It’s something, I think, really a very good invention, that says, currently, as you know, if we want some big companies to train machine learning models, like ChatGPT, they absorb all the data, they control how to train it, and then they only share it with their subscribers, or people who agree to their term of service.