oben
co-founder · personal ai · founded in 2014
i co-founded oben in 2014 around the idea that every person could have a personal ai: a digital version of themselves with their own voice and appearance.
we were backed by investors including softbank ventures korea and tencent.
we were doing early work on voice cloning, digital avatars and conversational ai, well before the current generative-ai wave.
by 2016, our technology could create a 3d likeness from a selfie and learn a person’s voice from a short recording. it could then generate new speech in that voice, rather than simply play back the recording.
we were also developing the conversational side. by 2018, our in-house work combined computer vision, speech and natural-language processing, including chatbot capabilities for personal avatars. the aim was a digital identity people could interact with, not just an animated face.
these are selected records of that work, using the ai tools available at the time.
selected public records
- voice cloning and digital avatars — nvidia, november 2016
- talking and singing avatars — htc vive, april 2017
- s.m. entertainment collaboration — june 2017
patent — co-inventor
creation and application of audio avatars from human voices
US 9,324,318 B1 · granted april 26, 2016
co-invented with julian bunn (caltech) and yi zheng. voice-transformation technology that preserves spoken content while changing the voice.