I’ve had a little more than a day to absorb the news that Volkswagen is going to put billions into its erstwhile competitor Rivian in order to gain access to its software platform. It’s bad. Real bad. It’s obviously not great for Volkswagen, which appears to be in the middle of the kind of Wolfsburgian struggle that was supposed to eventually stop. It’s worse for other automakers, though, as one of the biggest and smartest of them shows it’s not ready for a software-defined universe.
Remember all those Lidar SPACs that bubbled up during the pandemic? It’s ok if you don’t; that whole time was pretty dark and I’m sure Lysol-ing bags of wasabi peas broke my brain, but there was a sense that all cars would need Lidar and, combined with shortcuts to going public, these companies would be goldmines. What actually happened? So far, Chinese suppliers dominate the industry.
China also dominates batteries, though it’s not the only player. South Korea has a large battery industry and a major company there experienced a scary runaway fire that claimed the lives of 23 people.
And, finally in today’s installment of The Morning Dump: Will the CDK Global ransomware attack materially impact June sales? People don’t seem to agree.
Rivian, VW, And The Cariad Fuckdown
A recent installment of The Morning Dump Saga included a breakdown of various CEOs and one of the interesting points made in a linked article was that maybe VW CEO Oliver Blume should let go of Porsche because he seems a little busy. I disagreed and I still think I do, but there’s one funny bit from that Bloomberg Opinion piece I think that’s worth highlighting:
[R]unning VW is one of the toughest corporate gigs around: It involves overseeing 10 brands and almost 700,000 employees (many of them unionized), while mollifying the Porsche and Piech families and VW’s home state of Lower Saxony, who together control a majority of VW’s voting shares.
In fairness, Blume has so far shepherded these various constituencies rather well: Relations with the trade unions have improved, and he has the backing of VW’s family owners.
There has been less of the smell of rot coming from the State of Wolfsburg (well, technically, the state of Lower Saxony but that’s less poetic) lately. The bit of obvious brawling has come from the parts of Volkswagen unhappy with the company’s overarching software unit, Cariad.
Volkswagen, like most automakers, realized a few years ago that what Tesla was doing was shifting the desirability of cars purely from their stats/comfort/style (hardware) to their added electronic abilities (software). This is how we get the industry term Software-Defined Vehicles or SDV.
At the time, VW did what a big company is going to do and created a company called Cariad (originally Car.Software, which, lol). The company staffed it up with thousands of engineers to build a software platform for all Škodas and SEATs and Porsches and Bentleys and everything else.
Believe it or not, the move-slowly-and-don’t-break-things approach of VW didn’t work. The company lost billions of dollars and produced software that is fine, I guess. Despite smoothing the waters elsewhere, Blume was reportedly not happy with Cariad from his time running Porsche and this carried over into his term as CEO of both Porsche and the wider VW Group.
From a Business Insider piece last year on his ouster of Cariad’s leadership:
From the outset, Cariad was seen as a dispute between the brands, and produced mostly negative headlines. It was about wrangling over competence, schedules could not be met, delays kept occurring, and costs exploded. The software debacle was a major reason for Herbert Diess’s replacement. And for his successor, Cariad is probably the biggest construction site in the group.
With the announcement that VW would be giving Rivian a few billion dollars for access to its software, it seems like Blume has all but given up on the idea of Cariad being in control of SDV at VW.
The timing is interesting, as there was a recent report in Germany’s Manager Magazine that talked about Blume traveling to China for the Beijing Motor Show earlier this year to drive an Audi with the Cariad’s next-gen software and, well, it didn’t go well:
During a group acceptance drive, new models were tested, compared with the competition – and the boss was not very enthusiastic. Blume’s reaction? From the Audi people’s point of view, a “catastrophe!”, as one participant groans. The new Q6? “For Blume, the last in line,” says another.
It wasn’t just the design that was disappointing, as was the interior, presented by Gernot Döllner as a sparse basic version and not exactly luxurious like China. The infotainment system in particular was not up to scratch. Once again, the group’s own development subsidiary Cariad was unable to deliver anywhere near what was expected of it. Instead, it lived up to its reputation as the group’s problem unit.
When Rivian announced the VW deal there was a lot of surprise because, so far as I can tell, both companies managed to keep the secret quite well.
