It's already here even without pervasive video AI. We had text slop before LLMs too.