Sramana Mitra: I’m going to switch back on that and say that in the CRM world, you will see very solid adoption and very quickly IF it works. Sales people are motivated by making money. If you can create algorithms and technology that will help them in making money faster and quicker, they will adopt your technology like crazy.
Rich Green: The if thing is extremely important. In fact, one of the things we’re bearing down hard on very aggressively in this first generation of release is accuracy. When you go out and build a dossier on an individual, you’d better be sure that the thing you’re providing to that salesperson is accurate. It isn’t enough to provide bulk. It has to be highly tuned. They’re much more interested in fewer bits that are absolutely accurate than larger bits that have greater risk. The IF factor is a very thoughtful notion. That’s where we are spending a lot of our time.
On the adoption side, I agree with you somewhat. Logically what you’re saying is 100% true. If you can provide technology that will expedite the work of these individuals such that they’ll make their numbers quicker, it’s a no brainer. This is outside the Silicon Valley and all over the world that I’m alluding to. We cover a lot of the planet. When we polled them, there’s this sociological notion of their capabilities being impinged upon by this technology. They will end up in the scenario you’re describing.
It’s only the speed and course they will take and thus, the type of rollout that you provide in this. When you use Facebook, for people in the Valley it’s hard to know that they’re using deep learning to figure out that the five photos they put in front of you is a lot of machinery. Many people would find that a little creepy if they deeply understood what was happening.
Sramana Mitra: Some of it doesn’t work very well. Some of the judgements they’re making are not the right judgements. Part of me wants to look under the hood. If you take up my time and if you have that many deep learning capabilities, show me stuff that matter to me.
Rich Green: You have to look at the slope though. What is the change in quality per unit time and how do you extrapolate that to something? To you and me, every time there’s a mistake, it’s deeply disturbing because you know how the machinery works.
Sramana Mitra: Partly that comes from the fact that I have trained my newsfeed to do better.
Rich Green: I’m with you. I find it entertaining actually. I’m the kind of guy who runs up at high speed to a Google self-driving car, and I stop just before I get to the car to see what it’ll do. I’d like to experiment with other people’s things. In general, the rate of improvement is quite high. The slope is quite positive. The results continue to improve at the same time that the datasets become larger and larger.