• athos77@kbin.social
    link
    fedilink
    arrow-up
    13
    ·
    1 year ago

    I really hate the ones that make you select pictures that contain particular images.

    In a way, those are interesting because you can use them (sometimes retrospectively) to tell what Google or Google’s clients are working on. First it was all the text stuff as they digitized old newspapers and books and magazines. Then there was that period when you wanted you to identify stop signs and house numbers and businesses and other stuff like that - all of that fed into Google Maps. Then it was traffic lights and speed limits and stop signs, which was early self-driving. Now it’s motorcycles, buses, bridges, and bicycles - all things that were (maybe still are?) proving a challenge for advanced self-driving. The traffic lights and crosswalks fit into this somehow, though I’m not sure if it’s self-driving cars, map directions, both, or something else entirely.

    I have absolutely no idea what they’re doing with fire hydrants, staircases and mountains, though. It’ll probably be obvious in retrospect. But anyway, how do you like your life as not only a data point that Google can sell to anyone interested, but also as a cog feeding data into Google’s many businesses and helping them solve their identification issues?

    • burningmatches@feddit.uk
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 year ago

      It’s illegal to park in front of fire hydrants so you’d want a self-driving car to know that. However, I think Tesla is pretty much the only company using cameras for self-driving cars (rather than lidar/radar), so not sure this is the real reason for the captchas. Knowing where hydrants are would be useful for Google Maps too.