Media Interview
summary
Shows partnerships with Cameo, Cocomelon and Candle that turn voice models into branded nostalgia experiences and claims those prototypes are now real-life, scalable products. Describes a pipeline that ingests Stats Perform feeds to produce real-time synthetic play-by-play and says, with rights clearance, a John Madden clone is feasible. Lists limits: pronunciation and lexicon errors, cross-language ambiguity, and reliance on human-in-the-loop localization. Demonstrates voice control of crowd emotion via SSML sentiment control and an 18+ language output capability.
AI-written summary · about
chapters (9)
- 0:00Introduction and CES 2023
- 1:26Voice AI in real applications
- 3:13Building teams at Baritone
- 4:35Case study with John Madden
- 6:43Sentiment control in voice AI
- 8:27Voice AI limitations today
- 10:14Multilingual voice cloning
- 11:39Dream collaborations in voice AI
- 13:03Closing remarks and resources
related talks
07mVOICE 2021 | Allan Bekker, Snap
28mWhat’s Needed in Tech to bring Voice to the Future
25mConversation with Sean King (SVP, GM at Veritone) and @ianutile
50mSpeak To Me: Voice App Conversation Practices
21mSponsor Booth Interviews w/ James Poulter (CEO Vixen Labs) and @ianutile
17mBuilding Effective Voice Agents
20mBuilding the next interface - AI personas who feel as natural as interacting with a human
45mKeynote: Conversations with Machines