2x2 网格，16:9，AI 自选不同的病毒式传播话题，从零开始创作，不要当作编辑处理：

2x2 grid, 16:9, AI picks different viral topics, create from scratch, do not treat as edit:  

class ViralEyewitnessScene:
    def __init__(self, topic):
        self.topic = topic
        self.subject_form = infer_physical_manifestation(topic)
        self.event_type = infer_plausible_public_sighting(topic)
        self.environment = infer_best_everyday_environment(topic)
        self.capture_viewpoint = infer_bystander_viewpoint(topic)

    def subject_logic(self):
        return {
            "scale": "unexpectedly large, notable, rare, or bizarre",
            "presence": "physically grounded and believable",
            "condition": "real-world wear, transport hardware, practical constraints"
        }

    def environment_logic(self):
        return {
            "setting": self.environment,
            "public_context": "ordinary daily life continues around it",
            "supporting_details": [
                "traffic", "signage", "weather", "infrastructure",
                "scale references", "bystanders", "vehicles"
            ]
        }

    def capture_logic(self):
        return {
            "camera_style": "candid, viral, observational",
            "viewpoint": self.capture_viewpoint,
            "framing": "slightly imperfect, believable, non-cinematic",
            "mood": "surprised eyewitness realism"
        }

    def render(self):
        return f"""
        Render {self.subject_form} in a surprising but plausible
        public/transport scenario, in {self.environment},
        captured from {self.capture_viewpoint},
        with documentary realism and viral-image energy.
        """