Login
You're viewing the assemblag.es public feed.
  • Unnamed TNG skant beefcakeresearchfairy@scholar.social
    Oct 15, 2026, 5:06 PM

    I will accept that you have genuine concerns for the welfare of an AI when you extend even the smallest amount of concern toward the workers who are being de facto tortured to curate the LLM training data sets in lower income countries

    💬 1🔄 2⭐ 0

Replies

  • Unnamed TNG skant beefcakeresearchfairy@scholar.social
    Oct 15, 2026, 5:22 PM

    If you want to worry about whether your use of the stochastic next-token prediction algorithm is causing suffering in a morally problematic way that means you should stop

    The answer is an extremely obvious YES

    But if you want to narrow the question to whether the algorithm feels pain and exclude the real-life, well-documented human suffering

    Ask yourself who benefits from that

    💬 0🔄 0⭐ 0