returning a useless pointer to the start of the destination. If the %s is omitted, the function will work the same way provided the second argument doesn't contain a percent sign.
You can blame the radio station. And you'd be wrong. In computer security, principle number 1 is that everyone is responsible for their own safety. You can't just ask everyone in the world nicely not to brick your car.
You can blame Mazda. And that's a bit better. Not expecting a file parsing error? Major mistake.
But I myself blame someone else. The FCC. Couldn't you have come up with a standard where the same type of data is always transferred in the same type of way? The standard doesn't specify what to do with arbitrary files. So why the hell does my radio require arbitrary file transfer capacities? It's like the FCC was asking for a faulty implementation.
Tell that to the guys being sued for using Google Fonts in Europe. I've been told by everyone that it's their fault that people's browsers send referrers.
I'm pretty sure there is a standard, someone just forgot about it in this case
On the other hand once the ffc specifies something they will never change it even when it becomes outdated
HR Radio in Southeast Connecticut:
a) works UP TO 5% of the time,
b) available on only four or five out of literally dozens of stations,
c) drops out when clouds pass by overhead, or when you're descending a hill, or turning a corner, or driving under an overpass, or driving under trees, or when it starts raining/snowing, or you open a window, or you raise the trunk lid...
I'm sure it's WONDERFUL in flat, unobstructed areas, like if you were parked in a parking lot in the midwest, and your local station spends the extra money to invest in themselves. WHEN it works, it's a night-and-day difference in audio quality, and the subchannels offer additional listening options. The caveat is it requires a strong unobstructed signal that often can't be maintained while sitting parked. In motion, you're constantly in-and-out of HD which is annoying - you can't turn the feature off, so you either live with it, listen to channels without HD, or use another source.
HD Radio works almost as well as the RDS system. Wrong song title/artist displayed, garbled text, odd combinations of title and artist appearing as one word, and I had to turn off the "Auto Clock Set" function since my radio's clock would change with every RDS-capable station. Imagine my shock when the clock read 0:39 for the time - it didn't allow a zero to be set for the hour manually - and sometimes it's off by multiple hours AND minutes, as if the radio station never set their time correctly.
Sanitize your inputs is security 101. That dev needs to be writing PHP in Siberia.
I work on Mazda's for a living and have heard about this issue - apparently, a previously released software update fixed this issue - both the % and the extension-less image problem.
The problem being that to *install* the update, you needed to have at some point brought it into a Mazda dealer and ask them to update the CMU. If you hadn't updated in time, you'd need a new CMU.
You are meant to say Mazda sponsored you a new entertainment system right?
@@thisconnectd My car is old enough that I can install an aftermarket option 😂 and yes, Mazda corporate is replacing these under an extended warranty (although there's a relatively limited number of CMU's available in the PNW...)
I would hope self driving cars would decouple their driving OS from the radio (and all the other stuff like HVAC).
Yes you would. Trust your life to lowest bidder spaghetti coders.
Anything else would literally be insane. But given that the head of THE pioneering EV manufacturer is Elon Musk, it's nowhere near as unlikely as I'd like.
The system that controls the car's physical movement should be completely isolated from the non-critical creature comfort systems of the car and especially from any external (radio, WiFi, Bluetooth, cellular) interface, but of course, 'flashy' futuristic functions like remotely starting and "summoning" the car would not work without some connection (= attack vector) between the "smart" systems and the drive control system. Yep, it's creepy!
Be nice. But i doubt it will happen until something major happens and someone "important" gets hurt from it
That would make it more expensive, so no. To make computerized things actually reliable you would need airplane style redundancy - not gonna happen. Throttle, brakes, and steering are already electronically controlled - in any newer car you're basically just using a remote controller that is mounted near the driver seat.
And oh yeah, they're going to mix all that up together in one system with the computer and the radio and another computer for absolutely no reason!
AFAIK no car with ADAS runs it on the same hardware as the media.
I love me some carburetors and point ignition.. and the little computers I build in the basement from component parts... and program myself using skills I started using in 1984. I'm livin' the life. Ain't got no flies on me.
