• Sludgehammer@lemmy.world
    link
    fedilink
    English
    arrow-up
    5
    ·
    edit-2
    6 days ago

    I’ve said this in the past it’s be pretty easy to do a blind or double blind test.

    Just have someone choose some random product(s), prompt a LLM to generate conversations about said product(s), run those into a text to speech generator, then put a earbud up to the mic on the phone and play X hours of the generated conversations into the mic. Give the phone back to the owner and see if they can determine what the product(s) were in a set time span.

    • Goun@lemmy.ml
      link
      fedilink
      English
      arrow-up
      3
      ·
      6 days ago

      Good idea! Would love to see the results from anyone testing it.