I’ve been thinking about this more clearly lately. As Physical AI keeps developing there’s another problem that matters just as much as the robots themselves What are these robots actually going to learn from? Because pulling millions of data points from the internet is one thing. Capturing what a real street actually looks like is something completely different. That’s where I find @vangrid_io interesting. You can use a smartphone to capture real world data. People can collect the locations that are actually needed while the data can also be traced back to where and how it was captured. There’s Android web based capture USDC rewards and a live network you can actually explore. So this isn’t just an idea sitting on a roadmap. The question I’m really curious about is If Physical AI is going to scale who is going to provide all that Real world data? I think that’s exactly the problem Vangrid is trying to solve.