And it's funny to me, one of the most important aspects of it that I think stands up really well, and it's been 15 years, is that so many of these AI products that leaving the models and the underlying technology aside, these specific products that have taken the world by storm, you can really tell that the AI labs themselves did not know they were going to be as popular as they turned out to be. Obviously ChatGPT, they had no idea. Claude Code, Cowork. These were small experiments in the grand scheme of things. They weren't the companies like, "This is our big bet. Let's go." And that's just so classic that that's a universal aspect of product development, that you do not have the ability to predict the future. And when you pretend, you get yourself in trouble. If you hold everything like a hypothesis, you get the benefits of the scientific method. It's pretty helpful.
And it's funny to me, one of the most important aspects of it that I think stands up really well, and it's been 15 years, is that so many of these AI products that leaving the models and the underlying technology aside, these specific products that have taken the world by storm, you can really tell that the AI labs themselves did not know they were going to be as popular as they turned out to be.