• LainTrain@lemmy.dbzer0.com
    link
    fedilink
    arrow-up
    2
    arrow-down
    2
    ·
    edit-2
    6 months ago

    No profit in that, though, so doubt any corpos will do it.

    But in theory it should not be hard to do, there has to be a way to extract YT’s subtitles by scraping the page in something like Selenium (might be some trouble if they are really generated at runtime or in batches rather than once per video at upload and sent from backend during beginning of playback and simply made invisible in the source), then simply run a summarize prompt on the text to a local Mistral or w/e and have a result. If using 'Open’AI’s API, could even have it be a Firefox extension with mobile support.