When AI Tech Gives the "Ick" (stevedylan.dev)

OpenAI recently launched Sora — a video-generation model wrapped in a social app whose entire feed is AI-generated content — joining other controversial products like the Friend “AI companion” pendant that have drawn public backlash. The core concern isn’t just novelty: putting algorithmically curated, synthetic media into an endless-scroll social UX amplifies the same attention-extracting dynamics that made feeds addictive. For creators and everyday users this raises a cultural and practical question: if prompting replaces hands-on creation, what counts as authorship, and what human benefits (skill-building, emotional expression, real connection) are lost? For the AI/ML community the episode spotlights several technical and design risks. Algorithmic personalization plus an AI-only content supply chain can cause homogenized, lower-quality output and feedback loops if models are later trained on synthetic content, and it shifts optimization toward engagement metrics over human flourishing. Practical responses include stronger human-in-the-loop workflows, provenance and labeling for synthetic media, metrics that reward novelty and human value rather than just clicks, and alternative social UX models (e.g., curated blog-style feeds) that foreground real people. Sora’s release is a reminder that model capabilities must be paired with responsible product design to preserve creativity, diversity, and meaningful social interaction.