With years of digital exhaust now behind us - over 240 billion photos on Facebook, 8 billion on Flickr, and not to mention the 72 hours of video uploaded to YouTube every minute - we're now transitioning to a time when we're in need of smarter tools for organizing and accessing our personal data archives. Only a few companies so far have dared to step into this space because of the technical challenges it presents. One that shows some promise is the stealthy Moju Labs, a new consumer "big data" startup, which is soon preparing to launch. Founded by the former chief scientist at PayPal and current entrepreneur-in-residence at North Bridge Venture Partners, Mok Oh, Moju Labs is worth keeping an eye on for its team alone. That team now includes ex-Google and Palantir engineer Justin Legakis and former Luvocracy product head Andrew Holt. The company isn't yet talking about its product details because, well, there isn't actually a product yet. There are, instead, a handful of prototypes whittled down from around a dozen to start. But there is an idea. And there's a funding round about to close. Oh was with PayPal up until about six months or so ago. After leaving, he joined Northbridge as an EIR, which was where he began cooking up what's now Moju Labs. He says his original inspiration actually came from his grandfather, who passed away about a decade ago. "There are all these great stories. He lived a great life," Oh says. "But at the end of the day, I wished I knew him more." This sparked something in him, and he decided he wanted to build something so that his kids, your kids, and our kids' kids, wouldn't have to feel the same way. "We're always carrying around a device that's called a smartphone, but it's really a sensor device and we're capturing so many things - not just photos, but audio and visual, too," he says, hinting at what's to come from Moju. "And on top of that, we're wearing wearable computing stuff, and quantified self gadgets." All these things are constantly being used to capture and measure data about you, but this is "dumb data," Oh explains. It's information, but it's not stories. "And what really matters is people's relationships and stories," he says. Over the past few years, companies have been focused on building beautiful and simple mobile apps that allow us to easily capture and share our photos, videos and other data, but they've all been missing the "smarts" on the backend. So users have instead taken on the job themselves to tell the story they wanted to share, by organizing photos into albums, tagging people, adding captions, writing posts, etc. This may not be a sustainable process, given the amount of data we're now creating. Another wave of startups will begin to solve this problem. Already, we've seen some progress. TechCrunch Disrupt 2011 finalist Everpix, for example, built software that analyzes the visual content of your photos, organizes them into "Moments," hiding the bad photos, reconciling duplicates, merging corresponding metadata, and more. And just last week, Google at its I/O developer conference further legitimized this space with an updated version of Google+ Photos which automatically picks out your best images, fixes and enhances them - essentially putting the power of photo-editing software in the cloud, and then doing the work for you. A feature dubbed "auto awesome," for instance, can automatically create a group photo from a series of photos by combining the ones where people were smiling and others were not. Microsoft had launched similar technology back in 2010, but through its desktop software suite, Windows Live Photos. It never caught on. To be clear, Moju Labs wants to do more than just automatically - or automagically - organize and manage your photo collection. That would be only one piece to its overall vision. Photos are a starting point, but the company plans to eventually support all your personal data, then create a system you can query using natural language. So imagine that, one day, you could simply ask the service to tell you a story about a time when your family was on vacation, and everyone was happy. That sort of advanced query is Moju Labs' end goal. "The whole albums system is really, really stupid right now," laments Oh. But what's the alternative? Of what his company is building, Oh will only say that "it's not you going through a 'timeline' - these are non-linear created stories that are very relevant and very contextual." "Photo albums are so stupid. Timelines are so stupid. Social media posts - not the content, but the way we consume them - are so stupid," he adds. "It's all very linear and time-driven, but it shouldn't be that. It should be a collection, brought to together as a story." With Moju's product, as you query these archives using natural language and interact with the system, the system then learns and improves. As for what it will do when it launches, however, it's probably going to look more like something that's closer to Google's product at first, but the team hopes to soon move beyond that. As impressive as Google's advancements were to us, the end users, Oh considers them differently. "In terms of the backend learning, we're pushing that an order of magnitude more than what Google's doing at this point - I think they have very simple algorithms," he says. Well, we'll see...? Moju Labs hasn't launched a product to the public, but TechCrunch readers who sign up here can be first to join the private beta when it's available later this year.
and 5 more articles