F.M. stations were around commercially as early as 1941, and even earlier experimentally (wide-band frequency modulation was developed as early as 1933). The first F.M. band was actually from 42-50MHz, which the upper 6 MHz were claimed for T.V. Channel…ONE!
If that cache is in RAM, not FLASH, it should be fixable by pulling a fuse. Removing power from the radio will clear all of the RAM. If it's in FLASH this won't do anything but if there is a way to do a hard reset that should fix it.
That is a smart idea if it is indeed ram.
what pisses me off is... how can a freaking radio signal cause _everything else_ to also crash?? how dumb of an OS is this ICU running??
The software was indeed dumber than software produced in an earlier era.
Back in the mid 90s there was an image file viewing program called SEA from Photodex Corp, that software was made smart enough to SCAN the header content of files to determine which type that image file is and then proceed accordingly.
From the story, it seems such good practices have been abandoned.
If the only known fix is replacing the device, probably someone has figured that pulling the fuse doesn't fix the issue.
@@rashidisw Only on Windows. *nix has been using file information and magic numbers for years. No reliance on extensions at all. GIMP will happily open a JPEG named test.mp3 (assuming you're on *nix).
I’m not even surprised at the fact that glitches like these occurred, but why did these relatively small errors have to result in the entire computer bricking? Imagine that happening to your PC. I don’t much about software, but that’s why I’m wondering why failing to open one file had to lead to an infinite loop.
Maybe they were terrible at handling exceptions and failures
Because Mazda apparently has/had a completely substandard software development process.
@@PrezVeto I wonder if they outsourced their Infotainment software
That's what I dont get. Windows basically requires file extensions but if it encounters an extensionless file it just goes 🤷🏿♂️ and just holds it there instead of shitting its pants
Embedded systems like onboard computers are often written directly on the metal, without an operating system to help with stuff like that. Although specialised operating systems are also sometimes used, and I believe infotainment systems usually run some kind of Linux because they don't have to be streamlined for maximum efficiency (unlike safety critical stuff like, say, a chip that controls your breaks) and need graphics and audio support which would be hell to code from scratch. At least that's how we do it where I work as far as I'm aware
Anything but NPR purgatory!
Honestly though I love listening to the opening montage to NPR Viewpoints. It always amazes me the professions that people get paid for.
factgory resetting the radios would clear the cache and stop the system from bootlooping. it wasn't just an issue from the HD radio, if you had songs on a device with clipart that had a weird file extension, it would cause the system to loop also.
I was on the team who promoted and supported GIF on CompuServe, and it pleases me to no end that it's still being used - and debated - 34+ years later. :) -team soft g
so...i pronounce it giraffics interchange format? 🦒
And here I am on team H because of Spanish
@@cs8712 So I pronounce LASER as Lah sehr?
:D
I was just going to come back to this comment and ask you how to pronounce it... and then saw "team soft g" in your signature block and realized your opinion is wrong. =0(((((( 😁😁
Imagine the captive audience that NPR station has for their fundraising drives!
Man, that’s wild! If only they had cinderblocks in the passenger seat to protect them, like you alluded to. Please do a video on that and about cinderblocks in general
1:55
this is incorrect, RDS (radio data service) hides a signal in the transmission that allows compatible radios (quite common) to decode text with info about the station and its media
Ha! Jokes on you, the Mazda I drive near Seattle is too old to be fried
"... because hiring cellists to sit on your dashboard was getting too expensive." Ahh, yes. I can remember my grandfather telling me about the dashboard cellists they used to hire back in the day.
Well, it was the fashion at the time.
Note that a cellist also brings a windshield wiper...
This also explains why someone might actually ask a genie for a two foot pianist.
They put the drummer in the trunk..... 😂😂
The $15 motion alarm I bought from Harbor Freight, as it turns out, can brick the info console of my 2019 Ford Raptor. Lucky for me, it clears itself by power cycling the truck. The wireless detector has a ~435 mHz signal to the separate base unit/alarm.
The new car dashboards are on a size of a tablett with many functions.
