"Volkswagen’s troubled software division is getting new leadership. Again." -Article

[Deleted member 4581]

Volkswagen Group's troubled software division Cariad is getting new leadership again. According to Reuters, later today, VW Group is expected to announce that it's replacing Cariad's current boss with Bentley's Peter Bosch. The division was created in 2019 to consolidate software development for VW's 10 different brands under one roof. But it's been anything but smooth sailing ever since, resulting in delays to new models and even the firing of former VW CEO Herbert Diess in 2022.

https://arstechnica.com/cars/2023/0...are-division-is-getting-new-leadership-again/

 

[TDinDC]

I'm glad they picked leadership from a brand that is known for first rate user interfaces and innovative, reliable software.

 

[dtich]

yes, posted that (the original article) in another thread earlier. at least they clearly know they have an issue. hopefully this person takes the reins firmly in hand and make some good, strong changes.. even if it means getting used to another set of controls for users, if it's a lot stronger and things under the hood work much better, i'll take the hit and relearn some stuff. the system is dying for it honestly.

 

[Moonstone]

I’m very surprised to find that the PORSCHE PCM isn’t just a reskinned version of the software found in VW/Audi/Bentley etc like the all rest of the brands (that have reliable systems). Even more surprised to find how buggy it is.

 

[Hanny]

Think this debate and reason will continue to perpetuate:

• technology companies building cars = poor quality cars but great tech = Tesla
• car companies trying to evolve technology = poor tech but better quality cars = VAG, Mercedes etc etc

This article seems to relate depending on what strategies that have/are failing and has caused the article and noted points?

 

[andb]

Why not outsource the software to a big company like Mercedes did? Porsche must focus on drive-train tech not software.

 

[Archimedes]

Just bought a base GTI S as a run around town car to complement my 4S. The PCM works perfectly and Apple Car play is flawless. Most surprisingly, with a good source like Tidal, the base stereo in the GTI is noticeably better than the Bose in my Taycan.

 

[Hanny]

Just bought a base GTI S as a run around town car to complement my 4S. The PCM works perfectly and Apple Car play is flawless. Most surprisingly, with a good source like Tidal, the base stereo in the GTI is noticeably better than the Bose in my Taycan.

This this is more car operating system rather than the stereo, if I recall correctly, VW stereo might be Kenwood - so along the lines of buying in tech.

Why not outsource the software to a big company like Mercedes did? Porsche must focus on drive-train tech not software.

..and from what I saw in the Mercedes recently, it certainly was/is a success and one of the best I’ve seen.

 

[Archimedes]

This this is more car operating system rather than the stereo, if I recall correctly, VW stereo might be Kenwood - so along the lines of buying in tech.

Not sure what you mean. The stereo sounds way better in the GTI. The music separation is much better and clearer, the range is broader, and the bass is strong yet still clean. Could be partly due to a better sound environment in the cabin, but it’s significantly better. Yesterday, I played the same song off Tidal in both cars and the VW sounded so much better.

On the PCM side, I don’t hate my Taycan PCM like some others do, but the GTI screen works quicker and CarPlay works perfectly, whereas I still have intermittent issues in the Taycan.

 

[bluedonkey]

Volkswagen Group's troubled software division Cariad is getting new leadership again. According to Reuters, later today, VW Group is expected to announce that it's replacing Cariad's current boss with Bentley's Peter Bosch. The division was created in 2019 to consolidate software development for VW's 10 different brands under one roof. But it's been anything but smooth sailing ever since, resulting in delays to new models and even the firing of former VW CEO Herbert Diess in 2022.

https://arstechnica.com/cars/2023/0...are-division-is-getting-new-leadership-again/

Changing the management of the software unit is probably a good idea. Changing the head man for another non-software person, seems like a missed opportunity. No disrespect to Mr Bosch, but his background as a mechanical engineer doesn't really give him ideal qualifications for running a connected car software group.

I've spoken to several folks now in different legacy automotive companies, and there is a common thread from the people working on connected car programs: their management doesn't understand it, and are not really interested (one told me his manager's eyes would glaze over when he started talking about the connectivity features).

If VW Group wants to be automative software leaders they should really find an experience embedded/IoT/real-time software person who loves cars to lead their software team, ideally from outside the group IMHO.

From the people I spoke to inside VW group's US software team, they have also made some fundamental platform choice mistakes. With the approach they have, I believe they will struggle to get to a solid, stable connected car experience.

 

[Vercingetorix]

GM hires ex-Apple exec to lead new software unit
https://www.cnbc.com/2023/05/09/gm-hires-ex-apple-exec-mike-abbott-to-lead-new-software-unit.html

 

[daveo4EV]

the legacy car guys have no idea what they are doing…their tech is going to suck for quite a long time.

