• Irdial
      link
      English
      346 months ago

      LLMS are not (currently) involved. The article states that video content, trained to imitate the likeness of a celebrity, is generated to recite human-written information. Or so they say