Hacker News Reader

Comment by CPLX | Hacker News Reader

> But there is just no way Apple is going to beat Google to market with that technology.

Why?

Apple is possibly the world's most effective company at synthesizing industrial design and software in premium devices sold to consumers. Google is a web/cloud services and advertising company.

The reason presumably is that you can't see what Apple is doing in public, while Google has a car that still can't drive itself in the rain that is occasionally seen wandering around Mountain View. I question the methodology of that analysis.

View on HN

For one because there is no evidence of it. Apple are secretive but at some point you have to take a car out on an actual road to see how it fares. And get government permisison to do so. The fact that Apple has done neither leads me to suspect Google is further ahead.

Android was in development and publically known about for years before Apple unveiled the iPhone. Now all new phones are basically large touch sensitive screens.

So who knows?

Yes, but the regulatory and development hurdles involved in developing an operating system are significantly different to those of developing an autonomous vehicle.

Government permission was quietly granted already. It was just quietly done.

Source, please?

This logic is flawed. If you want to test something secretly you do it in plain sight and use mis-direction. Buy some Tesla vehicles and put your electronics inside -- can you tell who owns the software and hardware inside when looking at the outside? Use a Chevy volt for the really juicy stuff, nobody will even blink. If you need to test a fancy new sensor, buy a white car and slap Google on the side, everyone knows Google is testing this stuff, so meh!

That's a silly assertion to make. Apple is well known for secrecy, and any prototype vehicles driving on the streets are not likely to be branded with their logo. How do you know they aren't already driving vehicles on the roads?

Also, whether they are currently road testing vehicles or not, that doesn't change the fact that Apple is increasing R&D spending.

They seem to frequently fail at software. iTunes is a dog, Mac OS X is getting worse and worse, the Mac App store is horrendous. iCloud is a joke.

They do nice interfaces, and iOS is great. I have no doubt they would produce an amazing luxury car with a beautiful software UI, but I do not see them morphing into a machine learning software powerhouse anytime soon, at least compared to Google. Google Now easily beats Siri and Google Maps beats Apple Maps.

Google Maps' advantage has a lot more to do with data than algorithms. Google Now is probably a better example, though I've never used it myself.

"Google Maps' advantage has a lot more to do with data than algorithms."

Given literally everyone else collected the data well before google, and maps was still better, this seems wrong.

Google developed better ways to get the data and automate it, not just got better data. For years, literally everyone else had better data.

Not really. Google collect more, like wifi info :/ and images (street view). Google really did revolutionise mapping.

Google Maps was better before they did that.

Apple is not even trying, or is leaving their Maps team terribly under resourced. I've been running some experiments submitting the same edits to Google, Yelp & Apple simultaneously (since Apple uses Yelp data), and timing how long each takes to update. The last time I publicly recorded my results:

Google Maps: 40 minutes

Yelp: 3 days 20 hours

Apple Maps: 13 days

https://twitter.com/syneryder/statuses/668473075365642240

I know Apple's policy is to screen edits carefully to ensure accuracy & avoid vandalism, but it's actually making their maps more inaccurate, because it takes longer for errors to be fixed. Google has also started incentivising frequent editors with rewards (like free 1TB Google Drive subscriptions), which helps encourage people to make the corrections in the first place.

OpenStreetMap: 0 minutes

That wouldn't be good, either, since it would mean no quality control of the data going in.

Most first-user contributions to OSM are reviewed by the existing community: for example, many of the regional OSM IRC channels have a bot which pops up first-time edits, so a regular can check on them. It's just that they're retroactively moderated rather than proactively, which avoids new users being disillusioned when their edits don't show.

You can only really do the basics on a test track. If they want to tackle real-world driving they'd have to take it onto real-world streets, and that would probably be impossible to hide.

While obviously Google is historically a web/advertising company, they're showing a lot of strength recently in machine intelligence, like with Google Photos auto-tagging or AlphaGo.

If you say that Apple has more relevant experience, then why doesn't Audi, BMW, Daimler, Ford, GM, Google, Kia, Mercedes-Benz, Nissan, Renault, Tesla, or Toyota beat them both? All of these have more relevant experience than either Google or Apple, because industrial design is not really what car is about.

I would say the chances of one of those companies you mentioned being more successful than both Apple and Google are extraordinarily high. Nothing in your statement conflicts with mine.

>>Google is a web/cloud services and advertising company.

Google is actually an AI company.

Reply on news.ycombinator.com