Hell yes, video is data, according to Steve Russell, CEO and Founder, Prism Skylabs. He began by reminding us that there's a ton of video out there — something like 200 times the amount of video that's uploaded to YouTube every day is recorded out in the real world, just for surveillance in industries like retail, for example, every single day.

What are the barriers to more video being uploaded to the Internet and made use of? He cited four: access, computation, privacy, actionability (like mobile apps and data visualization). One application his company is particularly focused on is what in-store video can do for retailers. Gigaom says: "By leveraging customer data, online retailers have vastly improved the shopping experience. Until recently, brick-and-mortar retailers have been unable to similarly optimize at the same speed or scale. With new cloud technology, physical retailers now have the ability to mirror the strategies of their online counterparts, enabling them to critically understand store execution and optimize everything from merchandising to promotions."

Steve Russell is a Silicon Valley veteran with 15 years’ experience in building and managing video technology companies. His latest company is developing applications in marketing, branding, and visual merchandising. He believes privacy and insight can go hand-in-hand using "adaptive computer vision."

He closed with this point: "Sensors are now everywhere and are going to drive more video — it's an exciting future."

Ari Gesher, Engineering Ambassador, Palantir Technologies, just gave an animated talk, starting with the whole history of computing and operating systems (well, compressed a bit), and how it was all so custom (bespoke) — drawing an analogy to where big data is these days. He cited this classic tweet from last year:

@BigDataBorat In Data Science, 80% of time spent prepare data, 20% of time spent complain about need for prepare data.

Here's how Gigaom billed this talk: "The past decade has seen a proliferation of standalone-tools and technology for large scale data processing.  While powerful and transformational, the onus is still on the implementer to do most of the work – 80% of the time is spent on setting up the technology, leaving only a fraction to work on the actual problem at hand.  In the early days of computing, every piece of software had this problem – until operating systems heralded a revolution in building applications cheaply. What does the same innovation look like in the big data space?  How do we get beyond building prototype after prototype?  And what about the elephant that’s not even in the room yet – namely, good user Interface?"

Whew! To get Ari's take, watch the video of his talk on the livestream. And follow Ari on Twitter @alephbass.

This was an amazing session — the speaker's personal story made us all gasp! Wow, follow Ramona Pierson, cofounder and CEO of Declara on Twitter (@ramonapierson). Better than me trying to quickly recap it, watch the video here.

She talks about how machine learning can power platforms that make sure the right people and right content find each other without relying on who they know. Ramona spoke about some great work Declara is doing in Australia right now. (Shout out to my home country!)

Om Malik did a great on-stage interview (as he always does!) of Ben Medlock, CTO at SwiftKey, and Tim Tuttle, CEO, Expect Labs. Check out these two companies — I wish I could type fast enough to cover their insights! (Maybe their technologies can help?)

Here’s how Gigaom billed this session in the program: “Despite the hard work that goes into building systems for deep learning and other methods of understanding human language, users might never know they’re powering their favorite apps. And that’s kind of the point. Hear how voice and text messaging services are learning to predict what users will say to deliver a seamless experience.”

Tuttle especially had some fascinating comments. He said it will only be about six years until we smartphone users have a terabyte on our device! “That changes the game.” No lie — talk about the smartphone as brain…

Cloudera CEO Tom Reilly is on stage with Tom Krazit, Executive Editor of Gigaom. He's now talking about the $160M(!) his firm just raised — from T. Rowe Price (which invests primarily in public companies), Google Ventures, and Dell (led by Michael himself).

How soon will Cloudera be a $1B company? "In a large, accelerating big data market ($50B?), we believe we'll be the largest company?" How soon will you IPO? "We have a long way to go before that." (They just hired their first ceneral counsel!) He's done it before, so he's quite familiar with the process. "Gotta get off Quickbooks first." 🙂

Talking about product and their Enterprise Data Hub: "We have dozens of companies behind us building the platform. We don't need to build up," said Reilly. Cloudera is hiring more engineers to ensure the integrations with partners go smoothly. "Predictive support capability" is part of version 5.0. "Machine learning and SQL are working side by side."