157 replies
5d
CarPlay mostly works, and although it has issues it's definitely user-focused.
CarPlay doesn't prevent car makers from tracking vehicle activity.
In the end, consumers don't give a shit about the in-vehicle infotainment. It sucks, or it's AA/CarPlay. The first generation of iDrive showed that rich people people will buy cars in spite of the in-car stuff. In fact, most car infotainment sucks, yet people still buy cars.
Let's turn the question around: why would car makers want to spend millions of dollars a year rolling their own infotainment system? So they can make incremental revenue selling ads and user data? So they have control? Control over what, exactly?
I wonder how much of those frustrations has to do with Japanese market share of car industry; I think the touchscreen infotainment is not there because car manufacturers value it as integral and central part of car experience but simply because otherwise their product loses against one of Japanese brands.
Japanese road network is a disorganized weighted node graph and absolutely not a grid, and a bulletproof navigation unit has been a must for a car in Japan since its inception around 1990. It is also preferred that they are 2DIN compatible so it can be later upgraded. AFAIK, those are not high priority checkbox items elsewhere, but all cars nevertheless follow the Japanese manufacturer layout because of manufacturers' collective dominance. Cars before 2DIN navigation units seem to have had 1DIN AM/FM radio units with radio buttons[1], by the way.
That dominance leaves a 4:3 8" diagonal hole in immediate view of driver for all cars globally that must be filled with something of value. That doesn't have to be a touchscreen but usually are, and it ends up being a navigation-audio combo unit, and it's outsourced to the lowest bidder. It is not the primary interaction point for cars by overwhelming global demands or principles of automotive product design. That leads to jarring subpar experience that appear to be but are perhaps not intended to be part of core UX of the whole car. I think.
1: https://www.alamy.com/1956-mercedes-benz-190-sl-steering-whe...
Don't Japanese people have smartphones or tablets with navigation on them that they can use? I'd rather cars just have a place I can mount my own device, rather than include any kind of screen whatsoever with crappy un-updated, un-maintained software.
Cars can have better GPS reception, which is important when you're trying to drive in an urban area with a lot of buildings. Phones are bad at accuracy there.
(CarPlay can provide car GPS to the phone though.)
My iPhone 12 has much better reception then my 22 Tacoma even when tossed on the passenger seat.
Like waaaaaay better and more accurate
It just pretends to be more accurate by using AI. A car tends to be more raw, which is what you want when navigating at speed.
As someone who used to help out with Waze maps, raw GPS tracks from phones can be WAAAY off from actual roads and other GPS tracks.
Your phone looks for the most likely road for you to be on and snaps you to that road. Your car does a similar thing but has much higher GPS resolution than your phone so it is more likely to be right.
Could we please for once stop using the term "AI" when we mean "heuristics"?
Phones do use AI though, at least iPhones. They look at all sensors (accelerometers, gps, wifi, etc).
Sensor fusion is AI I guess.
AI can be distilled down to a series of if-statements. That's basically all it is.
This is true of literally any software and if you really wanna oversimplify it, our brains are just a bunch of if statements (condition based logic) too...
I noticed too late to add /s or to attribute the quote, but kinda sad that people are taking it literally.
It’s a play on the quote “Life is a series of choices and reactions” from “Before The Fall”
Right, expert systems are GOF AI. What isn't?
Gosh, so the moment you use an if-statement, or an aggregate function (like sum or average), or weighting coefficients to mark some inputs as more significant than others, it's "AI" already? And I naïvely thought it was just, well, "computing". You could do "AI" on a 8086, ffs.
Yeah, forgot the /s ... wasn't being serious.
heuristics and Meta-heuristics are AI. Heuristics are how children (and adults) learn to solve new problems.
no it IS more accurate. When i am off-roading i always record my track and have to unplug my phone from the trucks system to get an accurate un-smoothed track.
google maps does weird things but gaia just uses the raw GPS data from the phone and doesn't snap, however when you are plugged into carplay the phone seems to always consider the car/trucks data more accurate even when its clearly not.
this could very well my specific truck (dealer has refused to look and i don't have another to compare to) or a general tacoma thing but it both smooths the GPS data to make "nicer" looking lines and often introduces drift over time as likely some processing it does compounds an error
(i also realize i said the wrong phone, iphone 14 not 12 but even with the 12 it was notably better then the truck but the 14 pro has a much improved GPS and it shows)
AFAIK, and I could be wrong (though I did google this first), plugging your phone into CarPlay does not use the car’s gps unit. That’s just your phone trying to find a road because it’s connected to a car. Even if you are getting the car’s gps, it’s not going to give you raw values, but display values to show in an app, which should be smooth, IMHO.
https://discussions.apple.com/thread/250076351?sortBy=best
this happens when near roads or not. the truck is doing something weird and wrong with the GPS data and forcing the phone to use it
No, it does use the car's GPS unit if it's provided to the phone. Which is why I started this thread by saying it did.