You actully have to look on them when you want to switch the currently playing song, turn on AC, turn of the lane assist systems that wants to push your car into the incomming traffic because you are on a narrow street with no middle line.
But you are not allowed and gets fined alot (rightfully) when using a phone while driving.
there is something wrong with new cars and it is getting worse.
If the manufacturer builds a driver-distraction into your car, that's perfectly fine. But if YOU bring one into your car - GUILTY!
2021 VW Jetta w/ manual transmission. I downshifted on the highway, engine blew and I started warranty process.
VW said the computer reads I was over speed limit, what gear I shifter into, what song I was listening to, and used that to determine they wouldn't cover the warranty.
My insurance luckily stepped in and called BS but companies are now going to use your driving analytics to determine whether they will cover the warranties part.
Every computer engineer knows that computers are a mistake. Having seen the horror upclose I'm very concerned about more and more of them controlling larger parts of our lives.
Yet, you remain connected to the internet... Try living in a mormon community for a month. Bet you can't wait to surround yourself with technology afterwards.
I do it myself. Constantly complaining that decades ago live was simpler and therefor better. Yet, when my phone broke, I did suffer from kind of a panic attack... ;-)
@@2Fast4Mellow Huge difference between going Mennonite and wanting to go back to 1995, friend.
I hate that the world is completely dependent on them now, even people who want to disconnect pretty much can't. Not if they want to keep a job or maintain contact with friends and family. Even landlines are mostly VOIP now. If the internet goes down we're fucked, and we only have ourselves to blame. I miss the 80s and 90s so damn much.
@@calanon534Man I wish that all people who „want to go back to 1995“ just hold true to their promises and stay away from social media. They will be happier, and us others won’t have to listen to their miserable complaining.
"Computers make excellent and efficient servants. But I do not wish to serve under them" Spock, in the 1968 Star Trek episode 'The Ultimate Computer'!
What a funny coincidence. Your two AM examples of KDKA Pittsburgh and KBOI Boise are the two cities which I have, and still, live.
Was the person at KUOW that released the download called Bobby Tables?
Sam's mom: "Why don't you call more often?"
Sam: "You seeeeee...."
Soooo... basically what would happen if SQL injection and Windows OS had a baby.
So the baby is Mazda OS?
lol true
How has these software devs that work on these car systems not heard of the #1 rule: never trust external data? Verify it, sanitize it, scrutinize it, assume it to be malicious at all times. If they applied this philosophy then we'd probably never have these issues until someone found a way to get data into the system in a way thats not checked
This type of thing has happened to the best of the lot. IBM once had a program that only had one single instruction. It was discovered to have an error and was corrected. They then discovered that they had introduced a second error. It also was corrected. Luckily no further errors were detected with this program.
This is actually a very good simplified explanation of several relatively complex topics.
Good job HAI
This is ludicrous. I mean, apart from the fact that even very simple computers _decades_ ago could work out file types without extensions, if Mazda _really_ found it too hard to implement such a feature that… again… has been available for decades… any unrecognised file in an embedded car computer system should just be ignored, and probably also deleted. As it stands right now, you could literally drive around in a car with a pretty low power (hard to detect) pirate radio transmitter outputting on the same frequency as a popular local radio station, outputting tiny unrecognised files, and brick a good number of the Mazdas you drove past. To be completely clear I’m not endorsing that anyone try this, but most people with my (pretty good but not genius) level of computer skills _could_ do this, and if I lacked the ethical beliefs I have, I was bored, and I wanted to troll some drivers, I would _totally_ be doing this.
Mazda hatred will not be tolerated
Yeah, they'd literally just needed an empty else-statement/default case to avoid unexpected filetypes or lack of filetype information from doing any harm!
It's probably more an oversight in coding rather than laziness or cost saving, but yeah considering the multiple issues presented they need to be testing their code better
ludicrous today and scary tomorrow, imagine if this happens in self driving cars, who knows what can happen?
