TLDR · QUICK LINKS

OPENAI SAYS ITS MODELS ARE MORE PERSUASIVE THAN 82 PERCENT OF REDDIT USERS

arstechnica.com5 min read

Summary from the TLDR newsletter

OpenAI rates o3-mini's persuasive capabilities as a 'Medium' risk, meaning that the model has comparable persuasive effectiveness to typical human written content.