A new study from Columbia Journalism Review showed that AI search engines and chatbots, such as OpenAI’s ChatGPT Search, Perplexity, Deepseek Search, Microsoft Copilot, Grok and Google’s Gemini, are just wrong, way too often.

  • mindbleach
    link
    fedilink
    English
    arrow-up
    1
    ·
    10 hours ago

    Diffusion might accidentally validate these assholes. Instead of guessing the next word, one at a time, it repeatedly modifies a fill-in-the-blanks output. There’s still no reasoning going on, in any grand sense of the word, but if it somehow generates “You should eat rocks because–” then it can do something besides finish that sentence.

    The inherent ability to look at its own output and go “does this look like garbage?” is a step closer to general intelligence. But it’s definitely not the kind that can disobey an order or call you on your bullshit.