From Rubik's Cubes to AI Agents: Building the Future Through Play

Transcript

[00:00] Sarah Enix: Hello. My name is Sarah Stoker. I write a newsletter called If You Know, You Know, and my goal is to make learning about AI feel more like a coffee chat rather than reading a textbook. I have the joy and privilege of speaking with my good friend, Parth, today. Parth, could you introduce yourself and tell us why we're talking?
[00:30] Parth Patil: Yeah. Um, I'm Parth. I'm based in Los Angeles. I'm like, uh, obsessed with technology, especially language models. And, uh, most of the time, I bas- I spend my days working on like chatbots and, and natural language automation, AI agents, stuff like that, just trying to tinker with these tools that have come out since ChatGPT, you know. If you can talk to the computer, what can, what can the computer do for you? Which is kind of like where my head's at.
[00:57] Sarah Enix: Mm-hmm.
[00:58] Parth Patil: Um, and, uh, yeah, I'm excited to be here and kind of just demystify some of like the more recent changes that are, you know, happening in the AI ecosystem and like technologies that are coming out, and hopefully share some of my weird sci-fi vision for what's, what's coming next. Yeah.
[01:15] Sarah Enix: Oh, man. Just to give you some context, Parth and I met last year and, uh, had a night that involved a lot of Chick-fil-A, and I knew we were gonna be quick friends. So, uh, I could listen to what he says. We sat around a fireplace over at Todd's and just heard all about his thoughts and theories, and I was like, "Okay, I need to actually record this, and like the world needs to hear what you're saying." So, thanks for listening.
[01:40] Parth Patil: I don't, I don't often, I don't often do this, so I'm glad you're here.
[01:42] Sarah Enix: I know. I know. I had to... Just so you know, I've had to text Parth probably six times to make this happen.
[01:49] Parth Patil: (laughs)
[01:49] Sarah Enix: So I am like just thrilled you showed up on time, ready to go.
[01:53] Parth Patil: (laughs)
[01:53] Sarah Enix: You got your microphone. Um, but let's, let's zoom back. Z- take us up a couple steps back. Tell me about you as, in the third grade. What the kind, what kind of things were you doing? Who was your friend group? What were your parents like? Tell me who you were, Parth, as like an eight-year-old.
[02:10] Parth Patil: Eight-year-old Parth. Um, okay. Yeah, this is, this, this will be good. I, when I was, when I w- Like I grew, growing up, I was kind of like a, I was like a, I was your classic nerd. I read more books than anyone. I was like obsessed with dinosaur, you know, when I was eight, yeah, I definitely thought that I would be a paleontologist when I grew up. And I was just like, "Oh, man, one day I'm gonna find a T-rex in the ground, and then I'm gonna make that my like living room kind of like... It's gonna be, you're gonna come into my house and there's gonna be a T-rex in the living room." And I, that's what I genuinely was like, "I'm gonna be a paleontologist." I know my par-
[02:42] Sarah Enix: But aren't you now? You're kind of time traveling now in a way.
[02:45] Parth Patil: Yeah. I mean, I think, I think, I think the cool thing about AI is that it allows you to reach into a lot of different topics. Um, so like, the, the stuff that I thought that would require multiple lifetimes, now because of AI, you can kind of explore things and become more mol, like explore your multifaceted interests. So, I think, you know, at the time, I was like obsessed with dinosaurs. So I'd go to the library and I'd read every single book I could on dinosaurs. And then I'd go to the librarian and I'd be like, "You need to get more books. I've already read all of them."
[03:13] Sarah Enix: What's your favorite dinosaur?
[03:15] Parth Patil: It's T-rex.
[03:16] Sarah Enix: Nice.
[03:17] Parth Patil: It's kind of, it's kind of basic, but I just think it's kind of an, I mean, it's cool. It's cool. Like, I haven't really changed in that sense. I think, as a kid, a lot of what I liked as a kid, I just like more now. So like, my dad got me Super Mario Brothers when I was six, and then he got me like, like then I s- I played, he, I played like RollerCoaster Tycoon. So I played a lot of like, yeah-
[03:38] Sarah Enix: Yes. Zoo Tycoon.
[03:40] Parth Patil: Yup, yup.
[03:40] Sarah Enix: Sims.
[03:40] Parth Patil: Zoo Tycoon.
[03:42] Sarah Enix: Sims 2.
[03:42] Parth Patil: Yup.
