Sramana Mitra: I think where they’re doing phenomenally well is in advertising targeting.
Rich Green: Yes, is it because they’re just smarter or because they have a simpler problem to solve? Is it much more controllable?
Sramana Mitra: I think you’re right. It’s a simpler problem to solve. It’s structured data versus unstructured data. A lot of that advertising stuff is structured.
Rich Green: Some of the advertising stuff today is still algorithmic versus deep learning where it’s iteratively improving the result versus continuously tuning algorithms with the help of humans. They’re on their way. Sometimes the results you see are positive because it’s affected not only by machine learning but also by humans.
Sramana Mitra: That’s correct.
Rich Green: At the scale of Facebook looking for your friends, involving humans is virtually impossible.
Sramana Mitra: You’ve been doing SugarCRM for over a year but you were doing an IoT company before that. I know that company. What is your point of view vis-a-vis AI in the IoT space that you were working on before?
Rich Green: To be fair, we were not using AI technology in that startup. We were doing basic cloud computing. I have been quoted before saying it’s the second worst buzzword in the tech industry; the worse one being Big Data. IoT is such a catch-all phrase for sensor data gathering and processing. The fact that it has a reactive function as well as a data gathering function is a sideline.
The reality is as we scale the amount of data that can be and will be ingested, it’s quite obvious that the only way to deal with that and to make sense of it is through machine learning. It’s beyond the scope of the human mind to sense patterns in a dataset that is increasingly large. As you apply IoT sensor feeds to those feeds, it just keeps going up. A human cannot deal with that. IoT cannot succeed without machine learning capabilities because there’s no way to make sense of it all other than trivial algorithmic patterns.
Sramana Mitra: If you take these three trends – AI, Big Data, and IoT – and spin all the things you just said, which domain would you say has the best example of combining those three?
Rich Green: For now, the obvious domain seems to be in the security space because the number of sensors deployed is the largest. It’s a very focused function. The sensors are large and growing larger. There’s a lot of funding in that area. Whether we like it or not, that is the deepest well of near-term opportunity for the confluence of sensor data, Big Data, and machine learning.
My view is that dealing with human behavior is going to be the most interesting and challenging problems. One of the reasons I came to Sugar is not because of CRM per se, but because in my past, I have focused most of my energy on building out platforms whether it’s operating systems or Java itself. I’ve been involved in building out platforms and allowing others to create and do some interesting things on them. At some point, you go, “That’s pure technology. What fascinates me is how you merge technology and human behavior.”
It’s a very interesting challenge and CRM provides the opportunity to say, “If you could build insight into behavior probabilities and inclinations-based history, it’s a really interesting vector to analyze. How do you determine the likelihood and proclivities of an individual? I just find it fascinating. CRM is a more personal person-to-person or person-to-many business. It has slightly different applications than what we’re used to in terms of pure B2C play.
Sramana Mitra: It’s a very interesting conversation. Thank you for your time.