Reuters has the best background on this, and the timing, again, is what I care about:
The talks that led to the dramatic tie-up began, Scaringe said, when he and Volkswagen CEO Oliver Blume met privately at Porsche’s experience center in Atlanta.
Two sources said the meeting was in August last year.
Ahha, tell me more:
The companies got to work straight away, with a Rivian team visiting Volkswagen in Germany that fall.
The testing to make sure everything worked together was “like a scrimmage,” Scaringe told a company townhall on Wednesday, according to one source. Another trip to Germany followed early this year with lawyers and software experts, this person said.
[…]
To overcome the difficulty of integrating starkly different work cultures that often plague such deals, Volkswagen leadership agreed to embrace Rivian’s agility, its software chief Wassym Bensaid told analysts on Tuesday. He said “very clear rules and responsibilities” had been set for the JV.
Like I said in the headline, this is a bad sign for big automakers. Volkswagen may be big and lumbering, but it’s not dumb.
I just think the big-and-lumbering is kinda the problem. I don’t develop software and claim no expertise there, but with SDV it does seem like the smaller and leaner Silicon Valley model seems to consistently outperform what comes out of Germany, Japan, or Detroit.
VW threw billions of dollars and thousands of engineers and coders at the problem and, basically, gave up.
China Controls 80% Of The Lidar Market
Lidar, which is a sensor technology that fires lasers at things to understand where they are in near real-time, is a big part of the driverless car future that’s supposed to be coming. This was especially the case in 2021 when a bunch of them used SPACs to shortcut the IPO process and go public.
Here’s a fun Forbes article from the time: “How Will Pure Play Public LiDAR Companies Use Their Money?”
The companies listed were Velodyne, Luminar, Innoviz, Ouster, Aeva, and Aeye. Velodyne and Ouster had to have a “merger of equals” and the stock is not even 10% of its pandemic peak. Aeva’s share price is in the $2 range after being up near $100. Innoviz is less than a $1 per share. Luminar is a little less than $2 a share
As the article points out, at the time the range of values for these companies was between $1 billion and $2.9 billion.
What happened? For one, the autonomous car bubble burst with companies like Argo going bust and players like Cruise hitting a bunch of speed bumps. What else? As S&P Global Mobility points out, China also happened.
Specialist suppliers from mainland China dominate the lidar market. Their dominance is a by-product of the rapid advancements in advanced driver assistance systems (ADAS) and robo-taxi applications in the Chinese market, supported by benevolent government policy and subsidies. In terms of volume, the country accounted for more than 80% of global lidar sales in 2023, according to S&P Global Mobility data. Having found a large customer base in their domestic market, the Chinese lidar suppliers are actively looking to expand their presence in international markets to stay on a strong growth trajectory.
When China wants to do something it’s hard to compete, though S&P thinks there’s a future here for some of these companies:
And what about the lidar companies that went public via SPACs? Their time to shine is coming. The likes of Luminar, Innoviz and Velodyne are expected to gain market share as we move through the decade and as North American and European markets follow the >L2+ lead of mainland China, driven by evolving regulatory environment and rising consumer demand.
So it’s not all bad news.
Scary Lithium Battery Fire In South Korea Kills 23
Just FYI, the video above shows the start of a runaway lithium battery fire in South Korea this week. While no one gets hurt in this video, it’s not possible to know if any of the people seen attempting to extinguish the fire perished in the blaze.
If you don’t want to view the video, it basically shows how quickly a fire can happen in this kind of environment and raises a lot of questions about how the company stored these items.
There’s been some confusion online, and I’d like to clarify here that these are not rechargeable lithium-ion batteries used in cars but instead non-rechargeable batteries used for consumer electronics.
Still, it’s a reminder that batteries can store immense amounts of energy. According to The Korea Herald, the government is ordering inspections at all kinds of battery plants in the country:
As such accidents could be prevented with the right measures in place, the government ordered more than 500 battery manufacturing workplaces to conduct emergency self-inspections based on a lithium handling safety checklist on Wednesday. In addition, fire authorities are conducting urgent fire safety investigations with relevant ministries for more than 200 battery-related companies.
As the article points out (and as we’ve experienced with fires here), once lithium batteries get going they take a lot of work to extinguish.
Are June Sales Going To Be Materially Impacted By The CDK Global Ransomware Attack?
We know that the attack on CDK Global, a major provider of essential dealership software, was a ransomware attack and that it’ll take at least another week to get services fully online.
