2/ Ground-up EV design. Space for four passengers + luggage.

Built for short, urban rides, though highway-capable up to 75mph.

4-wheel steer (crabwalk). Bidirectionality is good for tight urban spaces (U-turns).

— Taylor Ogan (@TaylorOgan) December 14, 2020

4/ The vehicle has a 133 kWh battery and can charge wirelessly. In urban environments (essentially operating strictly as a LSEV), that would last well over a full day of operation. Zoox says 16hrs of continuous use between charges.

— Taylor Ogan (@TaylorOgan) December 14, 2020

6/ Mapping seems important for @zoox only for semantic segmentation & localization. As Zoox expands to new cities, it needs only 2-3 days to "map" new environments. Mapping of this scale in such a short amount of time shows how little its system relies on mapping.

— Taylor Ogan (@TaylorOgan) December 14, 2020

8/ Zoox CEO @aicha2evans has always said safety is the priority at @Zoox. With its AI, perception output bias always leans conservative. That means the system would rather misclassify a mailbox as a pedestrian rather than misclassify a pedestrian as a mailbox. pic.twitter.com/P8AXTFtv7C

— Taylor Ogan (@TaylorOgan) December 14, 2020

9/ Almost everyone's systems do this, but there are a few overly ambitious AV systems NOT doing this. I won't name names.

— Taylor Ogan (@TaylorOgan) December 14, 2020

The @zoox pods look similar to Didi/BYD's new D1–the world’s first EV built specifically for ride-hailing. There will be 100,000 of these soon-to-be L4 AVs on the road within the next year, and 10M by 2025. By 2030, they will be L5. https://t.co/pTX0RWK7aH

— Taylor Ogan (@TaylorOgan) December 14, 2020

Follow @snowbullcapital Follow @TaylorOgan