But the in-car unit has both GPS (with the antenna being on the roof) as well as wheel speed. Noticeable in longer tunnels with traffic jams that make the speed vary. Then it's clear Google maps takes a random guess at where you are while the in-car system knows how far along you are.
the only time i am paying attention is when off-roading and recording a track, and then its very obvious when the GPS is doing a good or bad job and unplugged from carplay i get a far more accurate track.
Phones also use WiFi access points and some Bluetooth beacons to get a better location. AFAIK only Google and Apple have up-to-date maps of these that are useful for navigation, and no one else does (Mozilla threw the towel a year or two ago).
Cars can theoretically have better GPS antennae, and can do better dead reckoning using wheel position and wheel rotation speed. But do they? I haven’t seen any evidence that any car actually does better than a modern phone.
Especially inside a dense city, WiFi location can be much more precise than GPS.
i only really care about accurate when off-roading in the middle of nowhere with no reception and recording my track - my iphone performs far better when unplugged from the truck then when plugged into it.
That's possible too, but the issue I was thinking of wasn't exactly reception - it's that GPS signals bounce off buildings and so it looks like you're somewhere else. Stronger reception might actually make that worse.
Japan has local satellites to help with this (QZSS) and local cars may handle that better than global phones.
Cars have wheel rotation and direction sensors so can do dead reckoning when the gps isn't being useful
Phones will use the accelerometer and gyro to do dead reckoning too, like when you're in a tunnel.
Phone navigation apps are surprisingly bad in Japan (Google ones in particular). Not unusable, but the in-car GPS is more competitive than in other markets IMHO.
Also with an aging population a phone screen is just too small for many.
BTW, to parent's point VW has been trying different approaches with a top mounted GPS/infotainment unit that can be omitted on cheaper trims.
That's not true. Google Maps, in particular, has by far the best route guidance. It used to be different - but not anymore. Most in-car systems don't even attempt to route the last 300m here because the detailed one-way street narrow layout of tiny streets isn't in their mapping data, or their algorithms are too weak. Google has mostly no problem with that.
Entering Japanese addresses IS tricky, though; here, custom-built Japanese solutions outshine. There are (mainly) no street names; instead, you specify your location by filtering down from Prefecture (Tokyo), City (Ota-ku), Commune (Kugahara), District (1-Chome), Block (26), House Number (1).
Japanese systems allow you to enter it this way - with Google (or, even worse, Apple Maps), it's a bit hacky. You would specify it as Kugahara 1-26-1 and hope for the best.
Why not use plus codes? They're well-integrated into Google Maps: https://maps.google.com/pluscodes/
TIL.
The weird part is Maps will accept them for search, but not display the plus code except when looking at POI (randomly selecting a point on the map doesn't show it for me, I only could get that from the /pluscodes/ map)
This seems genuinely useful, but as usual we're having the chicken and egg problem to get it adopted ?
In mobile at least, you can touch and hold to make a red temporary POI that will show you the plus code.
But yeah, nobody uses it in the real World. If I need to share a point with someone else I'll just send them the Google Maps share URL instead, which has its own shortcode
Google Maps fails in subtle but weird and sometimes costly ways.
On one side, it can't route through a bunch of valid paths. I first assumed it could be because of residents asking Google to cut traffic, but sometimes it's not even through residential areas. I see routes on the map that are avoided in favor of bigger loops, and when trying the shorter routes they're perfectly fine. Or perhaps it's the vehicle size and they optimize for SUVs ?
On the other side it will happily route you through paths that are restricted to specific categories of cars. It's up to the driver to carefully avoid them, but it really wants you to go through and reroutes you there when you deviate, so it's a huge PITA in areas you don't know and try to navigate while ignoring the navigation instructions. Cops seem to have noticed it, We've got fined the first time we fucked up, and now that I'm aware of the issue I see the cops in many of these spots basically waiting for the jackpot.
At this point at google, I wouldn't be surprised if they were selling your route to advertizers to "optimize your driving experience."
It should not be a surprise at all.
https://web.archive.org/web/20230129231747/https://www.waze....
https://www.waze.com/ads/
I notice that Google Maps does very poorly on my iPhone 13 Pro with the multi-level roads, like riding around in the tunnels under Tokyo or on the roads with highways above the local roads, whereas builtin navigation units usually have no problems with this. I also find their directions to be much harder to follow than the built-in navigation units when riding around the major roads where turning right requires you to exit to the left for instance. Also, Google Maps fails to provide the variety of route options with fine-grained toll-booth costs, that all navigation units I've used in the last 5 years have gotten spot-on.
