Reciprocity
Active Member
There are basically two kinds of data in this, or so the argument goes: the kind you can use to train a neural network - and the kind of data that might be useful otherwise to train or validate your auto-driving system. The assumption so far has been that it only sends the latter and that former is done by Tesla themselves.
Any comments on that based on this data amount? Or this summary? I am no expect, so I will not even begin to guess.
If they are only sending the data to validate as you assume, what data are they using to train the system? Employees driving around a hand full of cars? It does not seem to be enough raw data to train the system, but I could be wrong.
To me it seems like the right amount of data to send almost everything the driver is doing in 6500 miles over the 4-5 months. Unless we are all thinking they are not smart enough to compress video and raw data.
They are clearly using the onboard supercomputer processors in the car to make decisions about what to do in certain circumstances so it would also make sense that they could use it to decide what things it does not understand and mark that raw data to be saved long enough to transmit it over Wifi back to the mother ship, though based on the 1-1.5g per day for a guy who drives an hour a day seems about right for compressed video/image and data from 8 cameras and sensors.