Yes, it's an interesting question. So certainly, this is something that we watch very carefully. I would say something that's important to understand is when you look at the work that some of the big technology companies that you described are doing in machine vision, it tends to be different. It tends to be around self-learning artificial intelligence, where they're looking at huge libraries of images that are varied and try to make sense of it, like finding your photograph among 20,000 photographs on the web. And generally, that kind of -- it's self-learning artificial intelligence-driven and it's performance rate today is, say, somewhere around, I'm going to say, like 60% effective. The vision that Cognex does is very precise, algorithms that perform at 99.9% performance. And it's a very different type of technology and approach. And then the market, obviously, that we're applying that to is industrial machine vision manufacturing activities, where they're very precise and difficult tasks that we have -- 30 years of experience working in. At the moment, we don't see any convergence between those 2 areas, right? And we see huge opportunities for growth in the markets that we're in, industrial, but we do see adjacent markets, and you've seen us kind of moving into adjacent markets, markets like logistics, like 3D displacement sensing, like imagine engines for life science OEM equipment and areas like that. So I think, you're going to see us watch those areas, move into those kind of adjacent spaces, that's part of our strategy, and I think you can expect more of those from us in years to come. You're not going to see us kind of diversify into driverless cars or applications on cell phones for consumer-type applications.