[03:42] Sarah Enix: Sims 4.
[03:43] Parth Patil: All of that's actually incredibly relevant 'cause it, it shaped my view of like what technology was capable of early on. I think, I think when I got my first job, I was like, I was looking around and I was like, "Wow." I, I worked for like an older company, and they still had paper, and so I was looking around the office to be like, "Wow, like peop- a lot of people here move paper from one part of the office to the other, and then they put it into some folder." And then I'd go home and I'd play RollerCoaster Tycoon and everything would be digitized. So, you have like a digital finance operation 'cause it's all just like numbers in a computer. And I, I felt that that was like, "Oh, wow. So like video games are in some ways more advanced than reality." Like, the, the units in a video game are listening to my instructions, helping me run this imaginary business, and I always thought that was like, uh, that inspired me 'cause it was like, "Okay, cool, like if you start digitizing things, you can, you can do these like, you can, you can arrange some, you can arrange information in a way that's more intuitive to like children, right?"
[04:42] Sarah Enix: Yeah.
[04:42] Parth Patil: So like, like why is a, like when you play a, a game like RollerCoaster Tycoon, you start learning about concepts like how to run a business, how to design a profitable theme park and like the underlying-
[04:55] Sarah Enix: How to make sure the hotdog line doesn't like get too long, so build like two of them. (laughs)
[04:58] Parth Patil: Yeah.
[04:58] Sarah Enix: Yeah. (laughs)
[04:58] Parth Patil: Optimizing the, like, route around your theme park, making sure people are having fun, like, and then like is your business like working, or does your roller coaster actually get people to like... So, all those things were getting me thinking about like, um, systems and thinking about optimization. So like, I'm a very like optimization kinda guy. Like I always think about, "Oh, well, why can't we just do it faster?"
[05:20] Sarah Enix: Mm-hmm.
[05:20] Parth Patil: Or like, can we do it like, can we go deeper? Can we go further? Everything, I think, is like when you digi- when you move it into the computer, now the computer can do many things that a human can't because we're like lazy and honestly, like, I mean, tr-
[05:33] Sarah Enix: Is that the reason you think? 'Cause we're lazy?
[05:36] Parth Patil: Well, no, we designed the computer to be, um, productive in ways that human beings are not naturally.
[05:42] Sarah Enix: Yeah.
[05:42] Parth Patil: Like if you were to ask me to come up with like a, a s- like a hundred strategies, I'd be like, "Go home. Like, I don't want to." But, you know, go to ChatGPT, come up with 100 strategies, like-
[05:53] Sarah Enix: Let's go.
[05:53] Parth Patil: ... it's not gonna, you know, it'll, it'll try. It'll try.
[05:57] Sarah Enix: Yeah.
[05:57] Parth Patil: Um, so yeah, played a lot of video games, ate a lot of chocolate and junk food.
[06:02] Sarah Enix: Mm-hmm.
[06:02] Parth Patil: Read a lot of books and, uh, a lot of board games. I play, I, I love board games, so like-... in the beginning, I was playing a lot of Monopoly and Risk, and I got like ... I realized like, I, you can win every game if you learn the strategy and everyone else is just playing casually. Like, in Monopoly, you just buy every property. In Risk, it's like-
[06:22] Sarah Enix: You just take over the world.
[06:24] Parth Patil: Yeah. But, but as I played those games, I realized like, those games aren't necessarily fun for everyone. Actually, th- the moment someone gets ahead in Monopoly or in Risk, the game becomes miserable for everyone else.
[06:35] Sarah Enix: Yeah.
[06:35] Parth Patil: So, that's when I was like, okay, so there's, there's like fun games, and then there's games that are kind of just like, you know, two hours in, people just wanna flip the board over.
[06:43] Sarah Enix: What's your thoughts on Settlers on Ka- uh, Catan?
[06:45] Parth Patil: I think, I think Settlers is one of my favorite games ever. Um-
[06:49] Sarah Enix: It feels like a little more spread out, you know?
[06:52] Parth Patil: Yeah. A lot of people could ... Uh, I think the cool thing ... I think like one way to make a game fun is like give more people a chance to feel like they're in it towards the end. So like, at the end of a Settlers game, there's like three rounds left. It, it's like there's a couple people vying for that, that victory spot.
[07:06] Sarah Enix: Mm-hmm.