As this happens, dealers have claimed they can still sell cars, albeit slowly. Will this show up in monthly sales reports?
A recent release from S&P Global Mobility estimated the seasonally adjusted annual rate (SAAR) at 16.2 million in June, which is about 1.40 million total sales. That would be an improvement over last month and over June 2023.
And yet, J.D. Power and GlobalData say that might not happen (note: they have slightly different estimates from S&P so the numbers don’t perfectly align). From Automotive News:
J.D. Power and GlobalData estimate that U.S. light-vehicle sales for the month will fall by 2.6 to 7.2 percent from a year earlier, for volume of 1.27 million to 1.33 million. They originally had expected sales of 1.41 million before the ransomware event that began June 19 created a “significant” disruption.
This is extremely complex stuff and it raises a few possibilities:
- Nothing happens. Dealers are able to figure out ways to get cars off the lot.
- Sales data is delayed because dealers can’t file sales easily with automakers, so those numbers show up in July.
- Some automakers use this as an excuse for poor sales (the Stellantis approach, perhaps) while others show no downturn.
- The apocalypse.
- Cats and dogs, living together.
It’s the end of the quarter in three days so we’ll know early next week what happened or, at least, what automakers say happened.
What I’m Listening To Today
I’m still rocking the Shoresy Season Three soundtrack and this popped up and it’s been on repeat.
The Big Question
Can a large automaker be responsible for building its own software? Will GM pull it off? Will no one other than Tesla make it work?
“Can a large automaker be responsible for building its own software? “
Yes. If a smaller one like Tesla can do it, then there is no technical reason why a larger one couldn’t do it.
“Will GM pull it off? “
They will eventually… which is the typical GM way.
” Will no one other than Tesla make it work?”
Tesla’s success with their software is the result of work that goes back to 2008 (or maybe even a bit earlier).
There was a lot that was learned and refined over that 16 year period. How long have all the other car makers been at it? Not for even half as long.
Give it some time and the other car makers will get their shit together at varying speeds and to varying levels.
Sidenote: Software expertise is one of the reasons why I think Apple should buy up the remains of Fisker and the Fisker Ocean design.
I think it is too late.
The real disruption is self driving cars.
If a company doesn’t make cars that drive themselves with no supervision in five to ten years they won’t survive.
And it’s not just a matter of being able to drive themselves occasionally in predetermined conditions in top line models. It will need to be most of the time everywhere in ordinary cheap consumer vehicles.
I hate to say it, because I don’t like Musk. But Tesla is now so far ahead the only possibility for OEM survival will be up license the tech. But at what cost?
Nah… it’s never too late for a company to get serious about competing… unless they go bankrupt and get liquidated.
So it might be too late for Fisker. But for the other legacy OEM companies in automotive? Nah… they still have time and plenty of resources to get competitive with Tesla as well as the Chinese.
And speaking of the Chinese companies… there is a lot of fearful talk about how cheap they are and how they’ll destroy everyone else.
The same was said about Japanese cars in the past… and about Korean cars in the more recent past.
But I’d like to see how well those Chinese vehicles hold up in the rust belt. I’m gonna guess like the early Japanese and Korean vehicles, they won’t hold up too well.
And as a result, many of the current Chinese automakers will die and some will survive. Same thing happened in the Japanese and Korean auto industry.
Everyone looks at the success of Honda, Toyota and Hyundai.
They don’t remember Ssanyong, the original Kia that went bankrupt (and scooped up by Hyundai), the original Daewoo, Kurogane, Asia Motors, Keohwa, Shinjin and how Mazda almost went bankrupt with the rotary engines and how both Nissan and Mazda were in serious trouble with the Asian financial crisis and would have gone under without help from Renault and Ford respectively.
Tesla and some of the Chinese makers like BYD may be far ahead. But no one company can be all things to all people in all places.
So a company can still do well without competing head-on with Tesla by offering BEVs in a segments that Tesla doesn’t compete in… like vans, consumer-grade pickup trucks, off road vehicles, highly luxurious vehicles, etc.
Or if a company offers something that goes against the Model Y, they can mitigate a tech advantage by offering stuff that Tesla doesn’t offer… stuff like a broader colour palette, a CD player, traditional AM/FM capability, a spare tire, a design that is cheaper to fix/cheaper to insure, velour/cloth seats, etc.
And while self-driving vehicles is a big thing, I don’t believe 100% of all new vehicle buyers will want it… at least not in its current state.
