I always like to compare it with the gold rush, where millions of gold diggers rushing for gold, but who became rich? The ones that built the railroad and the ones that sold the shovels. So I've always looked at the picks and shovels pieces, potentially, at least in the near term, the best business case. For autonomous driving, that was actually the advances in machine learning, deep learning specifically, around 2012, 2013, 2014. Those were then translated and incorporated into autonomous vehicles. And all of a sudden, V as the research community made tremendous progress.