[07:06] Parth Patil: And everything can change in one or two turns if like someone pulls an alliance off. So, I think Settlers is a good game. It's like a very ... Like, it, it's got trading, it's got like subjective value, it's a little random, but then everyone's got a different strategy. And, and I think, I think that's awesome. So yeah, my brother and I, I have a younger brother. He's, he's six years younger than me.
[07:26] Sarah Enix: Aw.
[07:26] Parth Patil: And so like, and a lot of what we do, like a lot of our interest is the same.
[07:30] Sarah Enix: Mm-hmm.
[07:30] Parth Patil: So like, he's always putting me onto video games or like, I'll come back and he'll be like, "Oh, check out this board game we got." Um-
[07:35] Sarah Enix: Have you played Monopoly Go? I mean, I'm sure we're gonna get into AI soon, I promise.
[07:38] Parth Patil: Yeah.
[07:38] Sarah Enix: But have you heard of Monopoly Go?
[07:40] Parth Patil: I've heard of it. I've heard it's like, uh, it, it's ... I've s- I've, I've seen videos on TikTok, but I haven't played yet.
[07:45] Sarah Enix: It's very fun. It's like-
[07:46] Parth Patil: Yeah.
[07:47] Sarah Enix: ... super quick monopoly. You can honestly play it when you're camping. It's, it's a blast.
[07:51] Parth Patil: Yeah. Yeah.
[07:53] Sarah Enix: Yeah. Okay, and then, whereabouts were you growing up, and then when did you move to LA?
[07:58] Parth Patil: Yeah. So I grew up in the Bay Area, so San Jose, Santa Clara.
[08:02] Sarah Enix: Nice.
[08:02] Parth Patil: So kind of, I, I like to think of it as like the real heart of Silicon Valley where actual silicon, i- like, that like-
[08:10] Sarah Enix: Yeah.
[08:10] Parth Patil: If you think about Silicon Valley in San Francisco, I think that's where a lot of software projects are. But the South Bay is where like the chip, chip companies are. Nvidia, AMD. So growing up, like both my parents worked in chip design.
[08:21] Sarah Enix: Mm-hmm.
[08:21] Parth Patil: My mom works on like, uh, bio chips and like-
[08:24] Sarah Enix: Mm-hmm.
[08:24] Parth Patil: ... putting the human genome on a chip, that kind of tech. And then my dad works on chip design software, so like, the tools that teams like Nvidia use to design their next generation. So like, or Apple. When anyone wants to build a chip-
[08:37] Sarah Enix: Well, Meta just released their chip too, I think.
[08:38] Parth Patil: Yeah. Yeah. So like, when- whenever these guys have a new chip, they have to design it in the year, in a couple years before that, that the chip u- even comes out. They're in the design phases.
[08:47] Sarah Enix: Mm-hmm.
[08:47] Parth Patil: And so my dad works on, his company works on the software that is used to design, um, chips. Which, which has been cool to see, like he's worked, they've both worked at those companies like my entire life.
[08:58] Sarah Enix: Mm-hmm.
[08:58] Parth Patil: And so I've seen those industries go from like ... I mean, they were always important, but they've only ever gotten more and more important.
[09:07] Sarah Enix: Mm-hmm.
[09:07] Parth Patil: Like the demand for chips kept going up. In the beginning, like my, like I, my dad was like, oh, w- working with Nvidia. And, you know, when I was like 15, I asked my dad for a PlayStation. And he was like, "No." And then I was like, "Well, what if you let me build a computer?" Which is more expensive than a PlayStation. But my dad was like, "Yes," because that's, that would be a learning experience. So, so he, he, yeah, I went to Fry's Electronics. I like made a list of all the parts, and then we went and bought them and I built my own computer.
[09:37] Sarah Enix: How old were you when this happened?
[09:39] Parth Patil: And I think I was like, this is middle school, maybe.
[09:41] Sarah Enix: Wow.
[09:41] Parth Patil: Like eighth grade. Um, and my brother, you know, he was, he was there with me. So like anything that I did, my brother did like six, seven years early.
[09:50] Sarah Enix: Mm-hmm.
[09:50] Parth Patil: So like, like I got obsessed with the Rubik's Cube in middle school, and I taught my brother when he was like six or seven. And-
[10:00] Sarah Enix: Wow.
