 As machine learning continues to enter the mainstream, I think people are becoming more aware of some of the challenges around the quality of these models that are being trained. And as a practitioner, I myself find it challenging to, for example, create a classifier that doesn't have too many false positives, too many false negatives. And in fact, how do you even define too many? Like, what does that cost function look like? It gets a little bit tricky. And so it's understandable that there's a lot of concern over how is the government using machine learning, for example. As it becomes more widespread, is there an opportunity for... Is there a cost for concern over false positives, false negatives, the various traps you can fall into when you're training a model on data? So certainly that's something I start to see getting a little bit of traction, just on the web. This notion of are these algorithms skewed or biased in some way? Are there issues with the data? Are there issues with the way a model is trained? How do we guard against that? How do we... If it's something that affects our lives, how do we kind of peer into the black box to understand what's actually going on? Looking at the arc of the whole big data phenomenon, it really started at the end of 2008 with the founding of Cloud Era, which was the first company to commercialize Hadoop. You fast forward to today, it's a public company with a multi-billion-dollar valuation. And the other arc that we've seen as big data has become more widely adopted, the other arc has been on the machine learning side. So what do you want to do once you have all this data? Well, you want to make sure that you're getting really good insights out of it. And so Big ML hit the scene at the end of 2012. In 2015, Amazon and Microsoft released their own cloud-based machine learning offerings. And we're proud to have really kind of led the way, not just chronologically, but also I think in terms of usability, having an interactive interface that an analyst can actually use, having something that's very visual and that encourages free play with the data. I think that's really what makes Big ML special. What I'm really excited about is the opportunity to take these algorithms, which have so long been the domain of academia, and bring them into the mainstream. And the challenge there is not the algorithm. Many of these algorithms are well understood. The problem is this design challenge of how do you actually make them interactive, make them visual, make them something that a non-technical user can actually benefit from. And so I think that's really what excites me. I think that's what really drives Big ML is how do we solve this design challenge? How do we actually create an interface that makes machine learning really easy?