https://reddit.com/link/1ue5o4i/video/ozv6sohrc69h1/player Instead of only texting AI characters, the app shows characters that can talk through voice, lip sync, react with facial expressions, and respond to camera context during the conversation. The demo suggests a shift from text-based character AI toward video-native AI characters, where the interaction feels closer to a live call than a chatbot. For ML developers, the interesting part is the underlying stack: vision, speech, memory, ...