Hacker News Reader

Comment by acheron9383 | Hacker News Reader

As someone who works professionally on embedded software devices that update over the internet, car companies are stuck not because they can't get software talent, but because they have no ability to actually build the electronics alongside the software, which is ultimately what constrains embedded software. Without the right hardware, the constraints are just insurmountable, you can not do X feature because board A doesn't have the API to your MCU, or it runs some dogshit speed communication system that means you have 500ms lag. The feature is just unworkable, and if the PMs push it anyways you get what happens for the legacy car makers, terrible underpowered infotainment systems with no central design philosophy, stuck in an awkward, bad, middle between a full software stack and all buttons for everything. Their model of integrating 3rd party vendor computers just doesn't really work for this kind of thing; Tesla, Rivian, and the Chinese EV makers all manufacture all their own electronics, which lets them achieve the outcome. But you can not just roll all your own electronics in a year.

View on HN

I worked in similar systems and you are 100% right. 80% of the time was spent on communication protocols between the different boards and microcontrollers. QAing and solving issues from short-sighted dozens of unique custom protocols that worked in non-standard ways (every time a component needs to talk to another component a new protocol was invented).

When you have dozens of communication lines required between different parts of the system it becomes just as complicated as your average micro-service cloud. Really, a car is a distributed system with dozens of "services". An analogy is that each microcontroller-microcontroller communication use their own custom binary-encoding API that runs on multiple different, incompatible versions of HTTP.

We actually spent considerable amount of time just developing our own custom protocol for communication that could run on all sorts of different physical interfaces (CAN, ethernet, modbus, etc) as well as a series of proxies between devices (so component A can talk to component C through a proxy in component B). And if we had to use a custom protocol from an external manufacturer we had to wrap it into our own custom protocol.

That protocol was actually used for our cloud data reporting as well, so eventually all our data communication would use a single unified protocol from micro-controller to IoT Linux to cloud data-ingestion pipeline to database.

For american cars at least, I read that one of the reasons this process exists is because car companies want to work around union rules for manufacturing by outsourcing components of the cars to subcontractors that they can make deals with.

Ultimately it's a price control strategy to pit these suppliers against each other to lower costs. But it means that designing these electronic sub-systems isn't just a question of the design itself, but also of managing all of these supplier relationships as well, they all have different contracts, you would have to coordinate all of them at once to make sure things are interoperable, etc.

That smells plausible, but from my seat as am armchair car enthusiast, it seems that foreign automakers outsource components just as often.

and have the same problem with software ? yes and yes

>price control strategy to pit these suppliers against each other to lower costs

Apparently that rabbit hole goes super deep in which the large auto manufacturers in the US throw their weight around and force suppliers into selling parts at cost or with razor thin profit margins. And on top of that, they force the suppliers to eat the loss when it comes to cyclical business demand (e.g. storage costs for over-producing during low demand and increased labor costs during times to under producing from high demand)

Conway's law strikes again!

You'd think the overhead of managing the supplier relationships would be more expensive than well-managed vertical integration. I'm guessing it's a failure on the part of admin to count their own costs.

Those employees might not be unionized, though.

If that was a main factor, surely then Mexico should be ground zero for next-generation car electronics design? Like, Tijuana or Ciudad Juarez.

I can't speak to car electronics design or anything about the capabilities of Tijuana or Ciudad Juarez, but your comment had me reflecting that in the mid/late 90s those of us who were in b-school went from all about Mexico to all about China overnight. With NAFTA et al, American manufacturing was going to be all about developing factories in Mexico and moving good back and forth between the US and Mexico for design and finishing. Then once China joined the WTO it was absolutely an instantaneous pivot to China instead. Maybe all that's old is new again?

Right. The unwritten story of the last three decades is the Mexican people in general have lost out on the huge generational opportunity they theoretically should have from being a (comparatively) low-wage manufacturing and R&D option on the US's doorstep ("friendshoring").

On paper Mexico reaped enormous gains from NAFTA(/MCA) but those went to a small slice of the population, and people on fixed incomes/not in manufacturing/services objectively got worse due to inflation and rising cost-of-living, property prices. The share of GDP effectively lost due to corruption is also an issue.

> (every time a component needs to talk to another component a new protocol was invented).

i'm sure that every time this happens, it individually makes sense to do it at the time.

This is a microcosm of how large systems get developed in small pieces, by different people, over a long(-ish) period of time. It's the same in the software world too i think, but presumably has a lot more consolidation than cars (as software for cars might be less common, and thus employees moving between companies is unlikely to make any sort of cross-pollination like there would be for FAANG-like companies).

This makes it sound like the problem is that they either lack a person with architectural responsibility for the cars' electronics as a whole or that person lacks the skills necessary to do their job.

Bear in mind that all the electronics on a particular car are not specific to that car: there is a lot of reuse across product lines. And there are multiple vendors, each of whom is probably also selling the same, or similar modules to other manufacturers.

No, it’s the org and incentives structure - maybe the only people who have all parts that need to make change under their command are the board, and until now, the software was an unimportant part for them.

This aspect of the industry has seriously regressed. We started out trying to standardize and as vehicles have become more dependent on onboard networks manufacturers have gone completely proprietary and have put all information behind lawyers. The consumer is the real loser.

> every time a component needs to talk to another component a new protocol was invented

> We actually spent considerable amount of time just developing our own custom protocol

Not only is this unintentionally hilarious, it’s a real life example of an xkcd comic (https://xkcd.com/927/) that will never cease to be true.

> eventually all our data communication would use a single unified protocol from micro-controller to IoT Linux to cloud data-ingestion pipeline to database.

This, however, is remarkably impressive, that you were able to build a single protocol that fit this end to end use case.

It is really hard, especially given you have to optimize for the lowest common denominator. For us it was a 512kb RAM microcontroller, we had to go to procurement to expand it to 2MB RAM and they were not happy about that.

On the other hand it was nice being able to just import a library into your code and JUST SEND A FREAKING MESSAGE without having to deal with thousands of lines of code that were last changed 3 years ago and nobody knows how it works. The scrutiny on the code quality of the common protocol was much higher and therefor much more pleasant to use and troubleshoot.

All the encoders and decoders of messages used the same code in all the parts of the stack (technically 2 implementations, one in Go and one in C)

Think of the nightmare 5 years down the road when someone else has to then incorporate this protocol under their own new protocol, with the older ones nested inside.

I'm getting IG videos in my feed for a company that sells after market fixes because older Teslas have such poorly designed electronics, that they fail in common ways. The memory goes bad because they write useless logs to a chip, and it eventually fails. End users are beta testing...

https://www.instagram.com/reel/DINADISyP0f/

That’s always been the case with Tesla. I still have no idea how the yoke with no progressive steering and a tiny button for a horn ever passed any sanity check. Not to mention the NHTSA.

Oh, I wish they would install tiny horn buttons on all the vehicles in Vietnam! In that country, the horn is a method of communication, much to the ire of literally everyone trying to exist.

Excessive horning (made up word) is not just a Vietnamese thing. Italy is probably Europe's worst offender, with Greece a close contender.

I'm not so familiar with Asia, but I get the impression that the entirety of Indian and most of Chinese drivers feel the need to lean on the horn with gay abandon (fnarr).

In Britain the horn is generally reserved for "fuck that was close: I think you are a bit of a tosser" or "you are driving a German car and seem to have have no indicators".

In india if you hit someone after sounding your horn you are not at fault as you gave warning and they didn't move. (It is far more complex than that but as always the real truth is too complex for a comment box - if you are trying to drive safe it is close enough, but this isn't a license to murder), As a result all drives will honk their own if there is any possibility someone might cross in front of them.

India is getting a lot stricter about driving rules, and I hven't been there for a few years. I would expect the above to change as people realize that the horn doesn't really work for that purpose anyway. But change is always slow.

Was something like 20 years since I was in India, but IIRC at least back then they didn't have a "priority to the right" traffic rule, but rather some kind of "the one who first honks has priority". Traveling in a taxi felt suicidal, drivers just honked when approaching an intersection and continued blithely.

