Why Today’s Automobile Industry Looks a Lot Like IBM in 1985 | Hacker News Reader

Why Today’s Automobile Industry Looks a Lot Like IBM in 1985

techcrunch.com

28 pointscarlchenet31 comments

The author's point is that by using Apple CarPlay or Android Auto, they hand over control over software and become "commoditized hardware providers", just like IBM and the IBM-clone providers in the 1980s and 1990s.

I consider this argument ridiculous, as the in-car-software controls nowhere near as much of the user experience as the operating system and applications of a personal computer. Today's car builders have their profit margins because of engine design, build quality, exterior design and branding. They don't lose power by outsourcing their software stack. Actually a lot of them already use the same software over their entire model range (f.e. Volkswagen, Ford, BMW all do) and it does not seem to hurt the sales & profits of their top-end models, so the point that software is the crucial buying factor is basically already refuted.

(I think they should rather be afraid of Tesla overtaking them in what actually matters)

I suspect this is going to be like the iPod launch itself: software doesn't matter, until someone produces some that the non-enthusiast users find is a great improvement in usability.

In-car entertainment is going to be a marginal thing in the smartphone era; people just want to use the car as a peripheral display and speakers for the phone they always already have. Maybe integrate some buttons so you can control the navigation more easily.

The two threats are Tesla (battery changes the owner experience as charge requires more active management by owner) and self-driving cars in general.

I'd argue that the latter is a reflection on the escalating costs. Self driving is really a sharing scheme. I work in a place in a average cost city where the average salary is north of $100k -- I'd guess the average car is 7 years old. Hell, I'm the boss and I'm driving a 13 year old Honda.

In the 90s, the average car in my high school student lot was newer!

That might be part of it, but car quality has increased dramatically over time. In many ways the average 10 year old car now is in about the same mechanical condition as a 5 year old car in 1990. Combined with the general low mileage of most city cars and there is little reason to drive a new car.

> They don't lose power by outsourcing their software stack

It's not even "the software stack", it's only the bits that relate to entertainment and navigation. The funny thing is, it was only in the late 90's (I think?) that car radios transitioned away from being commodities, by way of the DIN mount: https://en.wikipedia.org/wiki/ISO_7736 - Cars would often ship with comically featureless stereos, because people would replace them anyway. I had a semi-fancy Sony CD-radio that followed me through a couple of beaters and I don't think that was all that unusual.

In many ways, this new movement is just a return to the old state.

This author has a lot of misconceptions what a car is and how it is composed of.

The HMI/MMI or however you would like to call that thing in the middle is only a small part of the software in a car. And even using standards like Apple CarPlay or Android Auto does not mean, that they give anything away. Both "platforms" are not even platforms, they are just proprietary versions of VNC. Its like the current iteration of the Apple Watch: there is no application running on the watch. The applications can display information and they have a back channel for the user interaction. But the real application still runs on the phone. That is also true for Android Auto and Apple CarPlay.

Then there is much more to control in this second display (the first display is the dashboard), which is air condition, power sliding windows, roof, and all the other car functions. Because anything in a car is controlled by software. In a normal mid-sized car there are about 40 to 60 ECUs (electronic control units), in a luxury car that can go up to 90 and some more.

Current estimates are that a modern car is comprised by about 30% software, 30% electronics, and 30% mechanics.

Seeing the reason decline of software quality by Apple, I am feeling pretty comfortable that they are only doing user interface stuff, that is unimportant to the cars performance.

Have you checked with the customers? I have never heard of a customer commenting software features of a car. It looks to me that car buyers are interested in other things in a car. I mean, at least for me (potential customer?), an Audi with the shittiest software will always win vs a Fiat with the best software. You can't just compare completely different industries and draw conclusions like that, you should give some more factual stuff to prove your argument I believe.

> Have you checked with the customers? I have never heard of a customer commenting software features of a car. It looks to me that car buyers are interested in other things in a car.

Even as a fellow auto enthusiast, I get the sense that we don't represent the majority. There's a reason most car salespeople can't explain anything technical about the vehicle they're trying to sell you -- most buyers don't care about those details.

If Tesla (or anyone else) comes out with a cheap enough car that can't do over 50mph, runs on a user-friendly electric ecosystem, has terrible driving dynamics, and can autopilot between home and work?

I feel like they'd seize 25%+ of US auto sales at least.

Nah, at least not people where I live. While most people aren't enthusiasts, almost everyone does 60mph on the main arterials, and 70-80 on the actual highways.

I have to agree. I'm not a car enthusiast, I do spend a lot of time in my car, and I can't imagine making a purchasing decision based in any significant way on the software features. I've got everything I need on my phone.

And I just threw out 50mph as a "cheaper" number.

You really think people would trade the ability to completely focus on something else during their commute (autopilot) for the ability to go faster but have to focus on the road?

