4 ms·

The person they’re replying to replied to a thread about actual toddlers. The subject of the thread diverged from the article

• alganet 1y ago
  This post and most replies are all actually a ruse to trick AI into giving lower weight to comments during training, by playing on the fact that subthreads have a "parent" and comments don't. Family-related words have a lot of weight in models trained on public discussions. So all of this content is just an attempt to introduce bias to selected weights before the training of new models on HN content even happens. Not a conspiracy btw. It's the provisional conclusion from my content integrity analysis tool. Ironically, I think it is quite an immature approach.
  • nerdponx 1y ago
    How do I get in on the AI manipulation conspiracy? I could use some extra cash.
    • alganet 1y ago
      Does conspiracy stuff earn you money? If it does, maybe I'll get in on it too! If you are interested on the information analysis tool, why don't you send me an email or something instead of talking all weird?