Tesla AI Director Discusses Full Self-Driving on ‘The Robot Brains’ Podcast

Tesla AI Director Andrej Karpathy has been working on Tesla’s Full Self-Driving (FSD) platform for several years now, and in a new Robot Brains Podcast with Pieter Abbeel, Karpathy details his path to working with artificial intelligence, how Tesla’s FSD is going, what it’s like to work with CEO Elon Musk, comparisons between Waymo and Tesla, and more (via Reddit).

When asked about working with Elon Musk, Karpathy said, “He’s an incredible person. I’m still trying to map out his superpowers – incredible intuition, even with a lack of information. Great judgment. He’s a double-edged sword because he wants the future yesterday. You have to have a certain attitude to tolerate that. If you can, you will thrive at Tesla.”

The interview also featured a few nuggets about Tesla’s FSD software, like how the company doesn’t outsource its data labeling, as well as its “low definition” maps model compared to Google company Waymo’s “high definition” mapping and sensor technology.

By using a set of low-definition maps instead, Karpathy explains, Tesla’s FSD can offer up simple commands, like “left and right turn,” to account for the many circumstances a human driver might normally encounter.

Karpathy also suggests that Waymo doesn’t have enough cars to collect enough data and solve FSD, saying, “Scale is incredibly important for dataset curation. I would rather trade sensors for data.”

In the interview, Karpathy seems to think that humans can expect major improvements to automation in all areas of life within the next few years – from drones, to cars, to warehouses, and more, Karpathy projects the next few decades to be incredible in the AI space.

You can listen to the full podcast with Andrej Karpathy below.

Self-Driving Technology for Electri...
Self-Driving Technology for Electric Vehicles: Tesla's Aim for the End of 2020 #shorts