Aging population prefers low quality tiny screens with awkward interfaces?
They're not tiny, at least not compared to a phone.
https://www.daihatsu.co.jp/lineup/move_canbus/03_exterior_in...
Damn that looks like a cool car. Shame it's Japan only
Japan gets way better small car options than the US.
They prefer something they are already familiar with.
That’s one reason I use CarPlay - the interface hasn’t changed for years. Apart from one really annoying change to Apple Maps when they replaced the “silence directions” toggle from a big square to a touch, wait, touch tiny icon interface.
Japan has some challenges with traditional GPS and satellites that can be low on and buildings interrupt the signal.
https://en.wikipedia.org/wiki/Quasi-Zenith_Satellite_System
...
Its got a neat orbit that does a figure 8 with the apogee over Japan (see also https://en.wikipedia.org/wiki/Tundra_orbit )
That definitely does not match my experience. I was in Japan in December of 2022 and used my phone to get everywhere, the directions were extremely accurate, more-so than in my home town in the US.
In car navigation came before mobile smart phones
Ironically, the nav unit I used on a rental car circa 2007-08 (birth of touchscreen smartphones) was exactly like the GP describes, a separate unit that had a mount like the modern smartphone variants.
This is sort of the goal of Android Auto and CarPlay, but not as a mount for your phone. Rather, it turns the screen into a dumb terminal for your phone, bypassing all the shitty built-in software and providing a UX designed specifically for use while driving.
And critically lets you directly access CarPlay interfaces to apps for streaming, podcasts, navigation, etc, all of which is much richer than a dumb audio only passthru with skip buttons.
The Japanese navigation units in cars have been really good for at least 15 years. The signs and lane markings usually match exactly what’s ahead of you in real-life, and there’s a radio system on the highway gives traffic and road closure information even if you’re not connected to the internet. It’s only recently that smartphone apps are as good.
With back up cameras being mandated in the US (and other jurisdictions?):
* https://www.cnbc.com/2018/05/02/backup-cameras-now-required-...
There needs to be a screen: I'm not sure what the incremental cost is of it being touch-sensitive. And I think a lot of designers (or cost cutters) may have figured that since they have to have software anyway, it may easier/cheaper to deal with software buttons than with moving parts like physical buttons and dials.
I'm pretty sure a touchscreen is cheaper than buttons, or at least awful close. You'll still get some buttons, unless the car maker really hates you, and hopefully something to turn for volume control (which may even work most of the time, if you're lucky).
Touchscreen is such a stupid low value move for driver. Physical knobs are safer in many ways, last much much more, I don't need to lose contact with whats happening in front of the car to manage these, just muscle memory. BMW has nice big physical knob (turn&press) to control all menus, much better than hunting on flat glossy screen with fingers.
Some stuff can be managed from steering wheel, but instead of putting there things like AC there are voice controls, while we all know very well this increases danger on the roads and calls shouldn't be made while driving.
That being said, I choose (used) cars based on many metrics, and this is just one of them and not the most important one. So its sort of self-inflicted move to worse.
None of which some up on a spreadsheet in the manufacturer's accounting department.
Similarly: open concept offices with hot desking lowers the amount of rent (per employee) that shows up on an expense report, but the productivity drop of open offices does not.
I used to agree, but the point I made in GP is that the touchscreen front and center could be a requirement, with what to do with one being open to local market interpretation.
IF that's the case, and that's one big if, debating touchscreen vs buttons might not be meaningful, because touchscreen MIGHT be a requirement that cannot be removed. With un-eliminable touchscreen, buttons becomes just added cost, rather than being a resilt of cost calculations between all-touch vs all-buttons.
Button-y interface can be handled on basically any IO-heavier 8-bit microcontroller with some muxes as all it needs is to handle interrupts and emit event messages over whatever-bus. The solution can even be hard real-time if there is such requirement.
Touchscreen will probably run atop whatever OS powers infotainment system. In theory touch events CAN be handled on the same 8-bit uC, but converting click coordinates to button-press messages requires knowing what is visible on the screen, which either hinders expressiveness of touch interface (fixed islands for controls) or complicates implementation (constant updating of element-coordinate maps). The moment safety-critical control runs along with non-safety software huge validation challenges arise - the performance characteristics of the whole system, including UI, must become at the very least bounded.
Hard button solution can be internally commoditized, whereas soft button solution requires constant babysitting throughout development lifecycle. One is much more expensive than the other, however, once the move is made incremental cost drops dramatically.