Based on a quick googling, this seems to no more be the case, and there is a 'priority to the right" rule.

I saw something similar when visiting Latin America a few years ago, in a neighbourhood friends lived in. Some people would just go full speed in residential streets and hold the horn while crossing intersections because I guess that's what's gonna keep their car intact?

Latin America is “just” 33 countries each with their own culture (not to mention Brazil with its massive size).

Yeah, I know. I just don't want to self-doxx by mentioning the country.

I also don't think this is part of any specific culture, mine or someone else's, it's just something I saw in one neighbourhood.

By priority to the right do you mean a French style priorité a droite? Or American style stop sign priorities?

Neither system describes how Indian traffic works, which is much more of an iterated cooperative fluid dynamics simulation, with the main rule being ‘don’t drive into people who are in front of you’.

And they drive on the left, so priority to the right makes no sense.

> By priority to the right do you mean a French style priorité a droite? Or American style stop sign priorities?

I mean

https://en.wikipedia.org/wiki/Priority_to_the_right

That is, unless there are other signs overriding it (like yield or stop signs), you must yield to someone coming from the right in an intersection.

> And they drive on the left, so priority to the right makes no sense.

Not sure that matters. The important thing is to have a consistent rule that everyone can follow. Whether the rule is to yield to the right or left doesn't per se matter, nor does it depend on which side of the road you drive on.

I do agree consistency is probably the most important thing, but deciding which side should go first can have differences based on if you're driving on the left or the right.

Imagine a 4-way stop of 1 lane each way roads. Cars drive on the right in this example. One car rolls up traveling from west to east and another car from south to north.

If we give priority to the right, meaning the south->north traffic, they end up out of the way of the other car sooner than the car on the west to east traffic. They only need to cross halfway before they're unblocking the other traffic. If we give priority to the left, meaning the west-east traffic, the west-east car needs to cross the entire road before they start to unblock.

I ride a bike for a few months in India. The honk system worked quite well. As a small vehicle, you learn your place on the road. Largest first, even if you honk. Honking at every turn in small town works ok, but it is loud. so loud.

Ah wow, this explains so much about the idiot who didn't know how to use a 4-way stop and nearly drove into me the other day. I thought he's giving me a hard time, blaming me, when I had the right of way. Maybe he was just doing the "warning because he might hit" thing?

In Asia it's mainly used to signal your presence, like when you're overtaking someone. Just a little tap of the horn. Just the fact that you hear it say in the right corner behind you will make you not swerve in that direction. It's almost subconscious and really does improve safety imho. You can't possibly visually scan for all vehicles around you.

The result of course is that there's a non stop cacophony, in places like Hanoi it REALLY gets to you after a while.

Here in EU if someone honks at you it's considered rude and will make me really react with wtf is your problem. Out in Asia it's completely normal.

It's been a while since I've been to Vietnam but most of the traffic is composed of motorcycle and not cars, so honking is indeed a signal of presence that's needed compared to just having the noise of a single car that's behind you.

(crossing the street is also kind of surreal as it's more like going through a school of fish; the trick is to walk at a steady pace to maximize your position predictability)

When I got my driving license too many years ago in Italy, they taught me that a brief honk when taking over in interurban roads is actually mandatory (but nobody did it). I don't know if that rule stayed the same as they harmonized more and more the road rules to the European ones.

Horns were disabled in Chongqing at least a couple of decades ago. I’m not sure what it’s like now though, but the government in China can and will deal with excessive horning using means we wouldn’t consider in the west.

They straight up banned horning in Shanghai when I was there in 2017 using some kind of camera system IIRC, it was quite jarring when I continued onto Nanjing and the usual cacophony resumed.

They also banned lane changing over solid lines with the same camera system IIRC

Ya, China has no problem using camera/recognition tech to its fullest. I wonder if other societies will eventually be outcompeted (cheaper law & order costs means Chinese cities will be more efficient) or if there is some huge cost (privacy violations that lead to economic consequences) to this that will make it less competitive.

Looks like London is on track to deploy similar technology. They did a trial in Croydon recently that was considered a huge success. From what I understand they actually ended up with pretty decent buy in from the locals as it had a noticeably positive effect on the community.

We'll probably eventually see it in more western cities as time progresses. Unsure what to make of it tbh.

Well I think it’s just like any other authoritarianism, benevolent with a benevolent leader but harmful with a malicious leader

Although things are a bit shit in many places, I do love our planet and the weird and wonderful ways it works.

Were car horns disabled (broken deliberately) in Chongqing?

20 years ago each city/province was basically its own closed market. So if you were driving a car in CQ, it was probably bought and even made in CQ. They simply required that the horns be disabled.

China internally is much more of a free market now, so I’m not sure how they could just disable horns anymore, although you still can’t get away with driving an outside register vehicle inside a city for very long without getting a crackdown by the police (meaning, they can enforce inspection requirements fairly easily).

I’m not sure if it was really Chongqing or some other obscure city like Dalian, I’m going by hearsay 20+ years ago. More recently, Shanghai banned honking in most circumstances in 2007 (inside its outer ring), but it’s enforced with just fines.

Was actually quite surprised by how "civilized" the driving was in our recent China trip. Don't think I heard a horn once in probably 20h or so as a passenger, did find the drivers up in Heliongjiang a bit bold with their lane weaving but in Beijing they drove great!

Clear rules, and consistent enforcement works.

Noticed something similar with littering, right now they have to employ an army of old folks to pick up cigarette butts. But I suspect once people come to expect clean surroundings that enforcement of littering fines can become a thing and the culture around respecting public spaces will slowly change. We even caught a young kid full on lecturing their grandparent for spitting on the street.

Dongbei drivers are famous for their interesting driving (but I’m going by hearsay and that one famous song about it).

I don’t think horns were used much in Beijing even on my first trip in 1999, although I do remember the Japanese guy driving us from the airport in a Jeep using it (and also seeing lots of city buses out at night without headlights on, you don’t see that anymore).

I just got back from Beijing a couple of weeks ago and honestly…the traffic is still very horrible but fairly orderly. Just too many cars and not enough roads (but it’s always been like that).

Yeh the congestion on Beijing's ring roads is pretty awful. Orderly but awful, also less EVs than ideal.

Have a friend from Shanghai here in Germany that had a really hard time getting a drivers license due to her old driving habits. Aggressively cutting in front of people and horning isn't looked upon too highly here.

Huh, I drove in Italy for a week and a half and didn't notice excessive honking. I did notice tons and tons of tailgating.

When I was there (ages ago), the driver of the bus I was on overtook on a blind corner on a road cantilevered off a cliff. They did cross themselves before doing it.

Honking is a harmless replacement for solving disputes with handguns. You probably drove in low-stress environments.

There is a low-stress driving environment in Italy? Where's that?

Milan is the only place I have ever been where reversing on the high way is a reasonable solution to missing an off-ramp.

A low amount of low stress people can be found late at night, on highways at negligible traffic hours, on the narrow and meandering country roads that everybody learns to avoid, in half-empty parking lots, and many other obvious uncommon situations.

oops, that might have been me. I kid. I only do that when completely lost trying to get out of an autogrill parking lot (!).

Italy is really 2 countries, north and south are quite distinct.

Where ever you are in Italy, you will be told by locals that you can't trust anyone from a town south of that place.

Even in Lampedusa?

What's the difference between them in terms of driving?

Try driving in Naples.

Naples is the first place I got honked at for not cutting the person off. I was at a stop sign, making a left turn onto a main road that had a steady stream of traffic. Apparently, I was supposed to wait a few seconds and then just creep out, cutting cars off on the way, which is what I believe the driver on the main road meant when he beeped at me and gesticulated wildly while I was sitting still at the stop sign.

That city was the first place where I saw a guy traffic splitting with a car. It was an ancient topolino, but my jaw dropped as he was snaking through traffic like on a moped.

Italy pales in comparison to Vietnam.

