Apparently someone got into a debate with a bot in Reddit. Maybe it's not a surprise that the bot was anti-mod, pro-Reddit admin. Make sure you scroll down in the comments to see the second screenshot (it's a screenshot.of the same conversation but in dark mode).

you are viewing a single comment's thread
view the rest of the comments
[โ€“] 13 points 3 years ago* (2 children)

I think they might just be trolling, the LLMs always say "As an AI language model" not program

  • source
  • hideshow 2 child comments
  • [โ€“] 5 points 3 years ago (1 child)

    The only thing they 'always' do is mimic human language conventions based on scraped data. It stands to reason that as a higher percentage of people talking about Chat GPT refer to it using different words, the model would take that into account when referring to itself.

  • source
  • parent
  • hideshow 1 child comment