I know that I’m not interested in self-driving until I can legally and safely have it drive me home while I take a nap in the back seat. In my view, self-driving is mostly pointless if I still have to pay attention.
I’m a silicon valley software person, though not in auto. I disagree this is bad news for the legacy automakers. Instead I think it’s bad news for their software divisions, but that just means the OEMs can outsource that work to specialized experts or acquire software startups, and focus on the metal. You know, the same way they have Bosch and battery makers and a bazillion other suppliers.
Blume seems like a positive change for VW. He’s also on the record dissing the user interface in modern VWs with no physical buttons.
I think a lot of auto makers miss the mark on software like Tesla because they bean count the infotainment and don’t go all in on hardware. Tesla’s software has it’s flaws(see early Model S’s where the excessive logging was killing flash drives).
But Tesla at the beginning knew they were going to need powerful hardware, mainly as they put EVERYTHING in that friggin center console, and it had to be responsive, and last a while for planned over the air updates, and Elon’s all about the tech and such, so that’s where they put the money instead things like buttons and switches.
Rivian actually does over the air updates like Tesla, their partnership with Amazon gave them a leg up on software engineering, so it’s a good idea to try and get some of that, but the question is will VW actually pay for the hardware needed to run good software? Maybe on a Porsche but I can’t see them doing it across the board.
When you frame the mission of an automotive software division as, “Make people pay monthly for things that used to be free, and resell whatever personal information you happen to acquire from them,” it’s not surprising that their output is awful. Building software is an inherently creative intellectual exercise. Creative people chafe under the yoke of awful management, and a “Software-Defined Vehicle” (barf) in its current state is not an attractive project for clever and inventive developers or managers.
If I can put on my old-man hat for a second, I’d go so far as to say the entire concept of an SDV is absurd. If we accept the premise that cars shouldn’t be defined by their engines, tires, suspension, cabin, and so on, then it follows that the thing that distinguishes luxury from penalty-box is invisible and intangible!
Is there a single piece of software in your life that is worth TWENTY THOUSAND DOLLARS more than another piece of software? Especially when said software enables the operation of something that, a mere twenty years ago, worked just fine with hardly any software whatsoever. It feels like they determined that the more of the vehicle’s price can be justified by the software, the greater the share of development cost that can be moved to cheaper departments.
When you think about it too much, it’s all just deeply disturbing. It suggests that the experience of driving a car should be more closely related to the non-car things that happen inside of it. It makes me feel like deep down, what consumers want is to spin the wheel around like Maggie in the Simpsons intro while an actual adult drives the car.
Car companies are getting $735 a month (quick search’d average car payment, 2024) and they’re like “let’s piss off all our customers to an insane degree get an extra $11.99 per month!!!”
Like, look at the big picture business boys.
“Software-defined x” (the phrase) comes from the tech industry, specifically “software-define networking” (SDN). SDN just means that modern network routing, firewalls, subnets, etc is now primarily done and configured in software, rather than by using dedicated physical routers, switches, load balancers, firewalls, spam appliances, etc. Of course, that physical hardware does still exist, but more and more companies don’t ever buy, lease, or interact with it at all, and instead they let cloud providers manage the hardware transparently.
Cars went from all-physical with electricity, then they added relays and solid state components (1950-60sl, then some dedicated electronic controls (E.G. fuel injection, 1960s-80s), then onboard computers (1980s), then x-by-wire (transmission, throttle, brake, even steer) with software-defined options like ride style (“sport mode” and the like).
The “SDV” term is just trendy jargon to refer to a waypoint on a continuum of technical evolution that’s been moving towards less hardware coupling and more electrons since the 1950s.
It’s not synonymous with subscriptions and disabling existing hardware unless you pay extra. Sure, it is certainly possible, but that’s not what SDV means. And the benefits are considerable: a recall can be an OTA update, no need for a trip to the dealership. Virtual BOMs (the bits and bytes in your car) aren’t fixed and immutable for years the way hardware BOMs are. Software can make your car can improve over time. Without software, a 3-y-o car will never be better than the day it rolls off the line. With software, it can actually be better 6 months later. That’s objectively beneficial.
Of course there are a LOT of well-known and unknown issues with software. Software engineering is not as mature a discipline as real engineering. So it’s not all roses.
But SDV isn’t a bad thing, and the phrase is really more marketing than anything.
