S15E1: Building the Infrastructure for Voice AI with Russ d'Sa, Co-Founder and CEO of LiveKit

The Room Podcast

Russ d'Sa, CEO of LiveKit, discusses the company's evolution from an open-source video project to the 'accidental AI company' powering OpenA

Key takeaways

LiveKit became an 'accendant AI company' after OpenAI utilized its cloud to build ChatGPT's voice interface.
The transition from text-based chat to voice requires a shift from discrete request-response models to continuous, low-latency computing.

Main topics

The evolution of LiveKit
OpenAI's use of LiveKit Cloud

Notable quotes

If OpenAI is going to build a computer that is increasingly becoming like a human, you're probably going to interact with that computer in human-like ways.

Conclusion

The episode provides a profound look at how the plumbing of the intern

Transcript preview

Speaker 1 (0:00) It turns out that if OpenAI is going to build a computer that is increasingly becoming like a human, Speaker 2 (0:06) you're Speaker 1 (0:07) probably going to interact with that computer in human-like ways. Speaker 2 (0:18) Welcome back to a new season of The Room Podcast. Hi, Speaker 3 (0:20) Mads. Hi, Claude. As you all know, we are opening the door to the room where it happens for today's technology founders. I mean, I Speaker 2 (0:27) can't believe this is season 15. Crazy. Probably one of our most exciting seasons telling stories up and down the stack. After six Speaker 3 (0:34) years in The Room, we're just delighted to be back together in San Francisco more and more, both being based here live from Vercel HQ. I think you have an update as well. Speaker 2 (0:43) Yes, I'm also now at OpenAI, which is fun, pulling us deeper into this revolution. Thank you, everyone who's been Speaker 3 (0:49) listening to us since the beginning six long years ago. And for those who are new here, please like, follow, subscribe wherever you listen to podcasts or, of course, at theroompodcast.com. Today on The Room Podcast, we sat down with Russ Dessau, the co-founder and CEO of LiveKit. LifeKit is the company that built the open source infrastructure layer powering real-time voice and video for AI applications, becoming the go-to plumbing behind the next generation of apps where AI agents actually talk and listen. And Speaker 2 (1:17) now a message from one of our sponsors, Ripley. One thing we constantly hear on the room is that great operators don't just need more data. They need a faster way to turn that data into decisions. As a founder, I know how painful it is to wait days or even weeks for a report. while a real business problem is unfolding. That's what makes Rippling AI different. You can ask Rippling AI, is our sales team at risk of missing quota? And Rippling AI instantly pulls real-time data from Rippling and your CRM into presentation-ready dashboard data showing quota attainment, headcount trajectory, and monthly revenue to quota by region. But it doesn't stop at the insight. Rippling AI can act on what it sees. recommending and building an automated recruiting workflow when open roles and slow time to fill are putting quota at risk, all while following your existing permissions and approval processes. Question answered, action taken, problem solved. Ready to rule your business? Head to rippling.ai slash the room and sign up for exclusive access. And Speaker 3 (2:14) now a word from our new sponsor, Gradial. Whether you're building a company, launching a product, or growing a brand, you know the idea is just the beginning. And marketing's bottleneck has never been creativity. It was always the work it takes to get an idea out the door. Gradial handles the operation so that marketers can stay on craft. An agent will learn your process, work across the systems you already use, and delete point solutions you don't need. Designers will ship on-brand pages without counting prompts or tokens. Product marketers will launch across markets in days, not weeks. And digital experience teams test customer journeys, catch what's not working, and fix it. all before launch. Gradial learns that you're a brand, business, and process and brings that continuous context to every workflow. Every task gets done your way with brand, quality, and approvals built in. No more token maxing, slop, or runaway AI bills. More time to think bigger and create the impact people remember. Gradial. the system of work for Speaker 2 (3:11) marketing. Ashurst Perkins Speaker 3 (3:12) QE is a leading full-service international law firm known for providing high-value strategic solutions for startups and investors. Lawyers at Ashurst Perkins QE support the most innovative entrepreneurs and investors in fast-moving and high-growth sectors across the country. while addressing the myriad of legal needs that arise when fundraising and scaling their businesses. The firm's tailored legal services are informed by understanding of each of the client's discreet needs. To learn more about Ashurst Perkins Coie, visit their website, ashurstperkinscoie.com. Speaker 3 (3:48) First of all, Russ, thank you so much for joining us today in the room. We're delighted to have you. It's great to be here. We'd love to start with the beginning, and we do this with all of our guests, just to kind of hear more about where you grew up and then, of course, how that shaped your view of the world. Speaker 1 (4:02) Yeah, so I grew up, I'm Canadian. So I'm nice. No accent. No accent. That's right. It got hammered out of me when I came to the States and everyone made