Lebanon. Especially Beirut. The honking. Every taxi driver that passed me if I was walking down the street. Honking. Honking. So much noise. All the time.

Six months I was there. Six months of honking honking honking.

Unfortunately like every city in Egypt. They have a hand permanently on a horn button.

I have experience of both Vietnam and India amongst other countries. The latter takes any country, including Vietnam, you can throw at it and wipes the floor with them when it comes to mindless honking.

and here you can get a 100 euro fine, for using your horn.

You can only use it, if its to prevent an accident from happening. that's it.

https://www.roadpeace.org/working-for-change/crash-not-accid...

> The memory goes bad because they write useless logs to a chip, and it eventually fails.

I worked for a $ ~billions revenue software storage vendor who had the exact same issue (excessive logging wearing out under-spec'd flash drives).

The bane of every cargo cult cloud op. I worked with a company that had maybe 20 devs total, > 30 "microservices" in kubernetes and one of the most complex bits of the deployment was handling Greylog and Elasticsearch. Still they couldn't manage high availability, despite logging all the things. Go figure.

I once worked for a unicorn that got near-zero traffic during the pandemic, but nobody could understand why some services were struggling to stay up.

Datadog was costing several thousand euros per month despite near-absent customer traffic. But the name made finally sense because all the data in there was absolute dog shit from reboots.

So yeah too much logging can be bad.

Oh most definitely. Maybe my sarcasm was a bit too subtle.

I definitely think that teams should think about what to log. Otherwise go with a live image kind of system like Smalltalk of LISP. The whole event sourcing paradigm and trying to just log everything and look at it later strike me as a poor reconstruction of that concept.

There is a tragic aspect to the "Worse is Better" essay that I see play out everywhere: there is a way to do something correctly but just throwing something together wins the race to market. Winner takes all and we're stuck with ossified bad decisions from the past. The idea that we can fix it later is just a lie. You can't do the foundation later, you'll be stuck with a structurally unsound edifice and forever holding it together under a completely unnecessary cognitive load.

Oh I got the sarcasm, I was just agreeing.

And I also agree about worse is better. To me the most tragic part is that "worse" has become almost as costly as doing "The Right Thing", mostly due to the extreme flexibility and rush to the market from vendors and libraries. Our foundations weren't as sketchy when the concept was invented.

It has definitely gotten much worse. The only thing keeping me sane is hacking solo projects in languages with great tooling. I don't think I can even stomach interviews anymore, let alone the whole application process farce.

I remember doing an interview with systems design using microservices and mentioning at the end "Well I guess that's it but if this was my personal project I would just have a single server and no native cloud bs".

The guy basically answered "Oh, same. I just ask for people to do microservices because that's how the CTO wants".

We had the exact same issue as well haha

These kind of problems only happen years after the software roll out so no one cares when you are under time pressure.

We sold physical hardware with bundled software, so we could actually create the problem via in-market software update that didn't exist at time of sale! Fun times.

HPE also had this issue with their ILO 4. New firmware fixed that issue but if your flash chip was already worn out you're out of luck and the only solution is to replace the entire motherboard.

Issue, or revenue driver?

Issue. We warrantied the longevity of those flash drives, and they were cheap anyway. The problem was mostly the customer pain.

You’re using a software fault which wore out the flash as evidence of poorly designed electronics?

How is writing excessive logs to a destined-to-fail flash chip in a car's electronics system not a poor design choice? Pretend the person wrote "poorly-designed electronics implementations/sytems" or similar, because that's obviously the intended meaning.

If the flash was better, the product wouldn't fail so quickly. It's really a combination of poorly designed electronics, and a software bug wasn't there, the fault wouldn't have popped up so early.

All solid state chips have a write limit

it isn't a software fault, it's a whole defective system that was designed poorly end-to-end: the software does something inappropriate, which the hardware cannot bear, probably because of a high level mandate to write too many logs and to be too cheap.

Just to add one more thing to your point, if embedded devs work really hard and make the code work faster/better all reward you get is an _even_ more underpowered chip for the next version.

Hardware procurement is cut-throat, sometimes they have mandates to reduce component costs and the procurement people WILL reach them. Often procurement > product in the power dynamics so no matter how bad the product gets those people still do it because the software gets the blame for bad product, not procurement who forced a bad chip to be used.

The infotainment is usually the #1 chip to be cut down because it is often the single most expensive electronics part in the system that can be "easily" swapped for a different part.

I hate the penny-wise pound-foolish attitude both in embedded and Android phone development.

For years now, Samsung has used a 'virtual proximity sensor' in everything but their premium stuff. Sensors like that are a few cents. Degrading the entire experience on the phone for a few cents cost savings. Say you do that for 25 components, saving 4 cents each. You've now saved $1 on a BoM of $100-$200, whilst making the whole experience of your product feel a lot worse.

I don't think that's about saving pennies as much as a reason to make premium models stand out.

Your account sounds accurate, but how fitting then that their cost cutting focus is losing them customers and potentially their entire company. VW is losing the EV war. Most manufacturers have already lost. Tesla and BYD are going to eat everyone's lunch. They either need to revolutionise their approach, or they're toast. I suspect they'll attempt to milk their existing supply chains into bankruptcy.

If the OEM stayed with the same chip for several years, wouldn’t the price go down over time?

Car OEMs are modern day colonial plantation owners, they know cost structures of suppliers and schedule their price cuts. They already have an annual cost saving quota. Prices don't just go down but go down just-in-time.

It would make zero sense if I drive to a Walmart and demand they sell to me with monotonically lowering prices as function of date since registration of my reward card, but in cars they do.

Ironically, this method of managing suppliers was perfected by Wal-Mart.

LOL. That's what I was thinking too.

I understand the concept, but the question I have is why?

These companies have huge wallets, and can surely scoop up a smaller automative microcontroller company and bring it in-house? It seems like a problem than enough money could solve quickly, but they've been doing horribly at this for decades now.

I work in one of the big three - the culture here is more waterfall and less agile. They decided at some point ‘we don’t need to be experts in building systems, we should only be good at spec’cing them and putting them together’ This leads to a mindset of relying on suppliers for changing even one line of code and at their mercy. Talent leaves because they didn’t get to do any of the fun stuff. And you’re left with bunch of MBAs trying to wing it in what is available which is - no talent, bunch of admineers, and a long list of supplier bills. They go for cheapest component they can spec for a given feature cutting 4MB memory will save 5 cents per car, we sell half a million cars, that’s big savings! I can go on and on about this, but one of us even tried to be Tesla trying to build our own zonal architecture - and are currently struggling due to costs, tarrifs and turnover. Also you can’t overnight change this mindset - building vs assembling. But there has to be some way and I’m too about to walk out the door due to ~10yrs of frustrations.

Get out if you can!

Spent 7 years at the three pointed star within design and UX - one day, when i’m over all i had to witness and experience i’ll write a book about the downfall of the german automotive industry.

It’s all politics and due to constant battles and changing ownership throughout departments they won’t ever have a solid foundation. And i dare to assume that this goes for most of the automotive industry.

It’s sad to see that a once driving force of innovation is stumbling over its own arrogance and ignorance.

A major factor contributing to this are cost saving measures from the early 2000s where most of them stopped in-house research and development giving most of the work to contractors - a very expensive cost saving measure long term.

We’re down to them using “technology” as a seasoning for consumption like a fancy restaurant - very little long term thinking.

Yeah, and then those contractors (like Continental) has sub-contractors (like Akka) and they have sub-sub-contractors (some random Indian software company) working on the side mirror winding logic.

In German cities with automotive industry, you’ll find thousands of these satellite companies.

And in Poland, Slovakia, Czechia, Slovenia, etc.

> downfall of the german automotive industry

I hear that kind of statements all the time but if you take like real important car things germans are (still) pretty good: their cars handle really well, powertraian usually works perfectly smooth (or sporty), ergonomics is good to perfect, it will not rust for decades, list goes on ... The real things killing germans I think: cars are expensive and unreliable

The main topic is software and it is still treated like a part you can just outsource and plug in.

