Hacker News Reader

Comment by cdnsteve | Hacker News Reader

Let's hypothesize for a minute, why a technology giant, like Google cares about user photos?

Their machine learning systems are unmatched. In order for machine learning to be more effective they need to be given context. What's the missing key for photos? People. What happens when you add in an authenticated Google user, who they already know everything about due to using other G services (mail, search, adwords) to that machine learning with photos?

My assumption would be they can simply learn more about you. Personalization is ad money, nothing more.

View on HN

Yeah -- because when you provide a friendly, personal service for someone, you're doing it for the ad money and nothing else. Look, you have a point, but it's an extraordinarily cynical point. You're connecting the dots to reveal "the truth", that Google strives to make money, but it could honestly be a lot worse. Personalization is more than just ad money. It's friendly and great for business.

Wouldn't being friendly and great for business be providing customer support? Something that Google rarely, if ever, does?

Google is a product lab that happens to fund itself from its most successful experiment, ads.

I've contacted Google for customer support twice in the last couple of years, and both times they were excellent. First time was when I had an issue with my Nexus 5, spoke to a human got it sorted straight away. Second time was when they changed the VAT status on personal Google Apps accounts (here in the UK). I raised a live chat to query this and had a callback from a guy on the west coast within 20 minutes who seemed happy to take the time to walk me through what was happening and what it meant.

Maybe I was just lucky. shrug

I'm skeptical about this whole "Google doesn't do customer support" thing, especially given the internet's astonishing ability to amplify vocal minorities and some people's desire to jump on anything anti-guy-making-all-the-money.

The interesting interview[1] with Bradley Horowitz (Google's "VP of Streams, Photos, and Sharing") directly addresses that.

Obviously, machine learning will benefit Google. But it will also benefit all the users of this service, by improving the automated assistance of managing your photos (some of which is already pretty neat, like the automatic photo albums Google will sometimes make for you).

If you are not a photographer, and say, you have kids, you definitely don't have time to do that much curation of your photo archive. You might set aside some time time make a share a photo album after a major event like a birthday party or something, but you will never have enough time to manually derive all the enjoyment our of your photos that you could be getting. This problem needs smarter machines to solve it.

Relevant quote:

    Q: You use artificial intelligence to surface photos on a given 
    theme, or find specific people in the photostream. What’s the 
    percentage of getting it right?

    A: It’s good enough. It’s not perfect, in the same way that voice 
    transcription five years ago was not perfect. The key to getting 
    that last percentage which tips it over will come now, when we deploy
    it at scale. Getting all that data will create a virtuous cycle of 
    getting better and better.

[1]: https://medium.com/backchannel/bradley-horowitz-says-that-go...

I see no need to theorize nefarious purposes. Every large web services company* has some kind of photo storage service. It drives users to their other products.

* Google, Apple, Microsoft, Amazon, Facebook, Yahoo, Dropbox, ...

Sure they want to show ads, no question. But also the other perspective: The camera is likely one of the most often used mobile apps. Users create tons of pictures. For being a dominant mobile platform you need the accompanying photo "cloud" solution. If the offer this as part of their play services this is yet another benefit of Google-Android over OpenSource-Android versions (like Amazon fire)

Yeah, nothing is free, they will know everything about us. It's going to be the social data source of google, like facebook comments, messages, events, friends, etc (which are also unlimited). We are the data source for the ultimate ad targeting machine.

Reply on news.ycombinator.com