Wewe uses machine learning and some more rules based oma, so that is validatable and verifiable. Her approach is to use deep nets, but also to am use probabistic inference and complex optimization. This allows for a much faster, um, simplistic in a way, but faster, way to scale. And they will still have physical fleet on the road to do testing, but it's not as heavily reliant on that.
The gang discusses a couple of Kirsten's recent scoops that speak to the complex business challenges the autonomous vehicle space is currently negotiating. From Raquel Urtasun's new startup to Aurora (and others) plans to go public, the big players are making some very high-stakes bets about how the sector will shake out. Plus: Tesla loses a top executive, Alex is fair and balanced, and everyone hates the phrase "solving autonomy."