At Scale we built a tool for model debugging in computer vision called Nucleus (scale.com/nucleus) designed exactly for this, which is free try out if you're curious to see where your model predictions are most at odds with your ground truth. Source: over 2 years ago

To address your point about gathering edge cases, which can also be defined as cases of low model fidelity for our use cases, there is active learning and tools such as Aquarium Learning and Scale Nucleus which make it easy to implement into workflows. Source: almost 3 years ago

Foxglove CEO Adrian Macneil will talk about the recurring challenges he encountered while heading infrastructure at Cruise, and how that experience led to him founding Foxglove. Afterwards, we’ll have a live demo of the Foxglove platform, with some specific robotics development use cases. Source: 10 months ago

Let me challenge you on this one: We already know Volkswagen has CARIAD, Toyota has Woven (and TRI), Stellantis has STLA Brain, and GM has Ultifi and Cruise. Source: about 1 year ago

I think it's a shame that Cruise https://getcruise.com/ isn't mentioned thus far. They've been fully autonomous in San Francisco for something like a year, and are piloting in Austin and Phoenix(?) ... No need to own a car if it can be doing dozens of trips instead of paying for parking. - Source: Hacker News / about 1 year ago

I’ve experienced this at Cruise AI myself as an engineer in the Machine Learning Accelerators (MLA) team. Deploying big, bulky models onto hardware constrained environments like an autonomous vehicle with strict system performance limits remain a significant challenge. Friends working at various AI and robotics teams have expressed similar frustrations. Source: about 1 year ago

I'm not sure what you can do to ride in a Waymo specifically, but if you're just looking to ride in a driverless car, you can also try Cruise. They have an autonomous fleet in SF, but I think they only operate after 10pm. More info here. Source: about 1 year ago