Live data from Hacker News

Apple's risky bet on CarPlay

theturnsignalblog.com

401–410 of 537 posts

Re: Apple's risky bet on CarPlay

#401
post #81
post #4

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…

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 navigati…

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!

Re: Apple's risky bet on CarPlay

#402

Earlier quoted context omitted.

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…

Could we please for once stop using the term "AI" when we mean "heuristics"?

Re: Apple's risky bet on CarPlay

#403
post #4

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…

> "So they have control? Control over what, exactly?" 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 y…

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.

Re: Apple's risky bet on CarPlay

#404

Earlier quoted context omitted.

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.

So everything turns green?

Re: Apple's risky bet on CarPlay

#405

Earlier quoted context omitted.

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…

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).

Re: Apple's risky bet on CarPlay

#406

Earlier quoted context omitted.

>so the most sustainable design is one where the software is easily seperable and replaceable from the core durable item. 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 repa…

> 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. 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.

Re: Apple's risky bet on CarPlay

#407
post #383

Earlier quoted context omitted.

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?

Re: Apple's risky bet on CarPlay

#408
post #34

There's an interesting gap when discussing carplay/aa. Cars from the 90s up until about 2013 can be easily fitted with a $500 head unit upgrade, and support carplay quite well. With the right tools, it can be done in about two hours right in your driveway. Cars from 2018 and up pretty universally support carplay, and it's generally quite well integrated into the car's infotainment system. But, between 2013 and 2017,…

"...since only 21% of people will buy a car without carplay" That statistic is complete bullshit.

I don't have an iPhone but I wouldn't buy a car without Android Auto - and it's usually the same cars that support carplay.

Anyway it's pretty much moot because besides Tesla, all new cars support both Carplay and AA.

Re: Apple's risky bet on CarPlay

#409
I went from CarPlay (Kia e-Niro) to Tesla (Model 3) to CarPlay (Volvo XC40 Recharge) and I vastly prefer CarPlay. I've tried wireless CarPlay (BMW M2) and while convenient when entering and exiting the vehicle I found it to be quite laggy compared to wired. While Tesla does have the Apple Podcasts app it's a really bad experience. It won't sync playback state properly so I have to memorize what I'm listening to and how far I've gotten when I switch from car to phone or back. Similar story with Apple Music and even Spotify (which is usually great at keeping state between devices but not with Tesla for some reason). Yes, I paid for "premium connectivity". Tesla did have better route planning though.

I also really don't like wireless charging. It spares the port from constant wear but my device tends to get really hot and the charging is very inefficient. I plan to go back to slow wired charging everywhere to put less strain on the battery. I already have in my car and by my PC. Unfortunately my XC40 doesn't support dual screen CarPlay so I use the Google Maps in Android Automotive for navigation and use CarPlay for music, podcasts, etc. Switching from AA Google Maps and CarPlay and back is not a high friction interaction though so it's an okay solution. I'm also able to deactivate wireless charging so I'm pretty happy with this solution.

Re: Apple's risky bet on CarPlay

#410

Earlier quoted context omitted.

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…

Could we please for once stop using the term "AI" when we mean "heuristics"?

heuristics and Meta-heuristics are AI. Heuristics are how children (and adults) learn to solve new problems.
Post reply on HN