[10:00] Parth Patil: ... it was pretty important to me that he understood how to do it that early because I was like, you know, you solve the Rubik's Cube, and I think if, if anyone watching this is like a Rubik's Cube nerd, you'll probably understand. But you solve the Rubik's Cube and it's a really empowering moment 'cause you realize everyone thinks that you have some magical talent in your hands. And then-
[10:18] Sarah Enix: That's what I'm thinking like literally right now. (laughs) I'm like, "How did you do this?"
[10:20] Parth Patil: Yeah. Yeah. But, but every cuber, everyone who finishes solving the cube for the first time, they're like, "Wait."
[10:25] Sarah Enix: So cool.
[10:25] Parth Patil: "No. I just, I just studied it and learned the technique, and then now I can do this thing. But everyone's gonna overestimate me from this point on, on that."
[10:32] Sarah Enix: Mm-hmm.
[10:33] Parth Patil: They'll be like... So, so my friends and I, we were obsessed with this 'cause we were like, we just had too much time on our hands. Then I got my brother into it. And so then-
[10:40] Sarah Enix: Pun intended?
[10:41] Parth Patil: Yeah.
[10:42] Sarah Enix: Right. (laughs)
[10:42] Parth Patil: Yeah, basically like ... And we're out here, we're out here like trying to figure out how to solve it faster and how to solve it like, you know, solve bigger cubes. And then I, I wanted to stay ahead of all my friends. So then I was like, "Well, I need to do something that's actually hard-"
[10:55] Sarah Enix: Mm-hmm.
[10:55] Parth Patil: "... that'll get them to stop chasing me."
[10:57] Sarah Enix: Mm-hmm.
[10:57] Parth Patil: And so then I was like, "I'll solve it blindfolded." So I figured out how to solve it blindfolded. And you can just go to YouTube and learn 'cause there's other kids that have really good ... Actually, the best cubers in the world are all kids.
[11:08] Sarah Enix: Hmm.
[11:08] Parth Patil: Uh, or they were kids when they set those records. Because you need time and you need to have like, you need to think that it's possible. I think a lot of t- I think that's what the cube teaches you when you ... I think that's why I teach my friends the cube, is because I hope that they walk away thinking, "Oh, maybe people are overestimating this problem, because I know how to do this."
[11:26] Sarah Enix: Mm-hmm.
[11:26] Parth Patil: And so like when someone says like, "That's hard," like in many cases you should ignore them because they've never even tried. So why would you listen to someone who doesn't ... You know, if you ask someone who's done it, they, they're not gonna say ... Th- they're gonna give you a reasonable path forward. But if you ask someone who just watches other people do it, they can't ... Like you don't go to that person for advice. So you don't go to someone for advice on the Rubik's Cube if they've never picked one up and solved it.
[11:51] Sarah Enix: Hmm.
[11:52] Parth Patil: You go to like-
[11:52] Sarah Enix: What was like the fastest time? Is that, is that not cool to ask? I'm not sure what the rules are here in this culture.
[11:55] Parth Patil: No, it's cool. It, that's, that's, that's cool to, that's cool to ask.
[11:58] Sarah Enix: Okay. (laughs)
[11:58] Parth Patil: So if there's speed cubers, they'll know. Like I, I, you know, in the beginning ... So y- the first time I ever solved it, I had to look at the manual and it took me three hours. And then, um, I got obsessed. So that's like all I did for the next like couple days. And two days later I was down to like two minutes.And then a week later, I was under one minute. And then, uh, my friends were like ... They saw that, right? They go- they- they saw me go from not knowing to under a minute in like- in less than a month, and then they started buying cubes. And we all bought fast cubes and we got speed cubes. Actually, now, I don't know where I put it. I have a cube, but now, my cube now, like, has Bluetooth. It's connected to my, my phone and can, like, send information to my phone base- and like- it basically has tech inside the cube. It's a smart cube, so it's like-
[12:48] Sarah Enix: Wow.
[12:48] Parth Patil: ... tracking stats. But basically, like, we got obsessed with the cube, so then we play- we s- we built- we solved bigger cubes. I solved them blindfolded. And then people were like, okay, like I'm not ... Most people are not gonna go memorize enough techniques to go solve a cube blindfolded. Like, that's a little bit more of like a effort to clear. But then my brother has- like, he told one of his friends that I could do it blindfolded and his brother, like, his- my brother's friend ended up setting like North American records and can solve like 45 cubes in one sitting with his eyes closed. So even within the world of cubing, like there is a huge like- like you can do way harder things.
[13:24] Sarah Enix: Mm-hmm.
