 I'll speak from my perspective as a researcher who works in robotics, machines, and manufacturing systems, but a lot of infrastructure gets developed that is owned by specific entities. You have Boston Dynamics that does really cool robots, for example. They don't release in the public domain how to build their robots. You can buy a robot and use it as a platform, but you can't be the platform. And so at the same time, right now if you look at deep learning frameworks, you have like CMTK from Microsoft or TensorFlow from Google or any of the other ones, like which one is going to be the framework and can you be that framework or do you just have to buy into that system? Are they a monopoly? So if we all have to buy into TensorFlow which is owned by Alphabet and Google, then what does that mean for what direction? What does that mean for the politics of the infrastructure? And I think that in the case of Boston Dynamics, most of their funding came from federal agencies, DARPA or Department of Defense. And so why is that infrastructure owned by them? Why is that platform owned by them? In the case of Google, they're pouring in their own resources to build their own platform. So it makes sense that they want to retain ownership over it. Are there other places where researchers like us can exercise insight into the design of infrastructural systems such that they retain democratic platforms or become democratic platforms? If you are driving for Uber, you have to use the Uber app. Can you also just fork the Uber out of it instead and be the Uber? That kind of idea for AI infrastructure.