Also: yes, some software is worth $20k more than some other software. The manufacturer has to recoup the real costs of building the extra features and the infrastructure needed to support them.
(Disclosure: I’ve been writing software professionally for a long time)
I would submit that there is a vast chasm of difference between the software found in ECUs, ride options, and so on, and the stuff that frustrates us in our infotainment screens. I’m not an expert in ECUs by any means (I work in telecommunications) but my understanding is that even modern ones are similar to embedded systems inasmuch as they are programmed in assembly or a low-level language and have very few steps between accepting an input and producing the correct output – as far as I know each program in them consults an input value against a data table and returns the lookup value along with possibly the results of a calculation. In other words, no more than a few dozen machine cycles. They also presumably need to be certified and have critical time constraints. If they don’t produce an answer almost immediately after receiving an input, the operator of the vehicle could be in danger.
Then compare this to the utter clusterfuck that is Android Auto. Your phone, an advanced microcomputer running a consumer-oriented multitasking operating system, is supposed to exchange a large volume of multimedia data with your car’s infotainment system, which is probably also Android or similar. Both computers have other tasks to schedule in addition to data exchange, and they both have to monitor for, buffer, and act upon user inputs. This arrangement is woefully unable to meet any kind of serious time constraints – there’s a noticeable lag when I touch buttons on just about any touchscreen, phone plugged in or not.
But the selling point of the “SDV” seems to be on the user experience side of the equation. We had totally functional and even quite advanced drive control systems before “everything in a huge tablet” was popular. It must follow that the difference is in the infotainment, environment controls, and so on. It feels like massive pig-in-the-poke, because the automaker is saving money on physical components while providing us with an inferior alternative! Like, yeah, it looks cool, but it doesn’t work!
Anyway…. old man hat off.
(edit: I’ve said it before and I’ll say it again, there needs to be a company that specializes in in-dash computer systems, because the current approach isn’t working. I just don’t believe that consumer software and off-the-shelf electronics are the right paradigm to deliver the kind of performance that people actually expect from a vehicle dashboard. I thought Apple could have done a great job if they didn’t get lost in trying to make a whole-ass car. Whole ass-car?)
While the OEM’s are trying to catchup to Tesla on the infotainment screen and OTA updates( they are more than ten years back) , Tesla designed and installed, into millions of vehicles, its own computer vision AI inference chip and building the largest AI training computers in the world. They are on hardware 4.0 with hardware 5.0 coming in 2026.
Its really beyond a joke at this point
It’s not just software. Look at VW’s current interiors. They’re aging rapidly, and the pieces that are newer are ugly and unergonomic. They have a holistic usability problem stemming from a lack of clear design leadership. Plugging the software hole alone isn’t going to fix the problem. They need to appoint a design and usability chief with responsibility and authority for the entire experience.
I have worked with real software developers and I have worked with engineers who could ostensibly write code and there’s a difference. I suspect VW went the latter route.
As an engineer who writes code, I can confirm your suggestion here.
I am not a software developer!
And…. I know this. I would never offer up my services on anything more complicated than an end-of-the-line tester or data logger GUI. Software, like that found on a car, absolutely needs to be written by experts in the field, not hobbiests, which is what every “engineer who could ostensibly write code” is, like me.
As a software “engineer” I can also confirm most of us are acutely aware that our discipline isn’t really engineering except in a few sectors.
All this stuff is too complicated for a person to be good at software AND mechanical engineering. An orthopedic surgeon wouldn’t try to treat psychiatric patients, and vice versa.
relevant xkcd: https://xkcd.com/2030/
Accurate.
As a former software tester and current Agile Coach, I can confirm that developers can be very good at creating a ‘tree’, but will, absent great leadership and clarity of purpose, miss out how that contributes to the ‘forest.’
While I think it would be smallest of Venn diagrams, I would argue that VW needed ‘car guys’ that also developed software under a visionary leader that could seamlessly turn strategy into tactics. The description of VW in this article doesn’t feel to me like they had those things.
Yeah, the “engineer” term gets thrown around a lot. It can be redefined as “problem solver” as far as I’m concerned, cause that’s all we all do, name be dammed.
To be honest, I wouldn’t even put Tesla in this list, given the recent blather about Tesla being an “AI robotics” company and not just a mere automaker.
Not sure, but VW made some very good emissions software.
Even then they had to rely on a supplier, but at least they could agree on the brief.