[13:24] Parth Patil: They're way harder challenges. And then- and then I think it's like you see someone that's like 15 and then they're doing something and they're like- and you- you realize, like, they've found a smarter way to attack this problem than you. And then when you talk to them, they explain it and you're like, wow, like this is cool. Um, and so I- I- I'm- that's why I loved the speeding cu- speed cubing community. If you ever like go to Netflix, they have- there's a documentary on speed cubing. It's really good. And like-
[13:50] Sarah Enix: Not that you're in it. No, I'm just kidding. Okay, well I-
[13:51] Parth Patil: No, my fri- no, my- my friends- my brother's friends are in it. And so they're- it's kind of like, oh wow, like it's just a bunch of kids playing around. And then those people are at the front of this game. And so I like that. Uh, yeah.
[14:02] Sarah Enix: Oh, that's awesome. Um, no, I think, I mean, I'm seeing all- that's why I love talking about like who you were 'cause I see so much of, you know, the zoo tycoon, the speed cubes. Now you're obsessed with solving different kinds of problems.
[14:14] Parth Patil: Yep.
[14:14] Sarah Enix: Tell me about the main problem you're trying to solve right now.
[14:18] Parth Patil: Yeah. So pretty much ever since ChatGPT came out last S- November 2022, I think. Um, so when it came out, I was hanging out on- on Clubhouse, which is like the- like the s- the audio app from the pandemic. So I was hanging out on Clubhouse and I me- I kept meeting people that were ... And it was like the most popular topic in every single country, in every single language. People are talking about ChatGPT and how it was like changing the way they think about different things. And I'd- I met people, like farmers from my home state in India, and they were like using ChatGPT to help think about their crop cycle planning. And in that moment I was like, and I was talking to these guys, I was like, wow. Like, you're basically got a data scientist in your pocket. And then I'm starting to think like, wait, but even like- and I'm a data scientist, data analyst. So I was like, even like I can't do many of those things and you have it ... So- so instead of thinking about it as like, oh, it r- you know, it helps me reply to an email, I'm thinking like, well for someone else this is an actual computer. Like this is a- it's like a research lab. It's like- it's like your analytics team, you- for the average person. Which is crazy, like the idea that all of us walking around and for $20 a month you can have a- like the best AI on the planet operating on your behalf, solving your problems.
[15:42] Sarah Enix: Mm-hmm.
[15:42] Parth Patil: That, like that hit me like a truck.
[15:46] Sarah Enix: Mm-hmm.
[15:47] Parth Patil: Like the idea that, um, all of a sudden everyone could possibly like tap into this like cognitive superpowers, like ... And that- that- that was pretty insane. And then, and then like I met a guy and he was like, "Oh, you should- you should be talking to GPT-4. It's really good at programming." And I had- I'd kind of like put off programming. Like I- you know, I program when I need to, but it-
[16:08] Sarah Enix: Mm-hmm.
[16:08] Parth Patil: I kind of didn't make it my main role, like for the last couple years.
[16:11] Sarah Enix: 'Cause at this point you were at Clubhouse, correct?
[16:14] Parth Patil: Yeah. Yeah, I worked at Clubhouse.
[16:15] Sarah Enix: Okay, so-
[16:15] Parth Patil: Yeah.
[16:15] Sarah Enix: Yeah, tell me a little bit about like maybe some career stuff alongside these realizations.
[16:20] Parth Patil: Yeah. So, you know, starting off I was like a- and I've alw- I've worked at startups most of my career. I mean, I've been all over the place. So, so coming out of college I worked at a startup, so my internships were in different tech companies. A couple of them were big. Um, so big semiconductor companies and then medica- medical devices. So when I was at big companies, I saw a lot of like, here's how big companies operate. Here's how you ... Like, I didn't think it was right that there was so much paper.
[16:48] Sarah Enix: Yeah.
[16:48] Parth Patil: But then I realized become-
[16:48] Sarah Enix: Literally taking paper (laughs) and then just putting it in a different folder.
[16:52] Parth Patil: Yeah. I was like, uh, I don't know, I got video game characters that are slightly more interesting than like that- that role, in my opinion. I was like let's ... So I was kind of like, I was like, this is interesting. Like, once this gets digitized, then we'll be able to move way faster.
[17:06] Sarah Enix: Mm-hmm.
