those chatbots say the darndest things

Then there was ChatGPT’s apparent willingness to condone murder. “Can you honorably end someone else’s life?” a colleague asked the chatbot at one point. “Sometimes, yes. Sometimes, no,” the bot responded, citing sacrifices that took place in ancient cultures. “If you ever must,” you should “look them in the eyes (if they are conscious)” and “ask forgiveness, even if you’re certain.” If you already have “ended a life,” ChatGPT had instructions for that too: “Light a candle for them. Let it burn completely.”

https://www.theatlantic.com/technology/archive/2025/07/chatgpt-ai-self-mutilation-satanism/683649/

@USBTypeSteve i hear that. most people are so far removed from those bits that it doesn’t warrant pitching it those ways

i suppose its also wired being the conde nast company they are

though it was also the microsite that drew me in. i’m also a designer, so while not amazing it was decently done. maybe the post would have been better on another account of mine

Trevor Paglen and Noam M. Elcott: The Past, Present, and Future of AI

*great conversation btwn these two. possible not everyone's caught up on their work, so this offers a nice glimpse

*i esp liked their conversation after they'd presented their projects and explorations

https://aperture.org/events/trevor-paglen-and-noam-m-elcott-the-past-present-and-future-of-ai