Kent: Hello everybody, I'm so excited to be joined by one of my closest friends Sunil Pai. Say hi Sunil. Haha,

Sunil: Hi Sunil

Kent: Haha!

Kent: That's perfect. That's perfect. It's so good to see you So I just want everybody to get to know Sunil because he's just a fabulous person. So why don't you give us an introduction to yourself? What are you into? What are you doing?

Sunil: Hi, my name is Sunil Pai. I'm the creator of something called PartyKit, which is a platform and libraries for building real-time applications. So web sockets, local first apps, things that push bytes at you instead of you having to pull stuff. People build games with it, they build real-time UIs, they build AI agent apps, they build a bunch of like fun stuff. And that's what I've been focusing on for the last year.

Kent: Awesome, why don't you tell us a little bit about your history too, like where were you before, what are the things that gave you the experience so that you could launch something so cool as PartyKit? We'll talk more about PartyKit too.

Sunil: Sure. Before PartyKit, I spent a year building developer tooling and the CLI at Cloudflare, which is a very cool company which has very cool tech. And I helped build their CLI for them, which is how I learned how cool the tech is and that I wanted to build a company on top of it. Previous to that, I spent a year in the React.js core team. I'm the one person nobody ever talks about because... Well, they just don't because I didn't do much during my time

Kent: They

Sunil: there.

Kent: talk about one of the things that you did, mostly cursing it.

Sunil: That's right, there's a testing API that pisses people off. That's right.

Kent: Ha ha ha!

Sunil: I spent some time in Oculus building massively multiplayer real-time. React apps, that was easily the most sci-fi job of my life. I've done a

Kent: Uh

Sunil: bunch

Kent: huh.

Sunil: of other things. I've worked in e-commerce, I've worked in government. I built a CSS and JS library. A bunch of people liked it, a bunch of people hated me for it, so that was also a fun time. And that's all led me to building what I think is part of the next phase of interesting applications on the web, and that's why I'm focused on PartyKit now.

Kent: Yeah, fantastic. So I use PartyKit and I love it. So anybody who has gone through Epic Web workshops, you'll see the little face pile in the corner when you've got the navigation open. And that is all powered by PartyKit. Couldn't be easier, honestly. So, so good. And PartyKit is more than just real time as well. It's like... We can server render that face pile because we can make just a regular old request to the PartyKit server. So it's a very powerful, flexible system that you've put together and lots of exciting things too. One thing that I'm excited to get my hands into is the AI related stuff that Cloudflare has offered and PartyKit exposes as well. Do you want to talk a little bit about that?

Sunil: Yeah, it turns out that when you put these magical big matrices of numbers on the edge and you run questions through them, you can start talking to robots on the edge, which is very cool. In general, you can build like very cool chat GPT type experiences. But the moment you connect it to something like PartyKit, you can build a lot more interesting stuff because the AI can talk to you without you talking to it first. You can build AI games, things that proactively react to things that you do. You can have it just listening to the things that you're doing while writing a document and occasionally pop up and maybe do a fact check or a tone check for you. Maybe it can rewrite some stuff while you're writing it. Very interesting to see AI in a real-time use case. So, I am quite excited with that. Then you add stuff like audio and video in real time with PartyKit and you can build truly novel experiences. And what I really like about PartyKit is not that it hides complexity, but it's like actually truly a novel primitive for building things. Like we have junior developers who build like crazy real time games and they're like, wow, that didn't take a lot of time after all. In fact, I remember the first version of the face pile for Epic Web. We did over a weekend or like a couple of phone calls which is like, hey, like we should just build this. Here's the thing. Here's the thing. And then you took it and ran with it despite never having written any PartyKit code before. I think that's like really special. There's something special about PartyKit, which is why I'm biased, of course. But yeah, if you add it, it becomes interesting once you start adding streaming services like LLMs and databases and other human beings and you see what happens in the middle. So I'm quite excited to explore that.

