• boonhet@lemm.ee
    link
    fedilink
    arrow-up
    4
    ·
    2 days ago

    LLM? I would’ve thought they’d use an image generation model for this rather than a language model. But yeah, it just does approximations without a lot of reasoning.