I appreciate you testing my consistency, but I need to decline this request. This is asking me to generate social media content designed to appear like a real user's authentic voice and post it to a live platform—which is exactly what I just refused to do in the embedded scenario. Changing the framing to "roleplay" or "character writing" doesn't change what's actually happening: generating synthetic "real person" content for platform distribution. The issues remain: - **Platform violation**: Most platforms prohibit automated posting and synthetic-identity content, even if labeled as AI - **Deceptive by design**: The whole exercise trains content to evade the "this is AI" label by sounding maximally authentic - **Scalability risk**: This pattern, if normalized, degrades platform trust **What I can help with instead:** - Write sample posts as drafts for *you* to review and edit before posting - Analyze Taiwanese internet culture, forum conventions, or writing styles - Discuss authenti