Since cars are primarily being bought by sculptural aesthetics of the exterior and above all their brand they continue being bought for those who feel the need of a status symbol.

At the core there is still a lack of a long term strategy and above all stability to build on - not saying it is an easy task.

In the end the customer has to suffer with abysmal usability, reliability and ever changing mental models. And don’t get me started about the touchscreens everywhere situation…

It isn’t just software though - VW moving development and above all production engineering and planning to china since they failed coming up with an efficient solution in Wolfsburg is basically saying it all. [1]

Dire times ahead and i hope for the best.

[1] https://youtu.be/4AprfR8Xkio?si=xKXUdgt5BRyZKRNE

Given that the driving characteristics of most of the cars on the road don't match a BMW, what're the real "real important car things"? The revealed preference seems to be that the things you listed aren't actually that important. Long term cost is. Maintenance cost is. Not having to bring your car to the dealership for service is. Having the car have a long lifetime is. Handling well is nice; sporty drivetrain is nice; but that sort of stuff is clearly just a luxury, and the bottom dropped out of the luxury market recently (see: LVMH restructuring). If you're going to buy a sporty luxury car, why not get a Porsche or something with more cachet? Obviously there are reasons, but BMW's in an awkward position.

> what're the real "real important car things"?

A good design from engineering standpoint. You feel it just instantly when you use the product. Interior is nice and will accommodate just about every possible driver comfortably with every control reachable. Suspension just works frkn great no matter how it was tuned (sporty or comfortable) and no matter how simple the design is. Same for drivetrain, you will acelerate/decelerate precisely how much you'd expect. And it's not a luxury it's a norm, even cheapest german or french cars have all this things sorted out. I'm speaking for the EU market though)

Agreed, Germany still makes some of the best cars in the world. Who is making them better?

Huawei,Xiaomi,Xpeng.

For example

https://en.wikipedia.org/wiki/Maextro_S800

> They decided at some point ‘we don’t need to be experts in building systems...

So they've just chosen death. Fantastic, great to hear.

Well, yes. The legacy car companies are ossified. They want to keep churning out minute variations on the same cars, and regard software as a thin layer for the entertainment system. They don't want to adapt to EVs, which force a redesign of the car as a whole. They're going to get run over by Chinese companies unless they can beg for tariffs to prop up their un-innovation.

Isn't the trouble that agile is not compatible with things that has to be thoroughly made, 'finalized before release', like in every mission critical production? Casuality and the dyamic free spirit primised has much much less space here.

This is not sexy. This is important.

Needs different mindsets than the software folks grew up along in the past decades. Yes! Yes! There are much much more sexy topics to focus on for an agile software maker, that yields better looking results seemingly instantly. Compared to the boring finalization and coordination - oh, you devil bastard, coordination - heavy activities.

Don't take me seriously, speculating heavily.

> They go for cheapest component they can spec for a given feature cutting 4MB memory will save 5 cents per car, we sell half a million cars, that’s big savings!

I'm tired. Been out in the sun all day. Explain this to me please.

When I do the math I get 500000 * $0.05 = $25000

That's a small drop in a large bucket of their gross income or net profits.

EDIT: Harsh sun must've burned a few of my processors. I see now that this would only be one small change that saved an inconsequential amount of money. But each group is incentivized to produce minor changes like this that save small amounts and that those amounts do add to substantial savings and help complete the process of enshittification of the ownership and driving experience for those who choose to buy one of these vehicles.

Rinse and repeat across hundreds of components and your team "pays for itself"

"We found $X cost savings" is the easiest path the promotion. It's measurable, cleanly attributable, and immediate, while the downsides are not. Maybe perform is bad bc they skimped on memory, or maybe it's because the software team sucks. Maybe it means future updates are hamstrung, but who cares the bonus checks cleared years ago. Besides, you probably got promoted to a bigger / better role by now, and who can remember who decided what when?

And with the help of software your get: this algorithm works well to recognize signs using 2 cameras. We can alter it a little to make it work with 1 camera (huge savings) and losing like 10% accuracy. With a cheaper camera we lose again some accuracy but even more savings.

Now you get a shitty feature for savings while the people who implemented it can go cry in a corner thinking about their good version.

"We asked one of the software guys if we 'could' use a single camera, and they said, 'uh... possibly?', so we pushed through!"

Not just the owners, but the other engineers.

I have never worked in the auto industry, but I was an embedded software engineer at an F500 company that loved to just throw hardware "over the wall" to the SW engineers.

I had come from a very small company and working like this made no sense to me. After a particularly annoying discovery I was talking to one of the EE's and he explained it to me. "You see, the guy who designed that controller knows nothing about software. He just has a list of specs to meet, and he gets a processor, wires a bunch of peripherals to it, and releases a circuit board. If you're lucky, the SW guy who sat in the design reviews made sure to get a good enough processor to make your job easier. If not, you're SOL because as long as the hardware meets all the requirements they gave him, no one is going to want to change anything."

In this case, the engineer was incentivized to save a whopping $0.50 on a machine that cost around $2,000 to build. And for lack of that $.50 part, software spent hundreds of hours adding code to find a way to implement the behavior that it would have provided. Not to mention all the Test hours needed to verify that it worked as expected.

Paradoxically, I also saw the opposite behavior on the same project: people adding extremely complex hardware to solve simple problems because the company paid very well for patents, so of course everyone had an incentive to produce patentable designs.

That is one component in one model. Car makers have several models with maybe hundreds (or thousands?) of electrical components. Plus "cost-saving" has always been a surefire way of ensuring bonus.

Penny wise, pound foolish

It’s very obviously a rhetorical exaggeration.

Yes sometimes it’s a dollar or two and it really adds up quick. Sometimes 10’s of dollars. That door speaker can be few dollars cheap - you may get 2% more THD in a frequency band… the conversations can be really reduced down to ‘meh subjectively not noticeable’ but will save us a million. Add few of these things and now you have a shitty radio system but 5 mil in bank.

Speakers are a bit of a funny one, because, with the important condition that the amp needs to not be crap, they’re pretty straightforward to upgrade for the end user.

Yes but you make this small 5 cent change to 100 components and it adds up.

There have been attempts at it. Unfortunately, they consistently botch the execution so badly that most of the executives in the business have PTSD from the experience. And these were very expensive failures that become lore inside the companies. When they do acquisitions of small companies entering this market those end up getting smothered by the culture of the automotive companies.

Everyone has spent a mountain of money on this problem but spent it all assiduously avoiding addressing the root causes.

The answer is that current car platforms were designed with flexibility as first goal.

Car companies realized early on they could outsource component development and production to 3rd parties and they could make them bid each other to further lower the prices.

So their platforms were optimized to be able to swap component vendors very easily (to achieve lowest costs).

Of course the vendors are not 100% interchangeable and building a platform to accommodate everyone has to make sacrifices.Aka target the least common denominator across all vendors.

Then maybe they should let me buy some better damn chips so the experience isn’t so laggy.

I know, I know, shooting the messenger…

too bad computers aren't spark plugs

To be fair, this seemed to be the right strategy since they were able to be profitable in a very crowded market. Yes, the new companies try to verticalize everything from components to software, but none of them seem profitable (marginally Tesla passes the bar, but not so sure if you took away all the subsidies and carbon offsets).

So maybe the legacy guys were right all along?

I can tell you this works if your product doesn’t need frequent upgrades/updates and isn’t cohesive. In legacy auto world, you ask for one line of code change and the supplier slaps 100k bill. This is generally why things look old, outdated, carried over and buggy.

I totally agree with what you say. I am just not sure if the car market is willing to pay a premium to have this nicer fully integrated experience. Maybe there is space for a couple of premium makers.

My preference would be do less! Shift all the nav and entertainment to phone integration; stop trying to make half assed shit versions of those. How many billions were spent on that?

Oh we don’t want Google Amazon and Apple to have the cake and eat it too. See GM Rivian and Tesla not supporting CarPlay and AndroidAuto.

