Russ d'Sa: AI hallucination is a feature in voice interfaces
Russell D'Sa · Building the World’s Smallest Fusion Reactor | Robin Langtry on TBPN · Apr 26, 2025 · at 23:48
Russ d'Sa, CEO of LiveKit, discusses why generative AI models provide a superior voice interface compared to deterministic legacy assistants like Siri and Alexa.
0:00 / 0:11exact quote · 11.4s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“People talk about hallucinations as this bad thing, but in a lot of ways, for like, kind of this speech-to-speech interaction model, where you're talking to an AI, Hallucination is a feature.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Russell D'Sa
LiveKit CEO: Hallucination is a feature in voice AI interfaces
“Well, you know, it's one of those things that I think is funny, like, people talk about hallucinations as this bad thing but in a lot of ways for, like, kind of this speech-to-speech interaction model where you're talking to an AI, Hallucination is a feature.”
Russ d'Sa: LLM inference is now faster than text-to-speech generation
“Now like LLM inference can actually be done in less time than generating speech with TTS.”
Russ d'Sa: 300 to 500 ms turn latency crosses voice AI uncanny valley
“Getting that latency end to end down, you know, to 300 milliseconds or 500 milliseconds on average for, like, turn latency that, that, that kind of helps you cross this uncanny valley for voice AI.”
LiveKit's d'Sa: LiveKit built the infrastructure for OpenAI's ChatGPT voice modes
“Fast forward, like, about a year and a half, two years, and, ah, we built a demo when ChatGPT first came out, ah, where you could talk to it instead of text with it using LiveKit's infrastructure, and a few months later, OpenAI ended up finding it, ah, wanted …”
Russ d'Sa: AI models will implicitly match user pacing within two years
“Then there's another part of it, which I think will come in the next, you know, year or two where the model will implicitly be intelligent enough to pick up on what your pacing is or your state of mind is just based on the way that you're talking and expressin…”