It is really sad to see people leading Tesla's AI with no ability to think in a wider range. They have no solution for:
- dynamic obstacles (cats, dogs, small kids) suddenly appeared out of the camera's angle. There is no front camera so car cannot see what's really in front,
- the size of obstacle (typically the curb),
- various weather conditions (e.g. fog),
- dirty cameras (all cameras except frontal ones has no system to self clean-up and especially rear camera is getting superdirty very easily,
- no backups (brake the B-slope glass and you have no ability to drive until the car is repaired,
- etc.
BTW, once you work with more than one single camera, you already have to work with high enthropy coming from various data sources. Different cameras, different resolution, different view angle, different position, different glare, etc. It's the same to have data from another one camera or from radar in addition.
www.notateslaapp.com