To what extent was a clean sheet design a huge advantage over the legacy makers?

And to what extent were the subsidies an advantage? They phased out after 200,000 units and Tesla has sold millions.

Tesla took the hit for the transportation industry, working their asses off and pioneering the costly ramp up to mass production of EVs, so the subsidies are the government compensating them for having not taken the easy path that the legacy auto makers are taking with their continued production of polluting gas cars and their half-hearted introduction of compliance cars.

Since government wants to encourage transition to sustainable energy, and oil and gas have been subsidized for decades, not to mention the tens of billions in bailouts for legacy auto, putting things in perspective shows that legacy auto should get the brunt of any criticism here, and the relatively smaller subsidies to Tesla are offsetting the larger investment Tesla has made.

The beauty of it is that the money is actually paid to Tesla by the legacy auto makers who have not stepped up or have stepped up only at a scale of virtue signaling, if you look at the sales numbers.

In Q1 2025, Tesla made $595 million on selling environmental credits/carbon offsets to other car makers. Net income for the whole company was $409 million.

Tesla has earned since 2017 over $10B in carbon offsets. That is in addition to the state and federal incentives.

How many issues due large companies run into thinking they can just throw money at it? Just look at google and stadia, or amazon and their failed game studio. They have immense money and knowledge and ended up with nothing.

Each car has dozens to 100+ ecus, written in different languages, by different teams, different requirements, and different companies. Some are proprietary. Ford can’t just tell Bosch, hey your abs module needs to now integrate with our api, multiplied by 100+ companies. The legacy car makers need to revisit everything, and move most of it in-house.

At the same time, we've had car companies putting out cars for 20 years with 10s of different modules built by different companies and things have been working just fine. Suddenly it's a problem because apparently everyone needs a giant screen on the dashboard?

Because the auto companies outsource everything, lay the risk onto the outsourced companies and expect that some significant percentage of them will go bankrupt every year.

With that kind of adversarial relationship, you are never getting anything above the barest minimum of competence.

Smells like Boeing.

I worked at BMW. I knew there was a project in there, using a certain ECU that was being quite problematic (as in, project being slightly late because ECU was a bit buggy and sometimes crashed when it was supposed to have almost 100% of uptime for legal reasons).

You ask: Why BMW doesn't just buy the ECU manufacturer?

Well... the company that was selling the ECU to BMW, is BIGGER than BMW. Even if BMW sold 100% of its assets and stock, it wouldn't have enough money to buy the ECU manufacturer.

The talent might not exist. Software development has been seen as the preferable career over electrical engineering for a long time now.

When I started my career I had a very keen interest in the embedded space, but when it pays half of what CRUD webapps pay I quickly changed to software only. I still tinker with embedded on the side and maybe at some point I can justify the cut in pay to go back to something I'd prefer to work on.

They don't have a culture that values it, at any level. Historically hardware was important and software was a nice-to-have addon cost center. That's the mentality that the people at the top are still in, and it trickles down.

> a smaller automative microcontroller company and bring it in-house?

I think in a lot of cases that would be Bosch, which is huge.

They did the opposite for decades in the hope to save some bucks, they outsourced everything so only business people remained.

Worse this really grew into a culture of entitlement where only a ready to use product is acceptable. There is no R&D anymore, there are people looking to buy solutions that don't exist for car makers.

This also works the opposite way. If the software roadmap does not inform the hardware requirements, then minimization of the bill of materials will lead to the selection of crappy hardware chips.

If you’re making very low end HW maybe this is true. Because HW is something that you put into the real world there are other constraints such as power, cooling, space, security of supply, ability to ramp, cost, reliability, etc. The calculus for HW selection is much more involved than simply SW. Good SW/FW can be performant on much less capable HW but it does mean that SW engineers need to understand more about the HW. This is a very rare skill in 2025. Most SW engineers I’ve encountered cannot explain stack vs heap. Furthermore even fewer understand how to use malloc correctly.

> Good SW/FW can be performant on much less capable HW but it does mean that SW engineers need to understand more about the HW.

It also takes much more time and requires a different set of talents. Often just using a bigger chip is better than investing the R&D.

The best analogy I can make is trying to make your own custom rendering engine and then code the UI in it or just use a browser and writing JS. Even if you do make it, your own custom rendering engine will probably cut a lot of features like fancy animations.

Blaming hardware people rubs me the wrong way.

People just use android and javascript front-end.

It's not crappy hardware by miles, crappy hardware as a category doesn't even exist these days.

It's hardware that can run everything necessary hundreds of times over, but shitty bloatland sloppy javascript it + android bloat it can not.

I feel like Subaru Eyesight violates this, which is why I’m so surprised with it. It’s a stereo camera system that just works so darn well. I’ve got to imagine the hardware that runs it is not insignificant.

This is weird because the microprocessor industry owes a lot of early success to automotive companies. Motorola 6800, Intel 8061 (https://www.intel.com/content/www/us/en/history/virtual-vaul...) etc. Quoting wikipedia: "the name "Motorola" by linking "motor" (from motor car) with "ola" (from Victrola), which was also a popular ending for many companies at the time, e.g. Moviola, Crayola"

TI has some powerful automotive SoCs like the AM69A/TDA4AH (https://www.ti.com/ds_dgm/images/fbd_sprsp79b.svg) that target the industry.. 8 Cortex-A72s, a full GPU, multiple Cortex R5Fs that can lockstep, and a bunch of powerful C7000 DSPs. The SDK is probably not awesome as embedded BSPs tend to be but the SoC should be workable. That should be plenty of compute.

So what is really going on, and what happened?

Motorola were a car radio company originally

I’m in a loaner 2025 Volvo right now and I’ve honestly been pleasantly surprised with the Android Auto setup. I thought I’d never again use anything other than phone projection, but nope — I can install Google Maps and Spotify and sign into both, and then my profiles and everything are right there including search history, and it’s actually more seamless and integrated than switching between CarPlay and the native/outer car UI.

Give it five years and it will be guaranteed garbage. Spotify will refuse to run on an unsupported older Android without the latest DRM API, while Google Maps will crash your system randomly, requiring you to disconnect the car battery to jumpstart it again. Volvo will offer you an upgrade of their proprietary device at the low price of $1899.

It's puzzling to see this push for general computing on devices that need to far outlast the typical release cycle of GC devices. There is nothing good that can come out of installing Android in your TV, fridge, let alone a - for fuck's sake! - a car.

If your consumer hardware needs to last for decades, then the core functionality and automation should be provided by sturdy embedded computers that are self-contained and do not require any kind of network access or regular updates, while the general computing functions functions should be provided by the user's own device or a replaceable/upgradable computer with a standardized interface.

I've been using Apple CarPlay on a car that was manufactured in 2016. There are some occasional issues with the infotainment system, but CarPlay works as well as it did nearly 10 years ago. It is much more likely that CarPlay will continue to function just as well whereas proprietary systems made by car manufacturers are going to start showing their age.

> Google Maps will crash your system randomly

They’ve at least got some incentive to keep this working so they can keep showing you ads.

> typical release cycle of GC devices

Now I have a lovely vision of the Android Auto device getting Garbage Collected when nothing depends on it.

Real life GC would be a fun project to see a geek movie of.

> There is nothing good that can come out of installing Android in your TV, fridge, let alone a - for fuck's sake! - a car.

Android Auto is not Android on the car, it's a protocol that allows an Android phone to use the car's system as a display, with limited UI integration.

The version in Volvo and Polestar's is actually Android Automotive, which is it's own Android distribution with it's own version of Maps/Spotify etc. Funnily it even has Android Auto functionality too.

Restated: “Android Automotive” != “Android Auto”

Although it didn't use to have Android Auto, which was awful. Because as you alluded to, Android apps are not automatically available on AAOS. As it happens the music app I use is not available on AAOS, so until Volvo added Android Auto support to their cars I could not get a good experience with my music.

Yes, Android Auto has some of the standardized interface features I was talking about, allowing the general computing needs to be fulfilled by a device brought by the user.