fun of me for saying after every sentence. But I grew up in Canada in a small town outside of Ottawa. And my dad was my dad was in tech. So I he was an electrical engineer in England. Eventually, him and my mom made their way to to Canada where I was born in Toronto. And then after about a year. My dad moved to the family. We moved to Ottawa for a job he had. And so I grew up there and surrounded by computers. My dad has like a pretty crazy story. Went from like semiconductor industry, GPUs, and then DSL and high speed networks, and then into flash storage. So he's kind of like worked across all of these foundational kind of like hardware revolutions that have happened over the 80s and 90s. And then so I was growing up, you know, around him and he's like one of my best friends and my hero. And so just surrounded by computers, my dad eventually got a job out in Silicon Valley in the early 90s. And he went out himself to Cupertino. Wow. And so he was in Cupertino flying back to Ottawa every month for about a week. And after a couple years of doing that, he decided, hey, you know, let's move the whole family out. And so we all came out here in the mid-90s. Speaker 1 (5:36) And then, you know, I've been kind of growing up in Silicon Valley ever since. And my dad always took me to things. I think I was the youngest person at CES. Was it Speaker 3 (5:43) CES? Speaker 1 (5:45) I think it was CES. The youngest person at CES back when I was like 14, 15 years old. And, yeah, I think like I've always been fascinated with like programming and wanting to make video games. I think it was a classic origin story for a lot of programmers in the 90s, growing up in the 90s. And so, yeah, I've been steeping in this stuff. I kind of feel like I was always. going to do this no matter what. Yeah. That Speaker 3 (6:07) resonates so much. My family moved from Atlanta, Georgia to Seattle when I was four. And Speaker 1 (6:11) my Speaker 3 (6:12) dad also worked in tech. And just the fact that as a 10 year old, I knew what the cloud was and kind of that concept of like, you know, Seattle being the cloud capital of the world with AWS. Was your dad at Amazon Speaker 1 (6:22) or Microsoft? He Speaker 3 (6:23) was at Madrona, which was an early investor. Oh, I know Madrona. Yes. And that to me was just like the, what every dinner table was like talking about. Great companies. Crazy. Speaker 1 (6:32) Madrona passed on our Series A, so that sucks for them. Oh, Speaker 3 (6:35) yeah, that does definitely suck for them because we're going to get to Speaker 1 (6:37) some big numbers in a second. No, Speaker 3 (6:38) no, no. Speaker 1 (6:41) Well, Speaker 3 (6:42) I'm curious for you then, having that steep and knowing obviously so many people around the corner in the industry, did you yourself always think you were going to become a founder? Speaker 1 (6:51) Definitely. I think not necessarily a founder in the sense of how it's manifested now, but... I think I always wanted to start my own company, but mainly around things like video games. So when I was, it's funny because my dad was in startups and all of that technology. But where the real kind of motivation came from was when we were flying to California to visit him for the first time after he had moved. I watched this. Speaker 1 (7:26) documentary on the airplane. This was like, you know, in like 92. I Speaker 3 (7:30) was going to Speaker 1 (7:30) say like Speaker 3 (7:30) on what, the TV that they had on top of the whole. They had a screen, Speaker 1 (7:34) but it was like the, I don't Speaker 3 (7:34) know. The Sherrod screen? No, it Speaker 1 (7:35) definitely wasn't the Sherrod screen because nobody else would have wanted to watch this thing. It was a documentary about John Carmack and John Romero. Nice. So John Carmack, I think more famously now for Keen Technologies, which is the AI company, and then Oculus, where he was the CTO. He was also one of the founders with John Romero of id Software. Oh, interesting. And id Software, you know, they made Doom as one of their most famous games and Wolfenstein as well. But, like, I played those games, you know, growing up. And so there was this documentary about the two of them. And they tell this story about how they were working at this company. And then in the evenings, you know, they wanted to work on their own stuff together. So in the evenings, they would unplug. Speaker 1 (8:21) the computers, put them in the back of their pickup truck, drive the computers home, work all night on their own stuff, and then take the computers back in the morning before anyone showed up and then plug them back in. And that wouldn't fly, you know, IP rights and all these days. But I just thought it was kind of amazing how these two guys, nobody's ever heard of them, could like type words on a screen and influence the world, you know, hopefully in a positive way. just from their garage or from their house. And so I was really enamored with that idea, that ability to have this outsized impact as an individual. And so it was kind of from that moment that I decided, okay, I want to be a founder one day. I want Speaker 3 (9:04) to Speaker 1 (9:04) create something myself. Speaker 3 (9:05) That identity was steeping in you. Yes, Speaker 1 (9:07) yeah, for a long time. Speaker 3 (9:08) And have you read the book tomorrow, tomorrow, tomorrow? I Speaker 1 (9:11) haven't. Speaker 3 (9:11) Okay, it's worth a read. It's kind of a normie gaming book, though, I will say. But they're about to make it into a