…take my upvote 😀
GM? lol. Lmao even.
Another pump and dump, aka Tuesday on Wall Street
The Mythical Man-Month was published decades ago and still these giant companies can’t manage software.
Edit: ninja’d by David W Alderman!
This. Came to post it, left satisfied
TBF *nobody* does software properly.
*Chris Sawyer enters the chat*
“I wrote RollerCoaster Tycoon by myself in Assembly”
—–
Which, I think, is about as lean as you can get.
That just brought up a ton of memories
If you think throwing developers at a problem builds good software faster may I recommend a classic book called “The Mythical Man-Month” by Fred Brooks.
I don’t see the automotive approach of “design by committee” working well in the software industry. My immidiate answer is “NO”, but if they can remove that barrier, I think it’s possible. Both groups solve problems differently, and you need to allow them to do what they do, and not force them into a poor fitting mold.
I did a brief stint at a software company (it didn’t bill itself that, but it is essentially what it was) as a consulting engineer. The company was built around the SaaS model, but the UI was so nonsensical that they had to hire folks like me to be the interface between the customers and the software. My first meeting with the UX team illuminated why – it was a design committee who had no clue what their client’s knew, how their client’s worked, and designed everything around “capability” (a term no one ever defined). It wasn’t until the company was on the brink of folding that they got rid of the UX leadership committee and appointed someone formerly in a position like mine to focus the UX team and fix the usability problems (which ended up being too little too late, hence why my stint there was brief).
Rivian seems like it can think nimbly without over-thinking things, which is a common trait I’ve seen with a lot of German companies I interact with professionally, so it could be a good influence if VW allows it to be. However, that’s up to VW to allow the influence, and a ship that big doesn’t turn on a dime.
A large portion of the usability problems comes from companies building feature creatures. Product management teams load up all sorts of pointless but ‘clever’ crap to the point where the basic function is too obscured to be used.
Thats the way its be misdone for decades in my experience.
I accidentally pierced a battery cell trying to remove the battery from my old 2012 MacBook. I’m glad I had some idea what to do in that I didn’t douse it with water, but that was pretty scary and intensely hot. It also has a unique smell.
Funny enough, all it did was scorch the aluminum on the user-facing side of the keyboard, which I wiped up and then carried on happily for several more years.
After a brief search that didn’t turn up much recent news other than China levies some sort of fee on Chinese nationals working overseas, I don’t have much, if any, clarity on how tariffs on China would apply in these situations.
If a company based in Indonesia, Laos, or maybe Easter Europe uses primarily cheap Chinese labor, is that something that should be looked at by the countries enacting tariffs against the mainland? I don’t know, but it seems awful loopholey if it’s not.
As for the question, I think GM has mostly always done well from a technical standpoint. The problem for them tends to be the products they are packaged in. So, yes and no?
So does the US. It’s the only major country that taxes ex-pats, regardless of country of residence.
Tajikistan isn’t a major country?
142/177 Economically? I’d put them on the outside.
Their parents must be disappointed.
Expats, not ex-pats.
Expat is short for expatriate (one word), not people formerly known as Pat.
Carry on.
Why do you hate Pat?
like Tom Brady, he would be an ex-Pat
Regarding “What I’m listening to”: I have absolutely no idea what song this is. I tend to read TMD at work and youtube acts… wonky. Could y’all put a link or at least some text describing the video/song in question?
Sorry! Sunday (feat. Pouya) by Danny Towers
Thanks, Matt! For what it’s worth, this actually happens a fair amount on the Autopian across all writers and it would be nice if checking for text descriptions and/or links to the source video was part of the process.
I must be getting old. So much of what the yute are listening to these days confounds me. First everything was slathered in autotune, then there are feat all over and they are dropping things all the time. Are kids getting clumsy or something?
VW is really signaling that their meticulous German analysis determined paying Rivian $5 billion costs less than another generation of craptastic software and buys them several more years before trying to do things in-house again.
If software companies can do it better, what happened to the Apple car? I think it’s more a problem of company size than capability. Big means bureaucracy, long lead times, and slow reaction times, plus budget badgers at every step. Tesla (used to be) Jack, as in Jack be nimble Jack be quick. Not so much anymore. Probably still better at software than the old guys and gals, but that advantage is eroding with each billion in stock value.
