So Google relaunched Photos.
It would be a super big deal if it’s a brand new thing, but it’s largely a separation of Photos from Google+, a broadening of the service, and a bump up. If you’ve already been using Android’s photo backup service, with some G+ integration, you would not be surprised by any of the features at all. The unlimited storage, the animations, the easy sharing, the Story features, the gifs, all wonderful and fun tidbits which we’ve been enjoying for near two years. Lots of fun things introduced under the vision of Vic Gundotra who used photography as a way to grow the Google+ community. Incidentally he was a photo lover, and so were many members of the team. G+ has always been strong on serious photography, and that’s where the photo gene came from. But Photos has brought G+ as far as G+ could go, and Photos has outgrown G+. It is now time for Photos to continue to develop not just as a moat for G+ products, but for all of Google’s products.
If Photos would draw that many people to use G+, imagine what it could do for everything else. Google Drive. Gmail. Google Maps. Hangouts. Android. And it could be a persistent service that keeps users in Google, even when they are on iPhones and Macs and Windows and other operating systems. Family and friend photos are a big part of Facebook; taking photos is a big part of iPhone. Photos, could be Google’s way into the social arena, if they are super good at storing and processing photos.
But photos are also a dream come true for Google, the company who’s mission is to organise the world’s information and make it universally accessible and useful. A perfect fit. Use all their immensely powerful servers and network, stuff it with as much text data, location data, identity data, travel data, image data, etc etc, and process it with all the machine learning and natural language processing and AI as possible. If you had the 10x ambition of Google, playing with text and websites isn’t enough, image and video is the next frontier where you could dump a lot more computing power and logic into. Gigabytes of relevant data per person, to be processed. Wow. Why not. If even there was no upside, they would probably love to do this for free.
But oh the monetary cost sure isn’t insignificant, even to a company with billions of dollars. They had to limit the storage amount, and they still do. So how much data does Google Photos actually use? They’ll never reveal but if we guess…
For convenience let’s say there are 1 billion distinct users (android and iOS combined easily more than that given the activation numbers.), and each user stores 1 small gigabyte if photos and videos (power users easily hit 10gb or 100 GB, but I know there are lots of small users of smartphones)…
1 billion x 1 gigabyte = 1 exabyte.
Which is 1 million terabytes. Which is 250,000 4-terabyte hard drives. Given that all cloud data is usually stored minimally in triplicate for data security. Or even totally 6 copies given the myraid backup requirements, that’s probably 1.5 million 4-terabyte hard drives. And hard drives break, lasting probably 3-4 years max. And you gotta factor in server costs, manpower costs, electricity costs, etc. That’s only the storage, what about the servers required to compress the image and resize it, what about the servers used to store metadata, generate metadata, the large gif files, the AI servers and more. It is an impressive amount of computing power to bear on your puny photos. Google-scale as they call it, billions of users, not a few puny millions.
It’s mind boggling.
It also means that their internet pipes are taking in exabytes of data to store. If you thought you could try to make good use of your home fibre subscription, try an exabyte of data. A 1 GBps connection would take 1 billion seconds, which is 38 years. and that’s 1 gigabytes a second, not 1 gigabits. Multiply by 8 if you have a 1 gigabit/s connection. hah.
But what are people fixated over? Why is the service messing up my uploads, why are there duplicate files (apparently there is some backend function that doesn’t store duplicated files! smart), why are my mac/windows uploader apps not working properly, why is this so scary, what are they going to do with my photos, will they run ads, etc. Some justified concerns though.
Will they run ads? In some form or other, it helps with their ads business. Or do you prefer to pay $100 per month for photo storage? If you want to run a full fledged photo backup option yourself, or pay Dropbox, or iCloud $100 a month, go ahead. Yeah sooner or later your photo album size will reach that big! And Google Photos isn’t completely free! If you want to store Original sized large photos, it does generate revenue for Google.
In the same vein, if we were to pay for every Google service, I think we will be broke. $10 per month for Maps? $10 for Gmail? $10 for Calendar? Per person? So roughly $2-300 per month per family? Ouch. Probably more than your utility bills. That’s what gonna happen if the ad business fails. Facebook does the same thing, with less security because they seem to run a madhouse, and they are cheapasses so they store tiny photo sizes, and they won’t allow you to easily export your data if you wish. And there’s no payment option, they want you to stick around. Apple security is sad, look at the iCloud leaks, their Cloud and services mojo is still lacking much. Just buy their hardware and forget their software.
It’s true though, Google isn’t the perfect place for all your photos. Some belong on Facebook, some on Flickr, some of 500px, some on Medium, some of Exposure, some on Twitter, some on print, on the wall. But Google Photos is the smartest and funniest place they could be. If only those smarts could be shared across the networks.
Discussing Google Photos from the perspective of image recognition and computer vision. Moar photos, moar data, moar crunch. Yeah that auto captioning thing was really quite impressive too. Seems like they’ve got alot more to offer.
Photos seem to be one category of personal history that’s easy to share. Emails, phones, location, photos. Everything else isn’t easily digitized, conveniently digitized, or relatively unsecured by default.