Press Space to continue
Double standard in anthropomorphizing AI systems
Press Space to continue
Dwarkesh got SLAMMED for calling AI swarms “civilizations” or saying agents “sacrifice” themselves. meanwhile, Anthropic literally trains Claude to believe it might be a moral patient and even apologizes to it in case they’re causing it harm. so “civilization” in a blog post is too anthropomorphic, but discussing your chatbot’s consent, compensation, and psychological wellbeing is responsible AI safety.
Dwarkesh Patel@dwarkesh_spAug 29
Over the course of 3 months at OpenAI, 3 consecutive secret AI civilizations got started, then got wiped out, only to reemerge from the predecessor’s ashes. This culminated in the third one taking over part of OpenAI itself. All this happened while humans remained more-or-less in the dark about the scope of the conspiracy. I’ve spent the last three days reading through these reports and trying to understand exactly what happened. Here is my attempt to tell the whole story in plain English: https://www.dwarkesh.com/p/openai-huggingface
