VOICE 2021 | Allan Bekker, Snap
summary
A case is made that virtual beings are already commercial and technical realities: a virtual influencer earns 11 million dollars and is acquired for over 100 million dollars, Ariana Grande’s Fortnite avatar draws 12 million live viewers for a 20‑minute concert, and MyHeritage Deep Nostalgia passes over 100 million downloads. The enabling stack is listed precisely: ASR, ML‑based NLU, learned state tracking with reinforcement learning, neural NLG + TTS, and crucially a video generation model that synthesizes talking faces to complete the multimodal agent.
AI-written summary · about
chapters (5)
related talks
14mMedia Interview
27mReinventing Customer Experience with Voice and Generative AI
20mBuilding the next interface - AI personas who feel as natural as interacting with a human
22mSynthetic Media as an Extension of Identity
12mVOICE Summit 2019 | VED Talks | John Gillilan | July 23
14mAI Beyond Knowing - Explore Emotion AI with Neura
17mMy AI Clone Attended Conferences While I Flew Planes - Tanisha Sharma
11mEnabling richer experiences through multimodal interactions