 

[WasserGKuehlt]

Changing the management of the software unit is probably a good idea. Changing the head man for another non-software person, seems like a missed opportunity. No disrespect to Mr Bosch, but his background as a mechanical engineer doesn't really give him ideal qualifications for running a connected car software group.

[...]

From the people I spoke to inside VW group's US software team, they have also made some fundamental platform choice mistakes. With the approach they have, I believe they will struggle to get to a solid, stable connected car experience.

To be fair, the problems - in software, at least - rarely originate with, or get solved by the top/head people; some make the wrong strategic bets, but it doesn't seem to be the case at Cariad. They do have a problem of execution, and perhaps an issue with naivety - their aspiration to become a cloud provider for the automotive space was .. can't find a nice word for 'delusional'. One ought to know what they don't know, or make an effort to find out. They were pitching a programming model to developers ffs, assuming the inner layer was the easy part.

I do agree they need software people - not just at the top, but especially in the middle of the hierarchy, where the work is scoped and costed (and then done). The problem is that, with the sole exception of Tesla, it's going to be difficult to grow talent in-house or attract it from the pure-software companies in the quantities they need. And we're just not - as a species - cranking out enough sw engineers. (As an aside, I am actually hoping that a large part of the people affected by the cuts in the tech industry do find openings in the "next wave" of sw jobs.)

the legacy car guys have no idea what they are doing…their tech is going to suck for quite a long time.

I don't particularly agree - after all it's not that difficult, and they weren't that incompetent. But these are long/slow-moving projects, quality isn't as easy to define and measure as in the physical world, and wrong turns/bad bets take years to emerge. On top of this, there is a strong bias in the automotive industry (and not only - healthcare, aerospace would be the same) for safety and correctness, which obviously imposes a huge time cost. "Move fast and break shit"-type people need not apply.

 

[bluedonkey]

To be fair, the problems - in software, at least - rarely originate with, or get solved by the top/head people; some make the wrong strategic bets, but it doesn't seem to be the case at Cariad.

The problem with the person at the top not have a strong background in software is that they don't necessarily have the knowledge to ask the right questions of the people below them, or to tell when an estimate for the amount of work, or a proposed solution seems wrong.

I do agree they need software people - not just at the top, but especially in the middle of the hierarchy, where the work is scoped and costed (and then done). The problem is that, with the sole exception of Tesla, it's going to be difficult to grow talent in-house or attract it from the pure-software companies in the quantities they need. And we're just not - as a species - cranking out enough sw engineers. (As an aside, I am actually hoping that a large part of the people affected by the cuts in the tech industry do find openings in the "next wave" of sw jobs.)

Agreed. The other issue (and this came out in my conversation with the VW team) was that they were not willing to pay top tech salaries. With Apple and co offering 1.5x to 2x the salary + stock and other benefits (they still are even today), they were unlikely to win over top talent.

Honestly, with the right investment, I think they could attract the talent needed and build something every bit as good as Tesla's software (and probably better). It will take a willingness to pay at the top talent level, and to hire people who are not primarily car people, including at the top. There are plenty of software people who have a love for cars though, and given the right incentives could be persuaded I'm sure. I have plenty of ideas, but it starts with the right platform choices. Getting that wrong (and IMHO, based on my experience building cloud connected IoT systems, they have got that very wrong) will make it very hard to build a successful system.

 

[daveo4EV]

The problem with the person at the top not have a strong background in software is that they don't necessarily have the knowledge to ask the right questions of the people below them, or to tell when an estimate for the amount of work, or a proposed solution seems wrong.



Agreed. The other issue (and this came out in my conversation with the VW team) was that they were not willing to pay top tech salaries. With Apple and co offering 1.5x to 2x the salary + stock and other benefits (they still are even today), they were unlikely to win over top talent.

Honestly, with the right investment, I think they could attract the talent needed and build something every bit as good as Tesla's software (and probably better). It will take a willingness to pay at the top talent level, and to hire people who are not primarily car people, including at the top. There are plenty of software people who have a love for cars though, and given the right incentives could be persuaded I'm sure. I have plenty of ideas, but it starts with the right platform choices. Getting that wrong (and IMHO, based on my experience building cloud connected IoT systems, they have got that very wrong) will make it very hard to build a successful system.

I'd work for an automotive company if they were competitive - I think they desperately need more people that know what they are doing - their unwillingness to pay appropriately is simple more indication that they have no idea what type of war they are waging - and labor cost on the OS for their ENTIRE business proposition is not where they are going to save money - it's like not paying for top mechanical engineers in the engines design department…software _IS_ that important.