@@poopmeat15 I love Mazda, one of my exes had one (a sporty two seater) and she’d drive us around in it, it was awesome. That’s why it’s so disappointing to me. And being a really nerdy girl, bad computer design frustrates me 😂
I like that you put KBOI there. I’m from Boise and I love it when people realize we exist 😂
My 2017 Subaru outback had the album cover for Lynrd Skynrd and the name Sweet home alabama stuck on the screen for several months no matter what song was playing. I tried disconnecting the battery, rebooting, changing stations etc.. It finally went away after my car battery died and the car sat for a week.
You'd think that you could just reflash the infotainment system's hard drive or SSD to clear the software issue rather than replace the whole unit.
This is why most cars have separate networks for low priority signals (radio, infortainment) and safety critical systems.
Umm, not really. They have separate units, but many cars have common networks for both. There are software 'firewalls' between functions, but a bug in that...
I'd like to think that safety critical systems don't have any network access whatsoever!
But apparently sanity is at a premium in the car industry these days. And brain fuel prices must be at an all-time high, too...
Talking bout remote start and remote control/"summoning" of a two tonne car...
@@LRM12o8 Even in most drive-by-wire cars, the brake pedal is still mechanically connected to the braking system. Even in the face of a malicious firmware update, pressing the brake should always work.
@@TheCountExtreme Steering as well. Electrically-assisted, but mechanically connected.
Well, I do have a computer science background - and the fix for Mazda is relatively simple. Likely, the operating system used in their infotainmen.....car computers, is some form of Linux - which has an army of open source utilities available to it - including GNU File Magic, which offers the ability to inspect file contents to determine the type of file. A relatively simple modification to look at the magic result instead of trusting the file extension would/could have fixed this issue. Sadly, auto manufacturers are like phone manufacturers - once you buy the device, they aren't interested in doing software upgrades.
I'm wondering if a similar programming issue is what caused my Hyundai's infotainment system to reboot when I tried HD Radio for the first year I had it. AM and FM worked fine, but if I toggled to HD, it would crash and reload. Thankfully it didn't brick the system and seems to have resolved itself now, but still.
Software engineer here. The funny thing to me is that the JPEG standard requires a “magic byte”, which is essentially a bit of metadata at the beginning of the file where it says “hello! I’m a JPEG”. Reading this magic byte is necessary for correctly decoding the JPEG, so crashing is extra odd.
Presumably it tries to find the extension, which of course isn't there. A zero length string or null pointer later and...
@@loc4725 yeah, they’re probably using an open source / well established rendering library, this seems like an OS bug in the bit of code that’s handling the file descriptor. Most desktop operating systems are smart enough to just “deal” with missing or even incorrect extensions, I guess infotainmentOS probably isn’t that sophisticated
@@kyle-silvershould've just ignored it instead of crashing into boot loop.
Worth noting that files will often have a "magic", from which you can detect what file-type it is
Are you talking about file header? Because it has a flaw. What if I transmit the same text as a file header? That would effectively make computer think that it is not a text file and it would crash the computer.
Fundamentally no matter what, out of band data shouldn't brick your damn system. You should fail fast and fail robustly. Unfortunately robustness testing is usually the first form of testing cut short by date-obsessed bad project schedulers, right around the same time as regression testing.
@@aashd9245 Well, you don't always need the header to detect the file type. If a file begins with the bytes ff d8 ff, it's probably JPEG
If it begins with the bytes 89 50 4e 47 0d 0a 1a 0a, it's probably PNG
If it begins with either of the two strings "GIF87a" or "GIF89a", it's probably GIF
That's how detecting file type with "magic" works
@@4906 That's the definition of file header. I know this because I have experimented with raw files and image files in assignments.
What would happen if I used hex chars ff d8 ff in text file?
It all comes down to trust. The assumption should be, the data is malicious, it's trying to do something bad. Assume that and act accordingly until it is proven that the data is safe. It should not be possible to form any kind of data to do anything bad.
ah. ancient radios.. There is a video on youtube of some genius building a basic radio with basic stuff, some sand and a soldering iron. That was the most amazing thing I ever saw
I grew up in Seattle and I guess I moved away well ahead of this pandemic in the HD radio world.
Now in a Midwest city where I have a great stereo system in my car I know to simply not use HD broadcasting which might one day kill both me and my car radio.