Sure, there needs to be a screen, but it doesn't have to do anything more than be the monitor for the backup camera.
Honestly I would prefer to have a 2DIN hole where I can put whatever infotainment I want, instead of being stuck with whatever the car maker included, and that I can change without changing my car.
Yeah, it can't get pro-consumer and pro-right-to-repair than that. Disappointment with lackluster stock infotainment followed by disappointment in the move away from repairability felt a bit hypocritical to me knowing that world. 2DIN holes are fine!
I'm not quite sure what you mean by saying that the Japanese road network is not a grid. In terms of actual layout, it's quite grid-like. I would even say it's usually much closer to a nicely laid-out grid than the mess any European city is.
What Japan lacks are addresses that can be found easily without using a map. Apart from Kyoto, roads in Japanese cities don't have names (or number), so addresses within cities are not "{number} {name of street}". Cities are cut in areas smaller and smaller all the way down to a block. The last number will be the house on that block. So addresses within cities are "{name of area} {sub-area number} {block number} {house number}", with some variations from city to city.
An address might be "Nantokacho 11-16-8", which means the 8th house around the 16th block of the 11th sub-area of the Nantoka area. Good luck figuring that out without a map!
That's really cool! I didn't realize that. Thanks for explaining.
Reminds me the Township & Range system in the US, which is still used in public land management: https://web.gccaz.edu/~lynrw95071/Township%20Range%20Explana...
I don't own my own car here, but whenever I ride taxis or rent cars in Japan I noticed that they almost all appear to have aftermarket infotainment units. This makes sense now!
GM's plan is to sell a subscription service that covers all the things you already pay for on your phone (maps, music streaming, etc). It's why they're killing CarPlay, because they know that even if their service is good, nobody will pay $20/mo for shit they already get for free on their phones.
Basically, their goal in life is to be a worthless middleman who takes peoples money while providing no real value to society.
How is GM a middleman when they are producing the product I am buying which is a car? If GM bundled OnStar with a top tier infotainment / navigation / ChatGpt / I’m hungry where do I eat app, I’ll pay for that as long as it’s reasonable and lets me know I’ll be taken care of if I crash and need assistance.
Maybe for you. If it doesn't exactly mirror the stuff on my phone, though, it's distinctly worse than using my phone. If it doesn't start playing music from my phone when I turn the car on, it's distinctly worse than using my phone. If it doesn't act like my phone does today, it's distinctly worse than using my phone.
For me, there is no room for GM to become a player here. I have friends' addresses, etc. in my iPhone's contacts and they show up on Maps. GM isn't going to do that, so GM isn't going to get my business. I'm starting to look for a new car right now, and this is a hard no. If it doesn't support CarPlay and Android Auto (both--I've used both in the last five years, and might switch back again to Android in the lifetime of my next car), it's outside my observable universe.
I would never consider buying GM. I was, in the past, a UAW shop steward.
I once bought a GM vehicle. Never again.
My family had GM cars when I was a child, decades ago. Never would I seriously consider buying a GM.
That's fair! I've looked long and hard at the Chevy Bolt in the past, but I think other car models have surpassed it.
Why would you pay for that when your phone already does all of that and you always have it on you?
They are trying to be a middleman between you and the internet.
It’s analogous to a TV and monitor manufacturer saying they are no longer providing non-wifi inputs, app side loading, or supporting screen casting from local devices.
Then the manufacturer would charge a monthly fee for the monitor to keep working.
Hopefully, next, the CEO acts surprised when reporting the resulting massive drop in shipments.
GM doing top tier tech is a joke. Also, they monetize your car data.
Yea but that's the joke is they plan.
Toyota should also plan to expand their manufacturing capacity to pickup GMs lost sales to anyone under the age of 60.
Lol
Toyota offering a V8 (both gas and diesel) in its trucks would change the market in a big way.
I’m surprised Toyota doesn’t offer a hybrid pickup similar to Ford’s.
I wish any manufacturer offered a 250-class electric truck with a basic trim line. They probably want to sell more cars per battery for now.
Tundra Hybrid doesn't count?
No, that’s a half ton class truck (f-150 comparable). A 250 class truck (f-250, 2500, etc) is a 3/4 ton class and is what you need if you want to tow >10k pounds while having basically any load in the truck.
Tacoma and tundra both come in hybrid now
Doesn't Tesla do that as well?
No, Tesla sells internet connectivity for $10/month.
While Tesla does have a $10/m plan, you still get basic Internet without paying for it... Tesla's free is better than any other company's paid subscription.
That said, $10 is far too much for what it is, should be $5.
A Tesla costs $80k and lasts 10 years. Why are you saying $1200 for internet is too much and $600 isn’t?
