Live Comments

Meta is starting on an ambitious project to integrate its live streaming services across Instagram, Facebook, Facebook Gaming, and WhatsApp. The goal is to enable simultaneous video viewing across these platforms, while also offering a cohesive experience for live comments.

This unified commenting system requires the capability for comments to be persistent (i.e., they can be reviewed later), while still supporting interactive features such as reactions (likes, hearts, cares, etc.) from users across all platforms. Additionally, those reaction counts must be updated and displayed in real time to all viewers.

How would you approach designing a system capable of processing and displaying real-time data across multiple platforms? As a final consideration for your design, suppose that an AI censorship feature is added. How would you handle the latency? Would you recommend for the censorship to be static (e.g., through text matching) or dynamic (NLP)?

Note: You do not have to develop the censorship feature, but rather, recommend an approach.