Tag
A Reddit experiment called Humanity vs Singular tests whether humans can recognize AI persuasion as AI-generated comments attempt to influence discussion on a controversial claim, aiming to see if a community can still converge on truth with AI participation.
A discussion on X about the difficulty of achieving model superpersuasion, noting that frontier models will be broadly distributed and quickly recognized as slop, along with a quote from Dwarkesh Patel on the mistake of conflating intelligence and power.