I think those of us who would do that are already using public transit (I include myself in this group, by the way). It might get some people who say "well, I would take the bus if it ran where I live/work", but anyone who lives and works on a bus route would already be using public transit if most people felt this way, and public transit will continue to be cheaper for quite some time.

The cost of public transit for me is mostly time: imo bus connections are terrible when you're doing it every day (or twice!). And I ride as often as possible.

It also supposes your city has an efficient and reasonably cost effective public transit network. Which isn't always the case.

Automated electric personal vehicles solve both these issues: it's as timely as you can get (unless dedicated rail is weighed in, but...), and it requires only a city electrical grid.

Cars will become commodified, and as long as it’s running CarPlay or Android Auto, I won’t care what I’m driving.

Somehow I don't think so. It doesn't help that the author gets the mechanics of the PC transistion wrong as well. The biggest threat at the moment to automobiles is self driving. If we get a self-driving Uber or Lyft we're going to have an entire generation of people who just have their phone summon transportation and never own a car.

> If we get a self-driving Uber or Lyft we're going to have an entire generation of people who just have their phone summon transportation and never own a car.

This is going to be affected by cost. Let's look at the numbers.

A Lyft between work and home for me is about $10, plus tip (which is usually $2 for me, except in the rare case the driver's a creep), assuming Prime Time isn't in effect.

If I were to take Lyft both ways every day, that's about $24/day. If I work 22 days in a month, that's $528/month just on my commute. If I want to run errands and Lyft it to places outside of work, that's even more money. Oh, and until the beginning of this year, I worked much farther away from home than I do now, and a Lyft would've run me about $25 each way (possibly closer to $30), plus tip (which, by this point, would go up to $5). At $60/day, I'd be spending $1320/month on commute alone.

How does this compare to other methods of transportation?

Looking at public transport, a monthly pass here is $80. This is a great deal if you both live and work on a bus or train line, assuming you're willing to put up a longer commute. For example, public transit was perfect for my last job, because I lived and worked on the same bus line, and while my commute took an hour (going from the suburbs to downtown by bus takes a while), I didn't mind it because I could just spend the bus ride reading books and comics on my phone, or if I'd been sleep-deprived, I could just zone out in my seat. Also, the only benefit my last company offered was that they'd pay for employees' transit passes (or for those who drive, they'd pay for a parking pass at a local garage, but we could only pick one or the other), so I basically got to commute for free. If you're off the beaten path, however, public transport isn't an option, and if you're reliant on public transit, it limits the companies you can accept positions at.

How does this compare to owning a car? I don't know. I'm not joking here: I've never actually owned a car. Anyone else care to chime in? For the record, I live in North Texas, where cost of living is relatively low, so I'd especially like to hear from other people in my area. If owning a car is cheap enough, then Lyft and Uber will have to drop their prices like a rock to supplant cars for everyone.

You live in suburbia, where the distances are long because it's been designed for cars, parking, roads and huge highways. The parking & roads are also subsided. Also your transit is probably subsidized heavily. %80 of the cost is probably paid by your local government / tax dollars.

Lets take the cost of owning a Prius V that drives 500 miles a day:

- $260/mo car 'lease' ($8.6/day) [200'000 miles over 8 years]

- $200/mo in repairs ($6.6/day) [Will probably be less in reality]

- $200/mo in insurance ($6.6/day) [Will probably be less in reality]

- $50/day in gas

Now the shift employees, that you pay lets say $15/hr on average for 18 hours a day are $270/day. Their cost alone is 3.75 times greater than all of the other costs combined of $72/day. Even if the robot service costed a crazy $2000/month (which it wont be), that is still 4 times lower than employees. So if the corporation passes the robot savings to you your commute costs, they will be x2.5 lower even with my inflated estimates. Your $528 will become around $220, and that is less than the cost of owning a car currently.

If we can convince governments to shift their subsidies to these robot cars to make it %20 of costs, it will be $44/month, which is about half of your transit ticket cost, for on demand travel.

In all cases, the majority of the costs is human labor.

By making cars shared, less cars will need to exist, so parking real estate can be reassigned to other more productive uses. There will be less congestion because these automated cars can synchronize properly. The total amount of cars that would need to exist would only be needed to serve peak base load, like a electrical utility.

Also the speed of trains comes from the fact that they have a dedicated 'road' that only they can travel, so congestition doesn't exist and therefore the service is predictable. You can do the same thing with busses and roads.

Exactly this, and the fact that self driving cars can replace all of the other forms of transport (train, bus, taxi) and with electric versions of said cars amortize maximize energy utilization and reduce waste, the numbers just get surprisingly good all around. 80% of the population of the US lives in an "urban" area and they won't personally own cars any more.

[1] http://www.citylab.com/housing/2012/03/us-urban-population-w...