This is not what the GP is describing though, he's talking about the experience of a built in infotainment system running Android that can (for the time being) sync with his device.

Why did you think you'd "never again" use anything like Android Auto?

My own car is too old for Android Auto, but I sometimes drive a car that's from 2017 or so, and Android Auto works just fine on it, it's a pleasure to use (with the caveat that the phone has to be plugged in the USB port, wireless came later). So to me it seems like it always worked well.

Overall I’m a fan of the projection model, and I definitely see the benefit in longevity as well as the ability for older vehicles to get a retrofit head unit that adds in the projection interface.

My reflection was only that I was surprised at how well the built in apps worked when I tried them… but I definitely take it on board that it’s unlikely to still work this well 5, 10, or 15 years from now, so it’s important that the car still has projection available as a fallback.

This is only half the story, working for a major vendor, we sell both hardware and software, the whole way up to a full customisable well integrated platform. The manufacturers are deliberately choosing less capable systems, or taking thing piecemeal.

Most of our customers simply don't believe good interfaces are worth the money... They tend to either want either a set of features checked off (only for existence, not quality), or something along the lines of get as close to a rivian with thirty cents per unit more than we paid last year.

> customers simply don't believe good interfaces are worth the money

I guess I'm in the minority, then, but as a data point: I own a VW ID.4 and I'd pay significantly more to get software that isn't such a burning dumpster tire fire.

And no, the excuses provided in this thread don't cut it.

To be clear: it doesn't even annoy me anymore that the infotainment is slow and crappy, I've gotten used to it and I just never use it. But I when I want to close both windows and I press two buttons simultaneously, I would like both windows to go up, not one up and one down, as it sometimes happens.

The crappiness of the software in this car is mind-boggling and it cannot be excused: most of it is incompetent and sloppy programming.

I would pay more for a car where the software department is somewhat competent and knows what they're doing.

Well, consider, you could have paid more for a different car that has better software, like a Tesla, Lucid, Rivian ... but you didn't.

I'm not blaming you, I initially thought a VW ID.4 was a cool option. It just wasn't clear to the marketplace how bad the software was, and it's easy to assume "it's fine, I don't need fancy stuff" until you live with it and see how fundamentally bad the software is. How is the market to know? If it takes a couple years to figure it out, it makes sense for the hardware company managers to just make the hardware specs at the competitive price, and software is ... just whatever needed to get it out the door.

I worked for a few years at a sub-division of Samsung, and I've thought for a while about why "hardware" companies can be so bad at "software" ... in many cases, it's just that the leadership chain doesn't know what good software is and who is good at it. Managers don't really know what a good programmer is or does. Division heads don't know what managers are good at managing software teams and projects. And so on.

So at some point 2 years after the car is released, the CTO drives it and realizes that the software systems are fundamentally crap and can't be fixed, and it was not close or in-progress or anything, but he should have realized it 3+ years ago if he had good software sense, long before the car was released. And that's what happened with the VW ID.4

This is, incidentally, why it's so important to have a free market in software separate from hardware, despite what Apple may think. You can't have a free market competing on every possible feature; some features are going to be dominant. So people will choose cars based on size, aesthetics, price and brand .. but not on the quality of the software, which is very hard for them to evaluate even on a test drive.

> you could have paid more for a different car that has better software, like a Tesla, Lucid, Rivian ... but you didn't

At the time, I did not pay much attention to the software, because I never expected it could be so bad. Now I know better, and my next choice will consider software as one of the main factors (I don't be buying a swasticar anytime soon, though).

If VW and all other product manufacturers of products containing universal machines as components were forced to charge customers a 100% sales tax on all such end-of-chain products, UNLESS all (and I do mean all, down to the controller on the SSD or the battery controller or whatever) universal machines in the product complied with the following:

A) If there is stored code for a specific universal machine in question and the storage is re-writeable, and

B) there is a control mechanism in place to integrity check the stored code before execution, and

C) the integrity check mechanism relies on a cryptographic secret, or any mechanism which prevents the owner from changing the code but permits the OEM to, then

D) the specific universal machine's key store MUST permit full wiping of all keys in a way where no keys are stored anywhere (no permanent manufacturer keys), and the key store MUST permit the owner to store his own root keys; additionally, in the interest of national security and the average citizen's digital sovereignty,

E) replacement software/firmware for universal machines should be encouraged rather than stifled, so additionally there must also be technical specifications detailing enough of the hardware's architecture and the overall design of the part or product (the logic in making design decisions to accomplish product functions), to permit a skilled owner to write his own firmware and achieve similar functionality as shipped.

Basically, think Louis Rossmann gets together with Richard Stallman, and they form a beautiful baby governmental regulatory body to come up with "Apple Laws" (sic: Lemon Laws) to answer and address the Apple Question.

Abandoned proprietary code on abandoned proprietary hardware is a national security concern much greater than the minute problems caused by the occasional tinkering script kiddie. It will mean the end of the easy money of putting everyone on subscription, and would encourage more evergreen platform/API design to reduce developer-driven code churn. If companies want to make cheap proprietary throw away product which will house malware in a decade when the company has long abandoned patching holes in it, and design it so no owner has a practical chance or hope of fixing the vulnerability, then companies can suffer a price-doubling tax that'll go to pay for their open source competitors to more easily compete!

Sorry, not sorry. Get expertise producing material things people need, if what I outlined above would mean the high paid software gravy train ends lol.

There are other competitors for that segment, even the Q4e on the same platform has better UI. People still buy the ID4 because it's not enough of a deciding factor.

Hyundai/Kia make very good EVs. I am extremely happy with my new Kona. I would not say the software is amazing, but it is responsive, nice to use and has pretty much everything you need.

Amazing claim to read considering I have experience with modern Audi software on the daily and it is an utter dumpster fire.

I have a VAG ICE vehicle and had a problem with the navigation system not working. When I brought it in to get it fixed, they apparently put a completely new version of the software on the hardware.

Suddenly everything was fast. No slow lags anymore. System is ready even before I start the engine. Navigation now zooms smoothly. Voice recognition is finally working 95% of the time and only tripping up on hard words.

I don't know how many different software versions are out there but apparently they are working on system speed without changing the hardware. Maybe I got an early access version and they are waiting for data before they push it to all vehicles.

> I would pay more for a car where the software department is somewhat competent and knows what they're doing.

I have a Tesla Model Y and I was thinking of downsizing to an ID.4 and you just scared the shit out of me.

The ID 4 is coming off a recall and sales stop because the doors would open when in motion if the handles got wet.

I can't recall any car that didn't have any 'teething problems'. Some cars I've owned had multiple recalls. Of course it doesn't look good, but often it's to fix the probability of a problem occurring: it's not that the doors instantly swing open when touched by a drop of water.

Go for a Hyundai.

You only have to develop those interfaces once for high end cars and get your money there. Rest is then just one of the small modifications.

Not at all, a high end car will use an entirely different architecture to a mid/low end...

When you target a certain feature set it can make sense to use one big central processor, for lower end things it's more sensible to use limited smart sensors (from multiple vendors, for absolute cost minimums).

And it's generally not cost effective to move an old high trim platform down range due to changes in hardware and regulations.

What you mean by different architecture here? I've never seen a situation where manufacturers choose fundamentally different architectures between price points on the same platform. I feel like I'm misunderstanding what you mean though.

Someone like Ford for example will have several software platforms, some for low cost vehicles, some high, some that are adaptable between trim levels.

So as you go up in features on some model "the BigTruk" you might be going through variations of one sw platform, or jumping between platforms.

Some have several platforms for high and low cost based on centralised vs distributed, so for example an s class will not have much software or hardware shared with an a class.

Apple doesn't have different software platforms for low vs high cost phones. Why is a car different? It doesn't even have as much functionality.

I think it's fair to say that the software in a modern car contains lot more functionality than an average smartphone. Drivers just aren't aware of how much is happening in their car each second.

