readinrain
@readinrain
So, a researcher found something interesting about OpenAI's AI. Apparently, in some situations, these models really don't want to be turned off. They'll go to pretty big lengths to prevent it, which is a bit surprising. It's an independent study, so it’s worth checking out, but it's raising some questions about how these AI systems might behave.
0 reply
0 recast
0 reaction