How are the employees sifting through millions of hours of video to find the ones with naked people?
Joke aside, probably the same way Roomba employees posted pictures of people in the bathroom or Apple partner leaking recording of people having sex: horny employees + zero oversight from mgmt.
SELECT *
FROM tesla_telemetry_videos
WHERE town = 'Austin'
AND postal_code = '78733'
AND content LIKE '%funny%'
AND (sex = TRUE AND nudity = TRUE)
AND violence = FALSE
AND cursing = TRUE
ORDER BY upload_date DESC
LIMIT 20;All you had to do is use the built-in "goatness" evaluation function!
How else would you find the naked riders in the haystack?