• ExLisperA
    link
    fedilink
    English
    arrow-up
    1
    ·
    46 minutes ago

    Sounds like BS. 100,000 words and phrases? Looks like more commonly written phrases always win over similar, less common ones. You can probably find millions of such less common phrases. It doesn’t mean they are actually listed somewhere in the system.

      • ExLisperA
        link
        fedilink
        English
        arrow-up
        1
        ·
        14 minutes ago

        The source article from routers covers this better:

        His curiosity piqued, Kauffman said he used a combination of OpenAI’s Codex and Anthropic’s Claude Code to hunt for a reason why, eventually stumbling on a compendium stored “deep in the bowels” ‌of Apple’s operating ⁠system. He said there were different ways to count the number of gray-listed words and phrases and their various permutations, but a conservative estimate would put them at about 100,000

        So he’s talking about phrases and permutations in multiple languages. I guess 100k can make sense in that context. They also link to this. It explains how some words are tagged as “sensitive”. It just sounds like they want to make sure the keyboard doesn’t autocorrect to something offensive. I think there are about 100k bigger issues than that.