Thanks for the heads up. 😂
Fun fact, basic info like station identification, artist, and song titles COULD be embedded in analog FM signals using something called Radio Data System (RDS). Basically it was a small digital signal sent on a sub-channel that radios with the right circuitry could pickup and decode.
RDS is still in use.
@@littlemeg137 Sure is! I live in Canada where we abandonned plans to move from FM to DAB (although a small handful of stations do use IBOC/HD Rario), so RDS is still a mainstay here 😁
@@BroadcastBlueprint Yup, by my point was that the video mentions there was no way to broadcast song data with old analog FM signals, and that's not correct.
@@SpencerN.C.you should’ve kept going with DAB instead of HD. It was a mistake for us in the U.S. to develop HD instead of DAB.
@@jayrogers8255 DAB+ is much better. In my experience I've sometimes had better DAB+ reception than FM.
FM falls apart quickly as soon as multipath propagation is involved.
2:55
The Graphics Interchange Format, or GIF, was first developed by computer scientist working at CompuServe back in 1987.
And while it has swelled or dipped, the debate over how to pronounce the acronym for those minute looping animations became a thing once the GIF really took off. Is it a hard g like in graphics? Or a soft g like giant? Answering that question depends who you ask and whose authority on the matter you believe in.
Here is a timeline of how to pronounce GIF, one of the great debates of the internet age.
June 1987:
Steve Wilhite releases the Graphics Interchange Format, or GIF, while working for Compuserve. He called it a GIF with a soft g. “Choosy developers,“ he reportedly said, “choose JIF.” This was of course a play on the peanut butter brand Jif’s line “choosy mothers choose Jif.”
1994:
The pronunciation debate was apparently already in progress. Australia’s ABC reports that “[i]n 1994, the author of an encyclopedia of image formats said ‘most people’ seem to prefer saying a different way than the GIF godfather instructed: GIF.
November 2012:
GIF is selected as the Oxford Dictionaries USA Word of the Year 2012. The dictionary wrote that “GIF may be pronounced with either a soft g (as in giant) or a hard g (as in graphic).
April 2013:
The White House announces its new Tumblr page where, according to New York magazine, it threw down the pronunciation gauntlet with an illustration that told visitors to the page that they can expect “ANIMATED GIFs (‘HARD G’).” This was the same year that we got GIPHY, a GIF database people could search for the GIFs they wanted.
May 2013:
Wilhite receives a Lifetime Achievement Award at the Webby Awards and used his platform to make his declaration. “It’s pronounced JIF, not GIF.” Just like the peanut butter. “The Oxford English Dictionary accepts both pronunciations,” Wilhite told The New York Times. “They are wrong. It is a soft ‘G,’ pronounced ‘jif.’ End of story.” But it’s not so simple.
May 2013:
Many people on the internet disagree with Wilhite. Gizmodo declared him “wrong” and Twitter filled with people expressing their disagreement. “Pffffffffffffff,” posted one; “Graphics Interchange Format. Graphics. Not Jraphics. #GIF #hardg,” tweeted another.
May 2013:
At the time, the J. M. Smucker Company, which owns Jif, agreed with Wilhite’s pronunciation, tweeting, ”It’s pronounced JIF”.
June 2014:
President Barack Obama chooses a side, announcing his “official position” is that it is pronounced GIF with a hard g like grape.
June 2014:
A survey of more than 1,000 Americans were asked whether they “pronounce GIF as ‘jiff’ or ’gift‘” and “‘gift’ handily beat ‘jiff,‘ nearly 54% to 41%.” It was conduced by eBay Deals and a digital marketing agency.
August 2014:
Designer Aaron Bazinet launches website, howtoreallypronouncegif.com, arguing for the hard g pronunciation. “It’s the most natural, logical way to pronounce it. That’s why when everyone comes across the word for the first time, they use a hard G.”
August 2015:
Mental Floss wades into the debate with the help of a linguistics professor for a comprehensive analysis, ultimately deciding that both sides are correct and that the g in GIF “really can go either way.”
July 2016:
Newsweek declares it GIF with a hard G and has a linguistics professor to back them up.