I don’t think software companies can do it better- see Apple and all of the Google/Uber self-driving nonsense. Rivian is not a software company, they are a hardware company that has to develop software as a component of it’s final product. They just happen to be really good at that development, probably because they properly integrated software development into their whole product development lifecycle instead of tacking it on as some sort of extra that every other legacy car maker seems to be doing. This deal seems like VW paying 5 billion in consulting fees for that sort of integration.
I also don’t think the size of the company necessarily determines how responsive they are, as I’ve definitely seen major fortune 100 companies jump on opportunities and change direction faster than any startup could have hoped to match (having employees in all 24 time zones and long-standing political contacts are a major advantage when you can properly coordinate them), it mostly comes down to the leadership culture and whether executives have the authority/autonomy/balls to do things as quickly as they need to happen.
This is 100% correct imo.
Also: there’s a giant rift between what the Apple and Google cars were rumored to be and workaday car operating system and software features.
The Apple car was supposed to be an autonomous vehicle. That’s orders of magnitude more complex than the software than runs a car’s operations. The former has never been done before; the latter is mostly old hat with some shiny baubles attached to it.
So comparing Apple’s failure to make the Apple car to VW’s failure to write halfway decent infotainment apps isn’t meaningful and doesn’t let us gauge the magnitude of each company’s lack of success. Apple is astronomically rich and talented but couldn’t make something very hard that’s never been done before. VW is extraordinarily well capitalized but can’t write an embedded OS with a few user land applications that don’t freeze and require reboots. Meanwhile NASA can remotely repair a probe billions of miles away.
VW just sucks at this.
I will certainly not disagree that VW sucks at it’s software, nor that autonomous vehicles are incredibly difficult, far more so than merely developing a infotainment UI that doesn’t suck, but frankly I think Apple would have failed even if they just went after a “normal” car with some fruit logos on it. Hardware is hard, especially hardware with requirements more stringent than light duty consumer gizmos, and software companies are really not set up to build it in the same way actual car manufacturers are. I think a much better approach for Apple would have been to partner with an established manufacturer who knows how automotive development actually works (Magna Steyr is a name that comes up often and sometimes Canoo depending what substances have been smoked) and focus on the UI and branding part of it.
Fully agreed. Even with their operations and supply chain chops, even Apple ends up with manufacturing and mechanical design problems, with small devices with zero safety considerations they’ve been building for ages. Building a vehicle from scratch is a nutty endeavor.
Good Lord, the fumes from that lithium fire are extremely toxic, and the fire itself is really hard to extinguish. If I worked in a battery factory and saw so much as a spark or a puff of smoke, there would be a me shaped hole in the nearest exterior wall and a dust cloud going off into the distance.
No kidding! The air in the rest of the city couldn’t have been pleasant either. Yikes.
This reminds me of when my wife and I visited South Korea and our phones randomly rang out with the scary government alert sound. Checked the screen and there was this “warning” red box with bold font but it was all in Korean and neither of us could read it. Was the North testing missiles/attacking? No one was reacting outside so that gave us calm. I put Google visual translator over my wife’s screen and to our full relief it was for an air quality alert as it was around 300 (whatever units that measure it). Pulled up an air quality index and basically it was blowing in from China where a bunch of cities were at 999 reading for air quality.
“OH YEAH!!!” -Kool-Aid Man
Yes, but they’ll need to commit resources to it in a way that most of them don’t really want to do.
It’s GM. They’ll get pretty close to having great software, then axe the project.
GM coin flip:
Release a thing, see that it did well and shows promise, but decide that they won’t do it again, and kill it immediately to be replaced by a seriously inferior thing.
or
Wrap it in an entirely unappealing package, and wonder why it failed?
Admittedly, they seem to be trying the latter, mostly. They’re pushing their version of Android Automotive and have made the choice to avoid Android Auto and Apple Carplay, as well as require a paid data service for a lot of that same functionality. So they’re stripping popular features and trying to sell data plans, neither of which are going to be appealing.
But I’ve seen enough people willing to ignore that because of several years of connectivity included, so they may see some success and kill it.
It could be the rare situation where GM gets to do both their favorite things!
Look, I’m not going to sit here and praise the bureaucracy of many legacy automakers. But “move fast and break things” isn’t a great philosophy for automotive either. Hopefully VW and Rivian can find a sweet spot. Mostly b/c I’d love an EV Scout.