Why do they even bother charging - just bundle it in with the headline price. It’s a tiny portion of the revenue, nobody is going to think “oh that $87,345 Tesla is too much I’ll go for the $86,145 one”
Tesla costs $30k https://www.tesla.com/modely/design#overview $29450 with tax incentive
You pay for maps on your phone?
You pay for cell service on your phone, which lets you access things like Google/Apple maps, spotify streaming, etc.
GM wants to charge for a separate internet service tied to your car, that still goes over the cellular networks, in order to access these exact same features in their head unit. It is completely redundant. CarPlay already offers a near perfect experience in this regard, without charging an extra monthly fee.
Either you pay with your data (Google) or upfront for the phone (Apple).
I use Organic Maps on GrapheneOS and add businesses I want which aren't there (hence I had to look them up in Google Maps in Firefox Focus, which is painful enough that I don't want to have to repeat it).
Planned obsolesence. Without proper CarPlay/AA integration, car manufacturers get to decide when those whiz-bang infotainment features stop working. You'd have to replace the whole car to get those features back instead of just buying a new phone.
Heck simple bluetooth audio playback has degraded year over year in my car. After an android update a few years back I don't get to see the track name any more, pause/play sort of works, and thankfully audio still comes through and I can go to the previous/next track.
Without constant updates, software that is part a a larger ecosystem will eventually breakdown.
This is why we should be so skeptical of tight software integration with durable hardware (e.g., cars and appliances with operational lifespans 10+ years easily). Software has a pretty short half-life, especially software that integrates with internet services; vulnerabilities get discovered in third-party components and remote APIs shift out from under you.
Durable goods manufacturers have little skill or interest in long-term software upkeep (maybe they like the profits and the rent-seeking, but not the actual maintenance), so the most sustainable design is one where the software is easily seperable and replaceable from the core durable item. Manufacturer-specific internet-connected infotainment in cars ain't it.
No, the most realistically sustainable design is one where the software is built into the durable item and not connected to the internet, so it never changes and doesn't need to worry about security issues. You can keep using the thing until it fails from mechanical issues and can't be easily repaired.
If it's internet-connected for some reason, then the software should be open-source so interested parties can take over maintenance after the OEM decides they don't want to bother any more. However, I don't consider this "realistic": how many manufacturers are going to do this in reality? Almost none: they want you to buy a replacement device.
How would that work in a car, where people expect map updates and streaming music or radio over the internet?
I don't think it would. I was really thinking more about things like home appliances. An embedded computer is handy for a washing machine, for instance, but it doesn't need to be connected to the internet (no, notifying you it's finished isn't necessary: the machine should tell you when you press "start" approximately how long it will take).
For a car, I really don't know. I don't see any way of putting up-to-date navigation software into a dashboard without a significant security risk, which means needing the ability to keep software updated. Even CarPlay/AA are non-trivial and would need security updates.
Just like that. Robust, simple software, open standards a r/w partition for map data. Having well-defined, narrow access to the outside world is much easier to secure and keep working.
My car isn't connected to the internet, it has not received any updates. However Google enjoys messing with the Android BT stack for whatever reason (there was a period in the mid 2010s where BT on Android was dramatically ~broken~ rewritten every couple of years), so accordingly my car has lost functionality.
Do people do that? I just use my phone beside the old useless infotainment. Honestly, I wouldn't buy a new car that used an in-house infotainment specifically because they go obsolete quickly (<5-10 years) and cannot be economically (or at all) to the latest tech. Carplay seems to be long lived.
My vehicle (a mid teens Volvo) doesn’t support CarPlay or AA. Its infotainment system gets its data from 3G, so it doesn’t get data anymore. Music is through Bluetooth, which works well enough. Navigating means using the clunky in-car nav with no traffic data or a phone with the volume turned way up.
It’s far from ideal but at least I have zero expectations for my next car.
Why don't you use bluetooth to navigate with your phone?
I replaced the head unit on my car from the same era with an aftermarket Sony XAV-9000ES unit that has wireless carplay support. It takes about 10 seconds to connect after powering up the electronics but is otherwise seamless. Full steering wheel controls too (volume, next/prev track, pause/resume, mute). No car info, but the car is old enough to have a proper gauge cluster.
There are also dash-mounted "carplay screens" that are both much cheaper and much easier to install but lack integrations/require hacky solutions to get, say, audio out.
No degradation at all for me over seven years. It’s a game changer. Car still feels new
You don’t like paying $300 for a map update? On my car, Nav + traffic was $3k. CarPlay was $300 and I get multiple map choices like Waze or Google maps. Only infotainment that comes close is Tesla yet map data is always out of date with construction road closures.
come on. It is AOL of cars.