February 2020:
Online GIF site GIPHY teamed up with Jif peanut butter to have some fun with the debate. The two companies unveiled a limited-edition jar of peanut butter in Jif’s trademark packaging, but labeled “Gif”. While some may think the packaging implies that GIF and Jif rhyme, according to the two companies-and a series of accompanying GIFs on GIPHY-the opposite is true.
“At Giphy, we know there’s only one ‘Jif’ and it’s peanut butter,” Alex Chung, founder and CEO of Giphy said in a press release about the campaign. “If you’re a soft G, please visit Jif.com. If you’re a hard G, thank you, we know you’re right.” That’s right: according to Giphy, GIF is pronounced with a hard G, unlike Jif the peanut butter.
So in conclusion, please pronounce it gif with a hard g or I will come to your house and make you listen to a 24 hour loop of someone saying gif correctly and the bugs bunny turning into big chungus gif. Consider yourself warned.
Choosey parents choose GIF
I had to scroll too far down to find this. #hardg
Very well written. It's a hard G! Like the word it stands for. May anyone who thinks otherwise drown in the fleas of a million Wombats.
@@forgottensavage5584 ❤️
Soft g.
Here's another "glitch" involving car radio: If you have XM Radio in your car, sometimes your cities' traffic lights will emit enough electromagnetic interference to "swamp" the satellite signal incoming to your XM receiver. No damage is done, but you may have to sit through that long stop light with no sound from your car stereo (it gets muted with a bad signal).
Can't be any worse than losing FM reception momentarily going through a freeway tunnel, no?
@@ScubaSteveM45 that’s right
It's not just Sirius XM. Lights can also affect regular FM signals too. There are also these LED screen billboards in my city and whenever I drive past them, my FM station goes to static for a second.
The LED lights they've been installing in gas station canopies also seem to have the same effect on radio signals, since the radio tends to cut completely to static when I stop to get gas.
It happens at certain drivethrus, too.
I just realized the second half of your channel's title is "as I thought it would be"
Arguments that Electric Vehicles are good/ bad as instructed by Sam
EV good. Fight me!
EV good, cars in the aggregate bad, city design that necessitates driving a car terrible.
Death threat to Sam for mispronouncing gif as instructed by the editor
I don't know doesn't seem to thoughtfully-worded
The internet was polled and 70% pronounce it Gif and 30% Jiff.
Its officially Gif and those in the 30% (including the original creator) shall now be notified of this death threat for being wrong as requested by the editor at 2:50. Please enjoy your next 24hrs.
I have no words
June 1987:
Steve Wilhite releases the Graphics Interchange Format, or GIF, while working for Compuserve. He called it a GIF with a soft g. “Choosy developers,“ he reportedly said, “choose JIF.” This was of course a play on the peanut butter brand Jif’s line “choosy mothers choose Jif.”
Classic bootloop. My phone had a broken ram drive & it caused a bootloop. Kind of odd that they don't include simple debugging tools on the consumer end like you get with a personal computer/phone. It's SUPER easy to clear a cache or load a backup. An expensive mistake with a solution that would cost pennies per car if implemented proactively
What really annoys me, more than this happening or that it can happen, is the cost of the repair or replacement. Back in the day when you could easily swap it out with aftermarket head unit it wouldn't bother me so much. But $1500 is insane. And yet, its probably on the low end, compared to what you might pay to replace the entertainment unit on something like a Range Rover or a Porsche.
Cars have been dependent on electronics and software in safety critical systems for over a decade, them becoming electric is irrelevant. As cars gain a greater level of autonomy and over the air updates, that will become more problematic. However modern cars are already significantly vulnerable.
I can't believe that there is no "Hold these three buttons down simultaneously" or "Poke unbent paperclip here" or "disconnect battery for 60 minutes" fix for this.
I heard about the Roman Mars Mazda Virus when it came out (that was a great episode of Reply All, what a shame that podcast was doomed) but I had no idea there was another incident with a radio station. Oh Mazda
"Eternal NPR purgatory". Boom. Thanks for the new name for my indie band!