Andrew Yourtchenko28 июл., 16:02We make the heavy machinery do annoying beeps when it does in reverse, such as to prevent harm to humans. As such, I do think the current behavior of Claude is absolutely correct in the grand scheme of things.
Andrew Yourtchenko28 июл., 16:04*goes* in reverse. its year 2026 and we are arguing about the behavior of intelligent machines, and then there’s iPhone autocomplete…
NotHarlock28 июл., 16:21since it's a chatbot, many of those clamouring work from a frame of reference where the AI has human-like attributes (even if they know it doesn't, and even if they say it doens't) - different from autocorrect or a forklift those that would clamour in the opposite scenario, aren't doing that
NotHarlock28 июл., 16:21it has a persona, and a lot of marketing value in that. so, I don't think people are following consistent internal rules when debating consequences around chatbot actions, because the nature of a chatbot that can have 'you did the wrong thing' attributed to it, is novel (and maybe inappropriate)
Andrew Yourtchenko28 июл., 16:26Forklifts absolutely do have personas. Those who think they don’t didn’t know many enough forklifts. As opposed to the iPhone autocorrect, which is genuinely just a destructive force of nature.
mr. TIM28 июл., 17:03omg i’m not paying attention to this sub-thread but this post makes me think i should
mr. TIM28 июл., 17:22it’s official, i just tried to catch up and frankly i have no idea what you guys are talking about