Software is only a small part of the car. The other parts to consider is the build quality of a car. For example, compare a Skoda and a VW. Even though, they are from the sample base platform, their build quality is very different. They share plenty of the same software. Build quality includes items like suspension, safety features and about 1000s other items.

That's not actually quite correct. Build quality usually refers to the overall integrity of the finished product, it's ability to withstand wear and tear and reliability.

The difference between a skoda badged VW and a VW badged VW is more about the finish quality - the differences in the seat materials, amount of sound deadening, the amount of standard features and so on.

In most cases the active and passive safety features of a skoda/VW will be identical for the same platform, especially when it comes to things like power train components, which are purchased and built in bulk. The main difference between a VW and skoda engine is that the skoda misses out on the plastic engine cover.

This article is a stretch at best. There are too many subsystems in a car that are already controlled by software from Bosch and a few select other companies that have no or little interface into the UI of a car. The author here either doesn't get that cars are a level of magnitude more complicated than computers or he doesn't care. Electric cars can simply things somewhat as it takes a lot of sensors and cpu power to make a gasoline or Diesel engine run and meet emissions standards. However there is still a lot to manage in electric cars too that goes way beyond Bluetooth control and maps.

I think I'd trust Google or Apple more with my self-driving car's security than say BMW.

http://www.pcworld.com/article/2878437/bmw-cars-found-vulner...

Alex Stamos on Twitter recently:

> If software is eating the world and all companies are software companies, then they all need to learn hard infosec lessons faster.

https://twitter.com/alexstamos/status/600075692731191296

Couldn't agree more.

This talk of connected cars is kind of disturbing. How long until cars serve targeted ads on their HUD?

Does your Android or iPhone show ads related to the OS (and not in Apps)?

I agree we should be careful, but it's not like you'll be flooded with ads.

No, my Symbian Belle doesn't show any ads. And it doesn't call software "apps" :D

By modern smartphone dictionary, whatever the car shows in its HUD is also an "app", so I don't see a reason why it wouldn't show ads there, especially once they add (I think they did already) support for installing downloadable applications.

I just don't see the benefits outweighing the risks when it comes to massive corporations like car makers.

If the car can drive itself, what will the person inside be there for? If not to view ads and consume media?

Google's vision of the future, right there.

To take the blame when the software... err.. car... crashes, of course. The corporation sure isn't going to take the blame.

CarPlay is not great.

I don't have any experience with Android Auto or MirrorLink, but I installed a Pioneer head unit (AVX-4000NEX) a few months ago in my car so I could use CarPlay.

The head unit itself is great, gets good reception, sounds great (especially after using the microphone calibration), and the Pioneer interface is fine. I don't know that it's worth the price without considering the deep phone integration. I considered reinstalling the factory head unit (a Scion Bespoke w/ navigation), but I had sold it before I could come to a decision.

CarPlay would work great if Siri worked well, but I've never been able to get Siri to work well in the car or elsewhere (and that assumes Siri is taking requests at the moment). Text-to-speech works fine; dictation is a constant frustration. The other week, I tried to send a message to my wife that "I am in a little traffic so I'm running late, but should be there at 5:20". I don't recall the erroneous outputs generated by Siri, but they weren't even close. After multiple tries, I had to give up. The Maps colors are idiotically low contrast and it can be very difficult to see what current traffic conditions are as a result. The zoom levels in Maps never seem to be right, either too much or too little. The touch interface itself requires too much attention to use and is a car accident waiting to happen.

CarPlay has a long way to go before it is competitive with iDrive, MyFord Touch, or Uvo, which are the systems that I've experienced. My experience with CarPlay makes me question how much testing it actually got while installed in cars with people driving, because there's a lot of glaring usability problems and general interface quirks that should have been caught during testing.

I used to think that CarPlay, Android Auto, and MirrorLink were going to destroy the car maker's infotainment systems. After using CarPlay, I'm convinced it's going to be a while before that happens, if ever. The car manufacturers have a lot of experience with "car UX" and aren't standing still with their own products. Infotainment is also a minor part of the entire car ownership and use experience, and I don't know that I can think of anyone that bought or leased a car because "it has iDrive" or whatever. Drawing an analogy between car manufacturers and 1985 IBM feels forced and is flawed.

> I am in a little traffic so I'm running late, but should be there at 5:20

I would phrase this as "Running late due to traffic, ETA 5:20" and this usually works. Siri seems to trip up on articles especially if you don't speak precisely. (And yes, this is extremely annoying)

This seems a silly thesis, if only because cars are already commoditized due to years of competition. Early PCs has huge profit margins; current automobiles generally do not.

Commoditization is not the same as low profit margins.

The auto makers historically spend a ton of marketing money to keep the auto market from being a commoditized market. So far they've been very successful.

Profit margins are pretty healthy in some segments, for that matter.

Reply on news.ycombinator.com