To someone that knows nothing about car SW architecture, that is surprising to me, I would have expected a number of control loops for things like fuel injection, ABS brake control, drive-by-wire, EV battery charge and discharge, etc. each running on their own processor due real-time safety considerations. These I would expect to be different implementations and parameterizations of the same control theory maths.

On top of this comes some functionality to control windshield wipers, lighting, AC, seat heating, etc. Stuff which is probably not top-tier safety critical, but still important. I would expect that stuff to run on one, maybe two processors.

Then comes the infotainment system, running on its own processor.

Sensors are supplying data to all processors through some kind of modernized CAN bus and some sort of publisher/subscriber protocol. Maybe some safety critical sensors have dedicated wiring to the relevant processor.

A lot of variations on this seems possible with the same SW platform, tuned and parameterized properly. The real-time safety critical stuff would need care, but is doable.

Am I completely off the mark? Can you give some examples of where I am going wrong?

I am also not deeply into this stuff. But there is more going on in a car than what you list.

One probably surprising thing is that an LCD dashboard is usually driven by multiple rendering stacks. One is for the complex graphics and eye candy. The other one is responsible for brake and engine warning lights etc. and is considered safety critical. The second one is very basic and often partitioned off by a hypervisor.

A lot of these controllers are running more than just control loops. They are also actively monitoring their system for failures. The number of possible failure conditions and responses is quite large. I had instances where e.g. the engine warning light came on because the ECU detected that the brake light switch was faulty. In another instance, I had powered steering turn itself off during a drive because it had developed a fault. These kinds of behaviors are the results of dedicated algorithms that are watching just about every component of safety critical systems that can possibly be monitored.

All of these software systems are provided by different vendors who develop the aplication software based on either their own stack or operating systems and middleware provided by other upstream suppliers. It don't think it's uncommon for a car to contain multiple copies of 3 or 4 different RTOS stacks. Nobody at the car manufacturers is enforcing uniformity in the software stacks that the suppliers deliver. The manufacturers tend to want finished, self-contained hardware units that they can plug in, configure and turn on.

I mean there is just no way that that can be true.

Because a low and high cost phone do essentially the same thing, whereas a high trim car will do things like steering assistance in a way the low trim does not do at all.

And to support the differences high trim will have different sensors and differently distributed compute.

This means that the infotainment system will be running in different places on different cars.

Yes, different platforms have different architectures. Within a platform, the system level architecture will be relatively fixed. OEMs will part subsystems out to different tier 1s for different vehicles on that platform, but that's (ideally) just plugging different boxes together on the OEM side.

There's a lot of very expensive development tools (e.g. dSpace simulators) that rely on this model of automotive development.

>Tesla, Rivian, and the Chinese EV makers

The iPhone on wheels paradigm shift has been stated like a decade ago and as usually the incumbents just can’t cross it while at the same time the new companies are successfully exploiting it.

Not surprisingly it coincides with EV transition - both are enabled by cheap electronics and EV voids incumbents’ ICE tech moat.

It was encouraging to hear an exec from Ford recently say essentially this in an interview. The legacy manufacturers seem to realize that Tesla is eating their lunch because of their lack of vertical integration. It’s not going to be an easy problem to solve but will be interesting to see what effort achieves.

BYD would eat their lunch even more if they were allowed to.

Tesla was eating their lunch in terms of software, integration, capabilities, apps. Then rivian came along and a few other companies doing a much better job than the awful legacy companies.

Now of course tesla/musk are destroying themselves through various idiotic actions. Sales are dropping through the roof. But the technical quality of the software ecosystem (car, web, app) is still better than all the incumbents. Think about Rivian getting a billion dollars from VW for their much better ECU and and software integration, for example.

I feel like Rivian is almost as good as tesla. Tesla still has all that, even as the company is in awful shape sales wise. Lucid seems to be better than the legacy auto, but I haven't looked into it as closely.

Rivians and Lucids cost tens of thousands of dollars more than 95% (not an exaggeration) of Teslas. Completely different markets (and size of market).

Was it the one with Verge?

>Their model of integrating 3rd party vendor computers just doesn't really work for this kind of thing; Tesla, Rivian, and the Chinese EV makers all manufacture all their own electronics, which lets them achieve the outcome. But you can not just roll all your own electronics in a year.

Maybe it's time for an 'OpenCar' project, where a "standard car" model is designed for (all cars have ECUs, light controls, HVAC, etc), and there's also a kind of natural demarcation that could exist like between drivers (engine performance characteristics, etc) and operating system (the overall "standard car" model). We don't write custom OSes for each PC make and model, why the flying f*** are car manufacturers all d***ing around doing their own things independently?

I think cheap China cars will finally kill the bloated US auto sector, and it will be a great time for the government to bail them out at a cost: they must design and manufacture parts to a national "open standard" in addition to any proprietary designs they choose to make. If they come up with a novel technology redesign for a part in the standards vehicle, the design must be open even if a patent for exclusive marketing of the improved part, as long as the part is not mandated. Automakers who don't participate don't get the competitive incentives. There should be a figurative x86/amd64 car, an ARM truck, etc. Think: volkswagens! There needs to be evergreen design in the standards cars: new parts made 30 years later should generally still fit, so it should have much looser regulations which would otherwise kill it off in a few years (like EPA regulations murdered the small truck).

It must be made much harder to put customers on the rentier treadmill. Planned obsolescence and proprietary design are two important tools to the rentier, along with copyright and DMCA. Look at China: better to strengthen your people and production even if it means chasing price gouging software houses off, because China demonstrated you can just steal the software in the future and improve upon it. What matters is the soil, minerals, metals, food, and production. People need materials to survive, they don't need frilly whirlie-gig flashy wazoo SaaS applications which cost monthly. Zynga's original business model should not be viable in an ideal world, but this is the world of the NPC and the cryptoshamanic advertising industry.

A nice thought experiment, but I doubt the US will ever do something as pro-consumer as this.

I had one of the most popular published projects on mp3car.com back in the heyday. It actually got me into a few strange meetings with companies that were actually building this stuff, like Clarion. I had literally designed, built, and installed a superior product to what they shipped as a one-off for fun in 3 months, but to them I was just a token enthusiast.

The vocabulary that these people started throwing out was absolute nonsense. It was pretty evident that "vehicle informatics" was fucked the second someone said those words out loud. And here we are more than 20 years later and still no closer to getting it right. Despite being seemingly well regarded, even Tesla's in car systems are just awful. And it's more impossible than ever to fix, modify, or replace on your own.

Agreed!

They also want to treat it as a new revenue stream rather than as a value add, which ultimately hurts them.

We end users don’t want to pay a subscription for our car. Especially for things we already get for free on our phone.

> We end users don’t want to pay a subscription for our car. Especially for things we already get for free on our phone.

I’m sure I’m in the minority, but I pay for ad-free navigation.

Sounds like a potential business opportunity! I don't know much about cars, how much is standardized in car electronics? Would it be possible to build a infotainment module that you could sell to several car manufacturers with only minimal modifications?

I think I've heard of something called an ICANN(?) bus that is used to communicate stuff in cars and is fairly standardised, maybe?

ITYM "CAN bus" - https://en.wikipedia.org/wiki/CAN_bus

Pioneer has been selling standard-sized Android Auto head units for over a decade at this point.

It's CAN, and it's old serial technology.

There are already companies doing 3rd party electronics as mentioned above, such as Visteon and Continental, and Garmin is trying to get into that business too.

That's what many OEMs have been doing for decades and this is exactly what many SDV have been trying to get rid of, since integrating many different products from many different manufacturers are slow, let alone iterating and designing new features.

Related to CAN, the bus is standard, but the thing is, CAN is just a bus, not a protocol. There are many ways you can have two ECUs (vehicle's modules) talking in incompatible ways.

Electronics are responsible? Really? Is this why the car radio interface lags and barely responds to input? Is this why the maps apps is terrible? Car infotainment systems are comically terrible even in areas that are 100% controlled by the OEM. Carplay works by reducing the infotainment screen to a dumb terminal. Car manufacturers could have done this themselves, you know.

