Would it bother you if someone used your writing / tweets to simulate you for private conversations?
Should there be a particular exception for AI Safety researchers to remain un-simulated because the LLM has "motives" for "deception" to avoid being shut down by good arguments against AI?
haha ok tangential issue but needs a gender split... is it net positive or net negative if someone uses a chatbot to roleplay as you for their sexual gratification?
More notes on nearby ideas
Repost · Feb 18, 2023RT @bio_bootloader: what if we are in a sim run by AI safety folk, watching to see if AI escapes the box and kills everyone here before the…Reply · Oct 7, 2022@deepfates You know ai safety ppl were never serious because they never consulted aella on this topic 😈Reply · Jan 9, 2023@twofifteenam convo started as finding out what idle assets we have in the datacenter... might get a vm for running AI models... and then took a turn…Reply · Jan 9, 2023@twofifteenam we were talking about using LLMs on my tweets to make a fake-mimi, and they had a pretty accurate guess at the likely contents of my…