The “Move fast and break things” philosophy also wasn’t particularly good for Richard Hammond…
It’s so interesting that the Rivian investment was primarily for software. But if anything, companies like Vinfast and Fisker have shown that the software is just as hard, if not harder than the vehicle part. Completely from the outside looking in, it seems like GM does a pretty good job with the software part. They’ve also been at it longer than most. The Volt had a ton of code in it, and it worked great. I never had any weirdness with my Volt. They have had some issues with the software on the newer EV’s (Blazer and Lyriq), but they seem to be getting through those.
With VW looking to invest so much in Rivian, it always makes me wonder why there isn’t a buyer out there for someone like Fisker. You’d think there would be a buyer at a fire sale for a vehicle that is already engineered and being built, especially if it fit a need in a line-up. Unless it’s so bad that some minor investment in fixing the software bugs wouldn’t make it a good car. Maybe Stellantis has something competitive in the pipeline, but they jump out to me as a company who could pick up those pieces and slot that in since they don’t have any full EV’s in that market segment.
Fiskers are being built by a third party that OEMs already work with. There’s precious little manufacturing or operational value in Fisker, and their designs haven’t taken the world by storm. Fisker isn’t really worth anything.
Just glad to know that I wasn’t the only one dousing my bags of dried wasabi peas with lysol during the pandemic. I would have never publicly admitted it had it not been for Hardigree already admitting to the same.
Haha, I almost jumped straight to the comments when I saw that line. We had a newborn and a frequently hospitalized toddler (asthma) when COVID started, so my husband spent a LOT of time “cleaning” the groceries in the garage. We were so innocent then!
It felt right, even if it was so so wrong.
We always get those metal cans of wasabi peas and watching my husband *ahem* stroke those puppies clean felt VERY wrong. 😮
That’s nothing I used to power wash every individual broccoli floret and black bean from the store after a 1-hour soak in bleach-and-everclear.
What I see as the “large automaker” problem is that they will suffer from any or all of the following problems:
-Blandification imposed by marketing so as not to alienate anyone
-Scope creep that adds delays and bloat
-Extended timelines that put any released product 1-2 generations behind the current smartphone interface
-Being focus-grouped and groupthinked into oblivion that tries to do everything, but very badly
-Reinventing the wheel
-Being tasked to lock consumers into an ecosystem that recurring revenue can be drawn from, which no consumer likes
-Cost cutting of the core functions at the expense of the latest marketing trend (AI? Crypto? NFTs?)
A smaller company that has less management bloat and marketing hind-brain driving the engineering decisions is just going to have an easier time developing and releasing a product that does what it’s supposed to do well, in a timely manner.
This is why I disagree with Hardigree here. I think it’s smart for VW to admit defeat/deficiency and buy in on something already done that’s better from a company more capable of doing such a thing. It’s common in history for a bigger company to swallow a more nimble one to use what they’ve developed with less overhead and too many managers slowing things down and forcing the product in too many dimensions (also, see: military contracts for new products) in order to adapt to change (and absorb their powers like Highlander . . . until they squander what made the smaller company valuable and have to repeat the exercise with another smaller company later on). I’d be more worried about how Rivian might make out in the end, but they need money now and this isn’t a buy out (yet?).
Aye, but from a VW investor standpoint, they’re left holding a bag of Cariad that is now worthless.
So future VW might prove to be decent, but past/current VW is less good than they thought it was and it’s a public acknowledgement of such.
True, but I think it takes a good leader to recognize a better outside solution rather than falling victim to the Sunk Cost Fallacy and continuing to stumble along the same path (though it sounds like this is several paths from the beginning).
You just described the software industry in general, and Silicon Valley in particular. If Rivian cracked the code on developing EV software, then $5 billion might be a bargain from VW’s perspective.
It happened to a company I worked for (it was a cool place to work until then and the new corporate tools took over and then they shipped production to Malaysia, anyway). I think it’s common in a lot of industries where innovation is a key driver.
This. Unless Rivan is able to keep the bureaucracy at bay (unlikely) they will be subsumed by the borg.
VW’s investment in Rivian just leads me to believe they’re going to use their platforms to build the Scouts. Think about it, less time and money invested into a platform and a faster turn around time. Win/win for both auto makers.
Great idea. Accordingly, VW won’t do it.
They say they won’t do it. But, from Adrian’s article the other day on the design process, they very well could be doing it now.
That was my first thought as well.