movie. Oh, yeah. And just kind of what you're describing of like kind of how you fell in love with the whole ecosystem. I have to know what the first company was that you got. into YC with. Was it a gaming company? Speaker 1 (9:25) No, actually, it's more embarrassing than that. So back then, there was this moment where webcams were being built into these white plastic MacBooks. Speaker 3 (9:38) So all Speaker 1 (9:39) these students, when I went to college at Davis, a little bit up north from here, and there were people that were buying these white plastic, you know, kind of laptops. They had a camera that got built in, and it was the first time that a webcam had been integrated into a portable computer that you could take with you anywhere. And so I thought to myself, I'm like, oh, wow, you could broadcast yourself with this camera here anywhere you are. Initially, the idea was, well, why not? live stream yourself everywhere. And it was right, maybe like a couple months after that, Justin.TV came out. Oh, interesting. Which is so funny because Justin Khan is one of the first checks into LiveKit. Nice. So it's just like this crazy, weird full circle. But Justin.TV had just come out, and I'm like, okay, well, can't work on that idea. You know, and it's a YC company too, so can't work on that. Maybe... And instead, we'll work on like something where you can meet anybody on the Internet live, right, and chat with them. So, yes, it sounds like chat roulette. But it is chat roulette in a lot of ways. In a lot of ways, chat roulette many years before chat roulette. But the angle of it was around speed dating. So we needed something that people could kind of like. see as like a purpose or at least we felt like chat really didn't really have a purpose but we felt there needed to be a purpose for people to even do this thing so we're like okay well how about dating because that there's this analogy in the physical world of people doing that maybe people will do it over the internet instead. Yeah, Speaker 3 (11:12) and Speaker 1 (11:12) what year was this? This was 2007. I Speaker 3 (11:15) feel like that was a novel time to be thinking about how people might do this. Oh, very early. Yeah, nobody Speaker 1 (11:19) was, literally nobody was doing, except for Justin.TV, nobody else was doing video at scale in a browser. Right, I mean, it was almost impossible Speaker 3 (11:26) to even do it. Very difficult. From an Speaker 1 (11:28) infrastructure perspective, extremely difficult. I mean, when I think about what we do now, and I think about what the tools were like back then, yeah, it was... Yeah, dire straits up there. Speaker 3 (11:39) Well, it didn't take off to the levels that you would have hoped with that first company. It didn't. And Speaker 1 (11:44) ultimately Speaker 3 (11:44) you joined Twitter, correct? Yes. Speaker 1 (11:46) Tell us a little Speaker 3 (11:46) about that. Speaker 1 (11:47) No, actually I joined 23andMe first. What? Speaker 2 (11:49) Yeah, yeah, yeah. Okay. Speaker 1 (11:50) Well, tell us about that transition then. Well, before I did YC, I had gotten a job offer. So this random company of five people called 23andMe had sent me a recruiting email. And so I had happened to read about 23andMe on, there was this blog, like a tabloid blog called Valleywag. Oh, yeah. That's old. Old school. And Valleywag was writing about how the founder of Google's wife, or they weren't married at that time, but girlfriend is like starting this company, this genetics company. And so I had heard about it and then I got this email. So I'm like, oh, okay, it's interesting. I at least read about it. So I went and did an interview. loop with them and I got offered a job. And then I also got into YC like maybe the next week. And so, you know, this was my moment to kind of finally make my dreams come true, start a company, follow in my dad's footsteps. And so I ended up going and choosing YC instead. And then after, you know, about a year or so, my YC company didn't work out. I went crawling back to 23andMe and I think they were about a 30 person company and I emailed them and I said, hey, are you still looking for folks? And they were like, yeah, yeah, yeah, come do an interview. And so I went and I interviewed again and got the job. And that was my first job after my Speaker 3 (13:10) first Speaker 1 (13:11) company. Oh, wow. Speaker 3 (13:12) Oh, my gosh. Well, I mean, really different overlapping ecosystems, right? Like obviously everything at the forefront of what's happening on the Internet, but completely different verticals. Yes. How did that all come about? And then ultimately, yeah, like finding yourself in the Twitter ecosystem. I Speaker 1 (13:26) think that I was drawn to 23andMe because they, you know, I have on my LinkedIn, I talk about how I like, I always work on these ideas before, too ahead of their time. And I was attracted to 23andMe because I had actually personally never taken a biology or genetics course ever. They somehow fooled me into taking this integrated science thing in high school and not taking biology, which I heard here they're still doing in schools. So I had never taken biology, and I have long thought that eventually technology will go from kind of its own category into something that permeates across every discipline. Speaker 1 (14:12) and where every discipline can get leverage from technology in some way. And so I think this cross-disciplinary combination of like genetics and technology software was really interesting and compelling.