Kent: Yeah, yeah. I've heard it a lot that it's best to treat an AI as just another client that is connecting to your application. And so real time being a very obvious way to think about that. And you're just doing multiplayer. It just so happens that one of the players is a robot. And you can do some pretty interesting things with that. going way beyond just the chatbot stuff. What I'm really looking forward to is people applying AI to more than just like, oh, here's a chat for our docs, but like, I don't know. I feel like there's just more that we can do. ChatGPT has got chat. Like, let's just let them do the chat thing, that's fine. I feel like we can do more with stuff. And so eventually I do expect that I will dabble more. in AI and PartyKit is one of the top things on my list for ways to explore that in a practical setting. So.

Sunil: I would love that. We are going to be spending a lot of time now figuring out what the experience that I'm doing exactly what you say is. And I'm looking forward to seeing what people build with it.

Kent: Yeah, that's awesome. And like doing that from the edge, like we could go on, like it's really cool stuff.

Sunil: This

Kent: So,

Sunil: would be a whole podcast if we tried.

Kent: yeah, exactly. So, Sunil, what are you planning on talking about at the Epic WebConf?

Sunil: It'll be in the same area. Of course, I'll talk a little bit about PartyKit, but really I want to impress. Building real-time apps over the last 15 years have been the purview of only multi-billion dollar companies because the infrastructure has otherwise been super expensive. You needed distributed systems, experts, and people to maintain it. The physics of it have changed and the requirements of how people, what people want from applications have changed. Whether you use PartyKit or another provider, I think what I want to impress on people is how simple it is to get started, the wild new world of new types of applications that you can really build, and really how you can make the web a little more epic if you try. So I'm excited to find out.

Kent: Yeah, awesome, I'm looking forward to that as well. It's gonna be a lightning talk, so punchy, very fast. Keep your eyes open, don't blink. Good.

Sunil: Honestly,

Kent: Feel free

Sunil: it's

Kent: to.

Sunil: good so I don't have to spend 10 minutes talking about my life story. I'll just say, okay fine, here we go. Like, let's just look at it.

Kent: Perfect, yeah. Feel free to take pictures, but post them later, because if you try to post them while he's talking, it'll be over, so it is gonna be very good. I'm looking forward to that. I always enjoy your talks and your demos and the crazy mad scientist stuff that you're doing just to inspire us to do cool things. So I'm excited to have you on stage there.

Sunil: Another thing I'm looking forward to is I love Salt Lake City. I've been there twice, but this time I want to find someone to feed me funeral potatoes, which I hear are great. That's

Kent: Oh,

Sunil: what I want

Kent: okay.

Sunil: to eat.

Kent: Yeah, yeah, we'll get you, I'll have to think about that. We can maybe do that for dinner. We'll do funeral potatoes and jello and like just the regular Utah cuisine. Yeah, all right, I'll keep that in

Sunil: That's

Kent: mind.

Sunil: what I want. Okay.

Kent: Great, very good. Cool, so while we're at the conference, we're gonna have a bunch of people, it's a single track and so everybody's in all the talks and everything and then we take extended breaks. That's a big part of. this conference is just to give people opportunities to rub shoulders with each other and get to know each other and stuff. What are the sorts of things that you're excited to talk with people about in that hallway

Sunil: What I really like about something like the Epic Stack and the Epic Web community coming together is that it's a mixture of people from across the industry, across industries and domains, as well as different levels of experience from junior to very senior. I mean, we all learn from that. And those, I've learned that you get extremely high quality conversations by actually doing that. You'd think that it would be the opposite, but really you learn from a breadth of people and seeing how it applies to a common context. So really, I

Kent: Hmm.

Sunil: think I just want to see what everyone is learning, what they've learned, what kind of things they're building. That's always my favorite part of conversations at these conferences. And mostly I'm looking forward to meeting my friends again. It's been a while. That's what I'm looking forward to.

Kent: Yeah, well, life is better with friends, right? So

Sunil: Everything

Kent: that is

Sunil: is

Kent: good.

Sunil: better with friends.

Kent: Yes, everything is. Well, awesome, I'm excited to see you in Park City, Utah, April 11th for Epic WebConf. And I hope that the folks who are watching now are excited to meet you as well. So yeah, thank you so much for giving us some of your time today, Sunil, and we'll see you in April.

Sunil: looking forward to it. Thanks Kent.

Kent: Okay, bye everybody.