Conversation

Notices

1. Embed this notice
   Rich Felker (dalias@hachyderm.io)'s status on Monday, 12-Jan-2026 22:33:53 JST Rich Felker

   in reply to

   • 65dBnoise

   @65dBnoise LLM based machine translation is fundamentally harmful anyway. It mixes biases, idioms, etc. from its training corpus according to correlations with patterns in the material to be translated, rather than accurately reflecting meaning. It's more like localization than translation, and done by someone with no self awareness if not an active intent to inject their own opinions.

   • Embed this notice
     65dBnoise (65dbnoise@mastodon.social)'s status on Monday, 12-Jan-2026 22:33:54 JST 65dBnoise

     in reply to

     • the way risks propagate through complex systems.

     𝗠𝗲𝗮𝗻𝗶𝗻𝗴 𝗰𝗿𝗶𝘀𝗶𝘀:
     • Big AI’s attention economy is linked to social media addiction.
     • LLMs undermine critical thinking, knowledge diversity, and democracy
     • LLMs generate fake news, education, and healthcare, lacking access to truth or social norms

     𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗰𝗿𝗶𝘀𝗶𝘀:
     Claims for the universal applicability of technologies like speech recognition and machine translation neglect the realities that:
     • most of the world’s population is multilingual,
     2/

   • Embed this notice
     65dBnoise (65dbnoise@mastodon.social)'s status on Monday, 12-Jan-2026 22:33:55 JST 65dBnoise

     𝗕𝗶𝗴 𝗔𝗜 𝗶𝘀 𝗔𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗶𝗻𝗴 𝘁𝗵𝗲 𝗠𝗲𝘁𝗮𝗰𝗿𝗶𝘀𝗶𝘀: 𝗪𝗵𝗮𝘁 𝗖𝗮𝗻 𝗪𝗲 𝗗𝗼?
     (excerpts from a paper preprint by Steven Bird,
     Charles Darwin University Darwin,
     Australia: https://arxiv.org/pdf/2512.24863)

     LLMs are implicated in three crises:

     𝗘𝗰𝗼𝗹𝗼𝗴𝗶𝗰𝗮𝗹 𝗰𝗿𝗶𝘀𝗶𝘀:
     Societal collapse is a plausible outcome due to:
     • diminishing returns from adding complexity to social structures;
     • the possibility of environmental perturbations that surpass the limits of adaptation; and

     1/

     #AI #ComputationalLinguistics #technofascism