• Beaver [he/him]@hexbear.net
    link
    fedilink
    English
    arrow-up
    32
    ·
    2 days ago

    I think what really made me skeptical of LLM chatbots was asking it challenging technical questions from my field of work, and it giving me the most bullshitty reddit-know-it-all answers.

    • Hermes [none/use name]@hexbear.net
      link
      fedilink
      English
      arrow-up
      25
      ·
      2 days ago

      I gave it entry level questions for stuff that should really be in the dataset and it gave the obviously wrong answer that people who know nothing give.