• RightHandOfIkaros@lemmy.world
    link
    fedilink
    English
    arrow-up
    2
    arrow-down
    1
    ·
    3 months ago

    Could have been as easy as the programmer’s personal chats. Humans have so many conversations every single day that it would be very fast to gather up enough data to train a model on. Give it a few days worth of conversations from 50 people in a development team and youre already at a very large dataset.