This isn't "we're going to deprecate your car, buy a new one". People will buy them anyway.
It is "You're going to pay for AOL, even though we have the internet"
And when you sell your car, some other dumb schmuck will buy it used and sign up for AOL-of-cars.
For me, the fact that it updates with your phone every year is a game-changer! The end of those bad user interfaces.
My ideal "infotainment" is a button that lets me pair bluetooth and volume/skip controls on my steering wheel. I have a phone mount, like almost everyone does. I don't need a display on my dashboard.
AA/CarPlay (and everything else) are genuinely just distracting annoyances that take up space. Especially in newer cars where the screens are for some reason, no longer matte and often angled upwards so that they just blast reflected sun into your eyes.
If find the Bluetooth interface too limited because it requires interacting with the touchscreen on the phone. I want to use the large screen on the car with an interface that is optimized to work while in the car not small controls on a small phone.
No touch screen i should be accessable by the driner unless the car is in park. I don't care if it is the phone or infotainment, no touchscreen is allowed.
The person I was replying to was arguing for a Bluetooth connection to their phone instead of CarPlay. CarPlay is much better from a safety standpoint than a phone screen.
You just choose a superlong playlist before starting the car and then play it randomly. You don't have to interact with the phone.
Last time I had a car I had an old mp3 portable mp3 player constantly hooked up to the car with 4GB of music inside. I would just hit play and drive and would only adjust volume from the steering wheel if I needed to. If I had a car again I would probably do the same, that is the lowest distracting way to have music in your car.
It what if I’m in FSD mode? /s
There is absolutely nothing on the phone I need or want to interact with while driving that is not covered by bluetooth audio controls.
Perhaps you should reconsider your driving needs vs safety?
I'll think about that while I'm sitting at a stop light and skip a chapter in a podcast.
Likewise. I vastly prefer the physical buttons, as I can hit them without taking my eyes off the road. The only two things I use infotainment for is mapping and music control.
Bluetooth should let you select phone audio sources from your car buttons.
It’s usually terribly implemented. iOS doesn’t support it for third party apps or first party podcasts, and apple music’s support is a shit show. Android is better, but it is hit-or-miss depending on the car.
I wish the government would declare this a safety feature and mandate recalls for incorrect implementations (including phones and cars).
This is addressed in the article:
But I think a great differentiator for upmarket brands would be to offer as few screens as possible and as many buttons as possible. Electronics are a plague. Some features are very useful: mainly, route planning and the ability to play music in the car. But the rest is a nightmare.
I don't want to talk to my car; certainly not when passengers are sleeping during a trip, and even not when I'm alone. There's a subtle humiliation associated with talking to a machine. More importantly, I don't want to randomly target zones on screens to set up things, and I don't want to look at the screens because I don't want to take my eyes off the road. Give me buttons that have a fixed location, and that I can feel without looking.
I’m happy with the way Mazda has integrated CarPlay. No touch while moving, the navigation wheel moves a highlight around the screen to select buttons/links. I find it much less distracting.
Audi is similar. I think also Mercedes. It’s great.
The audi wheel with buttons around it is terrible. I have to look down to find the correct button, and then I need to look at the screen to wheel to the correct option.
Touchscreen is not better than physical buttons but at least I can still look forward while operating it.
This audi is the first time I actually started using Siri - and not because I like it.
On the new mazda3 it’s no touch at all. They took it out completely
Don’t you think it’s possible that car makers are moving in this direction in anticipation of a day when you no longer have to be paying attention to that extent while driving? Or do people still think that day is very very far off because based on my experience with FSD we’re not as far as everyone thinks.
That's the danger of FSD. It's just good enough to trick poorly trained, unskilled drivers into thinking that they no longer have to pay attention. This works fine ... until they encounter an edge case that FSD isn't programmed (or trained) to handle.
I thought about that as well. At the moment, CarPlay can't take over every function so carmakers still have to make their own. But in the long term, there is an opportunity for Apple to make the whole thing and monetize it as a cost-saver to carmakers
No, never. This would mean to concentrate only to customers with Apple devices. Why should a carmaker do this?
I assume Apple would also provide code for a fallback UI that offered basic functionality for when you didn't have an iphone attached.
CarPlay poverty mode. Sign me up.
Basically Valet mode for the headunit.
So everything turns green?
Drivers don't car, but I think car buyers actually do. Remember that many people are not buying cars for themselves but for other people, usually family. They fall into the trap of thinking that those other people might want such features, if not now then in the future. Look at automatic transmissions. I know many people who much prefer manuals, but they always end up buying an automatic because they believe that other people will want the automatic. And a few years later, all the cars are automatics. The same is happening with in-car entertainment systems. We buy them not because we ant them but because we think other people do.
