The ML Marketplace is moving towards increasing levels of Machine Learning abstraction. AUCNET built their own custom model to classify car parts and estimate the price. But that was a custom model built from scratch. Ocado was able to build off the natural language API. So, they parsed results from the NLP API and used it to route customer emails. They were not starting from scratch, they were starting from one level higher, they were starting from a natural language API and they could use the sentiment and entities from that API to help them [with] their routing decisions. Giphy, which is the third level of abstraction on this slide, Giphy uses the vision API to find the text in memes, and it uses Optical Character Recognition. So, here Giphy is able to use a vision API "as is" to find text in memes. A social media company, as another example uses a vision API to reject inappropriate uploads. That's another example of using the API "as is" to solve a business problem. And the final, on the right hand side, is a much higher level abstraction. For example, Uniqlo used design a shopping chatbot using Dialogflow and they were able to use Dialogflow to create a completely new shopping experience. And this is something that you will see over and over again where things used to be done with custom image models. Over time, we're going to go towards increasing levels of ML abstraction following the typical journey of most software, where you tend to build things at a very low level while the technology is still new and maturing. And as the technology matures, things start to get more and more abstract and much more high level. In this specialization, we're going to be focusing on building custom machine models. So, we're going to teach you how to do Machine Learning, and the level of abstraction that we are going to work at will be around building custom models. But be aware that increasingly, you will get to incorporate Machine Learning into your applications primarily in the form of APIs. Of course, somebody will have to build these APIs for the marketplace and perhaps, you are the builder of such Machine Learning APIs.