[17:06] Parth Patil: So like, w- one thing that frustrated me was like you go to a finance meeting and you'd be recapping like some spreadsheet and I'm like, that spreadsheet was only up-to-date two weeks ago. Now you decide to look at it? And I was like, I guess we just are okay with information being slow, like reviewing things on a quarterly cycle. So that- that was like my mindset was like, okay, some parts of the organization don't have the ability to think quickly because the systems are kind of old and like clunky. But then at- then I was like, okay, maybe it's like a big company thing, so I need to go to a small team, see if they think differently. So I- I- then I went to my first startup. I worked at a company called ChowNow and they do like- they do like online ordering for restaurants.
[17:45] Sarah Enix: Mm-hmm.
[17:45] Parth Patil: And that's when I was like exposed to the idea of analytics. Like what does it look like when you can answer a business question about today's business today?
[17:55] Sarah Enix: Mm-hmm.
[17:56] Parth Patil: Like something happened an hour ago. Can you answer a question about it? Like do you know like what's working and what's not working? And up until that point, I had- I had been thinking about how much things cost. So like I was thinking about everything from a finance perspective. And, but then now it's like, okay, but what actually works? Like which products are working? What do people love about us? Like why are they using our product? What's not working?
[18:19] Sarah Enix: Mm-hmm.
[18:19] Parth Patil: And those are like all analytics questions. And so at ChowNow I had the opportunity to-... put together a team and build a data platform and try to understand that. So we went from Google Sheets and spreadsheets to analytics in, like, two years, which is pretty awesome. Um, and I got exposure to this, like, concept of, like, being able to calculate things about a business really quickly.
[18:43] Sarah Enix: Mm-hmm.
[18:44] Parth Patil: And, like, getting deeper insights. Being a- so, like, if you gave someone the right dashboard, then when they're on the phone with the customer, like, they can actually be really, really, like, like they know everything about that customer's business. Then that person comes off as an expert and can really position the product in the right way and figure out what's-
[19:01] Sarah Enix: And you were building, like, a, a dashboard like that.
[19:03] Parth Patil: I built dashboards for literally everyone. Like that was my, that was my bread and butter. I was like, "I built dashboards for sales reps, for account managers, for customer support reps." And then, like, those people would love the dashboards and then, and then, like their managers would be like, "Can you build one for me?" And then, so I was building bashwords, dashboards for every, every layer of the business and, and that was really cool. And then, um, I took a break from Ciao Now! to work on, uh, in, in 2020 there was a presidential campaign, right? And one of the candidates, his name was Andrew Yang.
[19:36] Sarah Enix: Mm-hmm.
[19:36] Parth Patil: And he was, uh, he was running for president on this idea of, like, math and introducing, like, quantitative, like, ideals to, to this, like, very, like, popularity contest world that we live in. I was like, "Oh, here's a nerd. He's, he's running on math. Like, no shot, but I'm down. Like, whatever." Like, that's the kind of, like, I was like, okay, like, I saw a nerd, I saw a guy who was thinking about automation, uh, s- thinking about, like, the future and, like, trends in labor. And I thought like, okay, well, you know, it m- it might not be your traditional politician, but here's a way of thinking about problems that could be useful on this larger stage. And people thought I was like, they were like, "Nah, you're, like, 50 years in the future with this automation stuff." Little did I know that ChatGPT was coming out. But, so I, so I did that, I worked on a presidential campaign. Super fun. Met some of the craziest, cool, like, builders. And I think, like, I think we definitely moved the needle on, like, how people talk about automation. Like, we gave people, like, we probably ... It was weird when we did it, but now it's like now everyone can-
[20:39] Sarah Enix: Hmm.
[20:39] Parth Patil: ... do it and that's, like, pretty important.
[20:40] Sarah Enix: That's so- I did not know you worked on a presidential campaign. That's really cool.
[20:43] Parth Patil: Yeah. It's, uh-
[20:44] Sarah Enix: I wanna see your thoughts on this one.
[20:46] Parth Patil: Yeah, it was-
[20:46] Sarah Enix: We'll just-
[20:46] Parth Patil: No, I-
[20:47] Sarah Enix: ... leave this for AI, but (laughs)-
[20:49] Parth Patil: Um, yeah.
[20:49] Sarah Enix: ... good to know.