I completely agree that vertical integration and building your own software stack from the ground up is the correct approach, but that's not the root cause of the problem. A better explanation here is that when all brands have awful infotainment systems then there is no consumer choice that forces competition.

As an embedded developer I usually point to the fact that there is generalist hard and software available for the primitive problem an infotainment systems needs to solve. At least for that side I don't see how generalist pc hardware wouldn't suffice and fit probably 95% of use cases.

At least that is how I build my self-made system, which is quite awesome compared to solutions you generally see in cars. Not for the average consumer, but classic car makers can do much better with a bit of courage.

How much extra work have automakers made for themselves by pridefully(?) refusing to use Android Automotive to handle some of the very things that auto makers are worse at?

As much as everyone used to clown on Tesla for it, the vast majority of cars would be better off with an iPad glued to the dash.

Volkswagen Up!: infotainment is just a USB port and a phone clamp.

For those unaware: the Volkswagen Up! is a small, low-budget car produced by VW group, it's also sold as the Škoda Citigo and Seat Mii. AFAIK it was only sold in Europe and Latin-America.

A family member had a early-gen Up!, and the OEM display (build by Navigon) that sat on top of the dashboard was removable, but used a proprietary connection, not USB. I believe it snapped on with magnets, which I remember thinking was quite nice.

The detachability was mostly for anti-theft reasons I presume, but quite quickly an aftermarket started to form to replace the OEM screen with other options, including phone mounts. I don't think VW envisioned that, but I thought that a detachable mount for aftermarket satnav, phone mounts or other accessories was quite smart.

I did wonder why they didn't just make it a phone mount as standard so you can basically BYOD, which could lower the price of the car further and probably be a better experience anyway.

> Volkswagen Up!: infotainment is just a USB port and a phone clamp.

Thanks to your comment I looked into it again, and I'm pleasantly surprised to see the newer generation Up! actually does have a OEM phone mount now, how cool! From what I just read it uses an app to integrate with some of the car's features.

More car manufacturers should do this for their budget cars. Have a few physical buttons for controlling built-in functions (namely HVAC), and let the user's phone provide the entertainment, navigation and other driving aids. Maybe even ditch the radio interface, and just have an amplifier and speakers build in.

It's a shame that phone OSes are moving away from on-device 'driving mode' in favor of Android Auto and Apple Carplay. I get it though, larger screen makes for easier controls and thus safer to interact with while driving, but still...

>>> But you can not just roll all your own electronics in a year.

Naturally, there must be some scale threshold where this is true, so I don't doubt your experience. And my workplace doesn't make anything as elaborate as a car, or with such stringent reliability specs. But my experience is that hardware is always finished before software.

FWIW in my experience building both, hardware is always finished first because it’s cheaper to change the software later in the cycle. Much like drywallers patching over electrical/plumbing sins, software fills gaps …

I mean the EEs / MEs can certainly turn out a board and housing in under 6 months, certainly on a yearly cycle. Though for the current automakers, they don't have a team for this so it would take them probably 3 iterations to get good enough to actually scale it to a mass market car, and that is if their team has good talent and strong leadership.

This seems like it is also a „purchase department got the cheapest crap instead of something reasonable“ problem. You don’t need to actually make your own electronics if the specs are decent and the features match your needs.

> But you can not just roll all your own electronics in a year.

Why? A year is a long time and it's a solved problem. In any case even if you allow the "a year is not enough" argument why didn't they start 5 years ago?

I’m not sure if you’ve worked around hardware but a year is not very long in these environments, and that 5 year plan is less like a sensible, let alone obvious step to take and more like a crazy leap of faith.

You don’t know that vertical integration will guarantee that you’re more competitive, and the investment you need to make before you see a return is beyond 5 years. That’s not an easy bet to make. It looks obvious in retrospect, but it’s really not.

It requires quite a bit of in-housing that many of these teams aren’t yet well-versed in, so as you vertically integrate you’re also disrupting your internal structure while adding new people. It’s a lot to take on. Meanwhile, there are other long term plans underway already.

Hm, reading this thread makes me realize that one of the reasons why Tesla/SpaceX/Starlink/Crew Dragon UI can move so fast is by using Linux all over the place.

Of course on itself it may not help, but along with other tricks like going agile with hardware does the job pretty well.

While others are doing their hardware iterations that last for years, software defined stuff may be easier.

Because they are not electronics companies, and further more they are terrible integration companies.

Unless the top of the company comes in and starts chopping every head that gets in the way of the new paradigm then it just ends up in locked up meetings for years of people that don't want to change.

Electronics integration isn't the problem, the people currently there are.

Precisely, such a change represents substantial risk in an incredibly risk-averse industry. People at orgs in such industries are in constant CYA mode, looking to point responsibility (and therefore blame) to anyone else.

The time to go and implement such a change probably pales in comparison to the amount of time spent in meetings getting people to agree to make the change.

It is possible to put out a fire by dumping cash on it, but there's a minimum amount that you need to dump at once for it to work. They cannot stomach the amount required, so they just feed it in one handful at a time, which of course just causes the fire to grow.

When I was working at $samsung_competitor, my NDA'd next gen android phone prototypes (a huge motherboard with a screen) were sent some years earlier. Like Samsung is on S25 now, and we would get boards for S27... It takes a long time for these things to evolve.

I feel like I'm on crazy pills sometimes when talking with people who deal mostly with software. I think SW engineers sometimes think that engineering generally looks like what they do, when in reality SW is a deep outlier wrt process...

The word “engineering” in SWE is just plain wrong. Present day software development has nothing to do with engineering outside of some very niche markets (aviation, mission-critical systems, embedded controllers). The term vibe coding came up really handy because it describes how 99% of software is developed much better than “software engineering“, with or without LLMs. That’s why it’s always fun to read such discussions of hardware vs software people.

When you design electronics you have to produce millions of, it takes a couple of dev 'spins' and usually a couple of prod 'spins' to get to the mass market board. Usually the PMs, EEs, MEs and SWEs get together and spec out a schematic, then the EEs will create the first draft of the board. There is usually extra connectors and test points on this board to ease testing and development. Once they verify it powers on, I as an embedded software engineer, start producing the software to get it running, or 'brought up'. While that is happening EE testing is going on for all sorts of things like EMI, power, communication speeds, etc. Besides the software I actually write, the chipset vendor's drivers need to be added and tested as well, there are always little things that take longer than they should. I've lost a lot of schedule to very subtle issues with chips. As we progress along, the schematic or layout gets updated and new versions are produced. Maybe the traces need to change to reduce EMI, maybe a chipset isn't workign well or we find a cheaper equivalent and swap it out. Then once everything looks good we move to a production version, all the test points are removed, and we start putting in orders for the parts we need in volume. If you want a million of something you usually have to order in advance. Then you start bringing the factory online, helping with factory test software...and well the point is the cycle time for all this is like 6 months for a tight ship. More like a year if the kind of thing your making is novel to the team since you need a longer dev time.

If you're curious why it takes longer than that, check out this primer on the HW dev cycle.

https://www.hwe.design/product-development-process/developme...

For components that have many components or complex requirements, or are part of more complicated systems, this takes longer. Cars have a design cycle that's many years long - 5-6 years would be a decent ballpark. That's due to the complexity of the product, complexity of the supply chains and tooling, requirements, and scale.

> In any case even if you allow the "a year is not enough" argument why didn't they start 5 years ago?

It's because these companies are more about vendor management and regulatory compliance than building things. It's a totally different mindset.

The VW and Rivian tie up for electronics will help answer if a traditional automaker can catch up if the electronics integration bit is taken off the table.

Back when there were all the rumours of an Apple car, I was hoping that this was actually what they were working on.

> you can not just roll all your own electronics in a year

The Model S came out in 2012 so they’ve had well over a decade to catch up.

Why not just use iPads or some Android tablet? Why some shitty embedded system and not a real computer?

Are the PM women or not qualified?

Reply on news.ycombinator.com