Can you explain this? I guess maybe the devil is in "many"?
Unless you are single, and even then you carry occasional passengers, many/most cars are used by multiple people. So people who buy cars are thinking not just about the primary driver but about all the other people who will drive/ride in the car too. Nobody buying a car actually wants in-dash 4k movies, but they think that their partner/kids will.
This doesn't really ring true for me. I agree that when people buy cars, they are often not just buying for themselves. But the purchase decision is not usually made without input from the other drivers. Well, a purchaser will probably discuss things with their spouse, and ask if they care about particular features. But in the case of someone buying a car for their kid, it's usually "they'll get what they get, and they'll like it".
Kids are in the back and thus a screen in the back is what’s needed. An iPad or similar will do that, so a place to mount would be good.
By the time they are old enough to be in the front they have their own phone and headphones and care less about a mounted display than they do about a TV.
It's true for some features (the wife (and I, to be fair) were pretty strongly on the "heated seats" side of things) but things like infotainment were not dealbreakers if everything else lined up.
But CarPlay is darn close; I'd not say I'd never buy a car without it but having it means I don't need to worry if the infotainment setup is crap or not, because I won't be using it.
When (if?) auto-driven cars become a thing I'd expect some people will want a giant monitor in their car if nothing else, so they can watch movies, play video games, tele-conference, etc....
Disney's "The Magic Highway" might be dystopian to some but the relaxing vehicle seems pretty cool to me
https://youtu.be/Vo4-rYNGEwE?t=199
Impressive. Funny that one thing they got wrong is cars becoming smaller instead of bigger.
If you could get 5 self-driving single seaters instead of one 5-seater for the same price... probably the 5-seater wins.
But if you can call a reliable self-driving one-seater within 5 minutes guaranteed?
Just a monitor? I'm putting in a lazyboy and full on home theater setup in the back of my self driving car, well self driving van.
I think a key thing to consider is that there are in fact three separate questions at play here:
1 - Does infotainment/software UI differentiation matter in the car market? Is there a significant enough market advantage for having better UI that anyone should care?
2 - If there is an advantage for better UI, is it enough of an edge that would compel you to build your own? Or is it the case where it simply has to be good enough?
3 - If there is enough differentiation to be worth building your own, is your company good enough at software to pull it off?
Personally I think the answer to #1 is YES. I think cars with better UIs - while not sufficient in and of itself - have a market advantage.
Where car makers start veering off from each other is the answer to #2. If you believe that you just need a "good enough" experience to not be actively awful, then you buy off-the-shelf. You see this with Volvo/Polestar and Google Automotive. The "skin" around the stock experience is minimal at best, with only minor customizations.
If you believe that being excellent at it confers some advantage, you'd try to roll your own. This would include folks like BMW and Mercedes-Benz.
Now, where the latter strategy really goes off the rails is question #3. That said, if you believe the answer to the first two questions compels you to roll your own - would you easily surrender to a third-party? Or would you at least try to level up your software orgs to make a serious play?
Android Automotive is already way beyond what manufacturers are releasing as "their own".
So unless it costs a lot of money to integrate, not picking AA isn't about building your own that's better, but it has to be about control. Or at least independance from a big tech, which I can understand.
For what it’s worth I agree on #1. I really like RR/Jaguar’s current UI. When I’m in a rental or someone else’s car and I’m forced to use CarPlay I hate it. Feels like I’ve been pushed to kindergarten and given crayons… any car manufacturer that just expects me to use CarPlay is probably not on my potential buy list.
I might be unusual in my preference but I really expect people to have preferences as strong.
I mean, theoretically they could make money by selling additional services. Tesla sells "Premium Connectivity" for ~$10/mo, though it might be against the license agreements traditional manufacturers have with dealers to sell enhancements directly to the consumer.
IMO it's more about control over the user experience. You don't want your customers' UX to be dependent on the whims of Apple or Google, because now you're implicitly building a long-term dependency with a third party that may not be acting in your interests in the future. You're moving closer toward a future where the vehicle becomes commoditized, and now you have more trouble differentiating from competitors. And keep in mind: it's only very recently that the "Apple car" project was cancelled.
That said, traditional automakers are also famous (or infamous?) for sourcing tons of components (including infotainment systems) from the same parts manufacturers. But I guess at least that retains the ability to pivot and use it as a point of differentiation in the future.
Every traditional European car manufacturer sells services directly to customers. Not sure about the US/Japanese/Korean, though.
Historically, Hyundai/Genesis charged $300/yr for BlueLink connected care. Think my 2023 Ioniq 5 is one of the last (EV) models with that yearly charge, and that year's Ioniq 6 and subsequent models get it free.