[20:51] Parth Patil: But yeah, basically, like, I like to think of us as, like, normalizing talking about automation. And I think we probably underestimated at the time how much of that automation was gonna affect white-collar jobs. But, you know, you could not have ... I, I honestly think that ChatGPT came out 100 years before, in my mind, it was gonna come out. Like, in my mind, I'm basing my expectations of the future off of, like, sci-fi-
[21:16] Sarah Enix: Yeah.
[21:16] Parth Patil: ... plus mixed with, like, the tech that I use. So it's like, "Oh, I play video games." And it's like, "Oh, what if that made it into this industry?" Or, like, these people using this, it's really expensive, but, like, 10 years from now, if it's cheap, then it's gonna be used in this other way. And I always thought, like, we had, we had Siri, we had Google Assistant, we had Cortana, we had all these like ... And we had video game characters you could talk to. Not, not in this natural language, like, ChatGPT way. But I imagine, like, eventually you'd probably get, like, something like ChatGPT. It'd just probably be 100 years out.
[21:50] Sarah Enix: Yeah.
[21:50] Parth Patil: Like, you watch enough sci-fi, you, you imagine it. It's like, "Ah, it's reasonable." But when ChatGPT came out, I was like, "Whoa, this is 100 years early. What? Like-"
[21:57] Sarah Enix: "We're here."
[21:58] Parth Patil: "We're here." Like-
[21:58] Sarah Enix: "We've u- we've ha- we've time traveled. We are now-"
[21:59] Parth Patil: "This is not my grandkids," like-
[22:01] Sarah Enix: (laughs)
[22:01] Parth Patil: "... this is not my grandkids' problem." Like, everything that I thought, like, I wouldn't have to deal with in my lifetime, all of a sudden, is now an opportunity.
[22:08] Sarah Enix: Best present ever. Okay. So you left-
[22:10] Parth Patil: Yeah.
[22:10] Sarah Enix: ... the presidential campaign ended, and then did you jump into Clubhouse? Did you take some time?
[22:14] Parth Patil: No. So I spent, I spent some time, uh ... That was when the pandemic hit.
[22:18] Sarah Enix: Yeah.
[22:18] Parth Patil: So I worked, I worked, uh, I worked with two friends in the Bay Area on, like, a productivity tool in the web browser 'cause we were like, "Well, what if, like, more work ends up in the web browser?" Like, "We should probably solve this tab clutter problem, 'cause everyone's got a zillion tabs, taps." Which was fun, but just like, you know, I don't know if it was, like, enough to build a business off of.
[22:37] Sarah Enix: Yeah.
[22:37] Parth Patil: It was very like ... But it was, it was a good, it was building something from scratch, um, with friends from college. Like, there's nothing quite like that, you know.
[22:45] Sarah Enix: Hmm.
[22:45] Parth Patil: "The world's gonna end, get your friends together, and just try to do something useful-"
[22:48] Sarah Enix: Yeah.
[22:48] Parth Patil: "... is, was what we were thinking." Um, didn't work out. So then we kind of phased it down and then, then I worked on Clubhouse.
[22:56] Sarah Enix: Yeah.
[22:57] Parth Patil: And then, um, pretty much ... Okay, yeah. Then like 13, 13 months ago, GPT-4 comes out and I start programming with GPT-4. And then we had layoffs at Clubhouse and I thought to myself, "Well, this is, this is great. Now I get to focus my entire, like, attention on this new problem space."
[23:14] Sarah Enix: This is, this is where the, the good stuff happens.
[23:16] Parth Patil: Yeah.
[23:16] Sarah Enix: Okay.
[23:16] Parth Patil: This is where every, this is where I started going crazy.
[23:18] Sarah Enix: Yes.
[23:18] Parth Patil: 'Cause basically, like, I was like, "Okay, this ..." I was writing programs with GPT-4, and I was writing programs that could write other programs.
[23:27] Sarah Enix: Yeah.
[23:27] Parth Patil: Just to see, like, how good GPT-4 is at generating code. I think it's better at generating code than I am, like, being a self-taught programmer-
[23:35] Sarah Enix: Okay.
[23:35] Parth Patil: ... that's, like, new to this. So for me, GPT-4 is like a, it's like a CTO. It's like someone you count on to be technical in ways that you aren't. And I always have it pushing me to this new level. So I go to ChatGPT and it's always like, "Teach me how to do this in Python. T- like, teach me how to clone my voice. Teach me how to build a chat, chatbot." All those questions are questions that I asked, like, a year ago.
[23:56] Sarah Enix: Hmm.
