• ☂️-@lemmy.ml
    link
    fedilink
    arrow-up
    2
    ·
    edit-2
    8 days ago

    see, something i always wonder about with this type of tool is whether or not they are able to algorithmically discard it as fake data.

    maybe they can do that by, say, your entry points? other data that can’t be spammed (ie your identity and verification)? your previous data they already trained on?

    i think this approach is too good to be true, what do you guys think?

    • ButteredBread@sh.itjust.works
      link
      fedilink
      arrow-up
      2
      ·
      8 days ago

      app says it usea personas to make it harder to filter. It makes the noise more consistent by like making the niise or whatevs in base of the persona’s interest and then switches after like a week.