Unfortunately, I don't think they're planning on offering it free to my model or earlier, and I don't know if the same deal exists for ICE cars.
Top of my head, most car makers aren't rolling their own. It's either off the shelf with some white labeling or they buy it from another care manufacturer. i.e. Mazda default infotainment can be found in some Toyotas
Not sure if there's any more to it than this, but the Toyota Yaris is a rebadged Mazda 2. I suspect you won't find Mazda infotainment in other Toyota models, though.
I think you're right.
Also, the Mazda infotainment in the mid-late 2010s was made by Johnson Controls. The newer systems might be too, I don't know.
Yeah it's CarPlay or whatever Android uses, or I want nothing to do with it.
Every time I rent a car it's a HELLSCAPE of figuring out whatever crappy UI this brand of car created for that year ... until I get my phone hooked up. Man I just want to get to my hotel not futz with some garbage UI in the garage forever.
I just carry those vent phone holder and use my phone when renting cars.. it's not the best but does the job for getting to the hotel....
obviously i rather have carplay if i could but i am worry about security implication, plugging to a car i don't know
Every single car I have rented in the last ~5 years in the US has had CarPlay.
Regarding "it sucks", if we're talking mapping/directions I disagree. Google maps is really a pretty piss poor application. It hasn't changed in 15 years and it's obvious its maintainers haven't ever driven anywhere using the application, even around the Google campus in MT View, or downtown SF. It's deeply bad. In my experience some car manufacturer mapping applications are quite a bit better. Since they obviously suck at software, who knows how good it could get with the combination of (not Google) AND (competent team)?
Until Apple Maps got offline downloads (in iOS 17), Google Maps was way ahead for offline use, which was enough to have me use it normally as well.
Good. It works.
I use Apple rather than Google as I try to avoid Google as much as possible but it’s the same principle. I want long term stability in software, not new changes put in to earn a PM a promotion.
There's an interesting (and apparently often misunderstood) article called "IT Doesn't Matter" [0]. In it, Carr is largely arguing that IT, as a business differentiator, was over for many of the things people thought were differentiators. That is, things that helped a company (say American Airlines) get a lead on their competitors in the 1960s had become commoditized. Now every airline was offering flight search and booking online (directly and through aggregators). The IT edge had become table stakes, you didn't do it to beat out a competitor but just to stay in the game. And, even more importantly, many of the things that used to be IT differentiators became commoditized.
Car infotainment was once a differentiator for car manufacturers or for classes of vehicles within the same manufacturer. Today, it's table stakes. Not all the manufacturers have figured that out (have any?).
[0] https://hbr.org/2003/05/it-doesnt-matter and https://www.nicholascarr.com/?page_id=99
It's a bit aggravating that a slice of money I pay when buying a car goes to paying for development of the infotainment, which can be shoddy. It's like paying the Windows tax when buying a laptop with it pre-installed.
Imagine if car manufacturers started offering cars with no infotainment OS, and as users having a choice of open source OS distros to install on them..
Personally, the existence of a car infotainment system is a disincentive for me to buy the car. I certainly won't use it, and would prefer not to pay for it.
If that system is required in order to control a car's functionality (windshield wipers, etc.), then its existence is not just a disincentive, but a dealbreaker.
If it was user-focused, they wouldn't make unnecessary changes to the UI every release. If it was user-focused, they would put more effort into refinement and fixing bugs.
No, CarPlay is Apple-focused.
The auto industry is really at a crucial point with how it integrates tech. As cars get more autonomous, infotainment is becoming a key part of our driving experience. I'm wondering, are carmakers at risk of falling behind if they don’t embrace platforms like CarPlay? Or do they have solid reasons to keep developing their systems in-house to keep control over their tech narrative?
Siphoning user data of course. Remember: Over time, every company becomes an ad company.
Auto makers want what everyone else wants: recurring revenue. They want to find a way to sell subscriptions to something. The infotainment system is a potential angle for that. CarPlay makes that irrelevant.
Isn't that exactly what VW did? No more GPS. Just use CarPlay.
The tech giants are not component suppliers you can symbiotically partner with to add value to your product. They are predatory and parasitic goliaths. They are wolves more powerful than governments with designs on your hen-house. It is frankly insane to let them own the primary interface to your customer and auto will likely regret letting them get this far.
Google/Apple infotainment ventures cannot even be called trojan horses given how open they are in their investments and desires to seize the auto industry the moment tech/profit makes it feasible. Unattractive low-margin manufacturing keeps them at bay but they are gambling on "software eats the world" long-term. For now, they will drain every available high margin service for themselves.