That "raw sensor input" is massively, hideously spatially and temporally redundant.
Even the shallow encoding patterns of AV1 can easily crush a raw video stream down by a factor of 10^3 without losing either spatial coherence or semantic loading.
Humans live in self-similar environments and perform self-similar activities. Any practical training dataset has the "natural human data" beat on diversity by a crushing amount. And a system that can learn deeper patterns makes the sensor streams more redundant, not less.
Even the shallow encoding patterns of AV1 can easily crush a raw video stream down by a factor of 10^3 without losing either spatial coherence or semantic loading.
Humans live in self-similar environments and perform self-similar activities. Any practical training dataset has the "natural human data" beat on diversity by a crushing amount. And a system that can learn deeper patterns makes the sensor streams more redundant, not less.