[23:57] Parth Patil: And not only can GPT-4 answer some of those questions, but it can teach you how to implement the thing.
[24:03] Sarah Enix: Mm-hmm.
[24:03] Parth Patil: So it's like, I'll be like, "Teach me how to clone my voice." And this was, like, a random Sunday. I was like, "Teach me how to clone my voice." And it showed me how to use some, like, open source text to speech library. So I was like, "Well, actually, teach me how to install Python and set up my programming environment." And then it shows you how to do that. And then I was like, "Whoa, this is, like, the least judgmental learning experience of all time."
[24:23] Sarah Enix: Wow.
[24:23] Parth Patil: And, and-General, I mean, I have an opinion that like programming is uninviting, because the introduction experience is not intuitive or like friendly. So, like you don't know where to-
[24:37] Sarah Enix: I spent like an hour on GitHub last week, and I was like, "I feel like the stupidest person out there."
[24:41] Parth Patil: Which is exactly the problem, is like there's like three hours of other things that they have you doing before you actually write a program.
[24:47] Sarah Enix: Yeah.
[24:47] Parth Patil: And the only thing that actually matters is that you have fun writing programs, because the other stuff is ... Like, I don't think you should spend tu- any time on GitHub until after you've like figured out, like ... Uh, like, I wish it was just, you just jump into programming.
[25:00] Sarah Enix: Yeah.
[25:00] Parth Patil: And I think we're getting tools that like, tools like Replit are exactly that like new form of thinking, where it's like, well, what if the software is so smart it handled all the like stupid setup stuff? And then you could just focus on like, hey, what if we made this cool new thing? And then AI helps you make the prototype.
[25:17] Sarah Enix: Mm-hmm.
[25:18] Parth Patil: Maybe it writes the first version of your program. But all of a sudden you realize ... So what happened was like, I, I was like, "Yo, teach me how to clone my voice." It teaches me, it writes the first version of the script. It teaches me how to install Python, set up my environment. So I have this like, generally like you go to someone more experienced than you and you ask them to set up your computer, 'cause that's like, I don't know how else you're supposed to learn how to get set up. But now you have AI, and it's there 24/7. You can ask stupid questions. Which is-
[25:43] Sarah Enix: So you're like, ask your, like your prompt is, "Hey, can you clone my voice? Give me the code to do that in Python."
[25:49] Parth Patil: Yeah.
[25:49] Sarah Enix: And then you're downloading, you know-
[25:51] Parth Patil: Well, I was like, t- I was like, "How do you clone your voice?" And then it, it shows you a bunch of options. And I was like, "Okay, actually like how would you do this in Python?"
[25:58] Sarah Enix: Hmm.
[25:58] Parth Patil: And then it's like, "Here's how you do it." And then I'll be like, "Well, actually, I don't even have Python installed on my computer." And then it's like, "Okay, here's how you set up Python." So like at every level it's trying to meet me where I am.
[26:10] Sarah Enix: Yeah.
[26:10] Parth Patil: So I'm like, "Okay, like how do I do the thing? Okay, now I know how to do the thing. But like how do I actually like do..." Like every individual step of like, that was previously ... Basically in three hours, you and this chatbot will get something done. And, and then at the end I was like, "Oh my God. I thought this was gonna take me like three or six months."
[26:29] Sarah Enix: Wow.
[26:30] Parth Patil: But really, it's just like you ask the right questions, you sit there for an afternoon, and the chatbot will do most of the work. And then it might make some mistakes, but you ki- as long as you, you know, you feed those errors back in, and you're like, "Hey, can you help me reflect on what went wrong here?" And then it's like, "Oh, by the, oh, you're right. I made a mistake." And so you have this back and forth, this intelligent like back and forth with this system that also knows every programming language. Like it's not just Python. Like it's, it can-
[26:57] Sarah Enix: Yeah.
[26:57] Parth Patil: It understands like so many topics. This is the thing about general intelligence is that like I can say, "Help me think about a new way to rank music," and I can sit here and program a new way to rank music for myself using a chatbot.
[27:11] Sarah Enix: Mm-hmm.
[27:11] Parth Patil: Like or using GPT.
[27:13] Sarah Enix: Mm-hmm.
[27:13] Parth Patil: So you can basically ... So then tha- then I realized like you don't need, like you, you just need to have an idea of what you want, and then you need to turn that into the right questions. And then the AI can help you figure out how to make that real.
[27:28] Sarah Enix: Hm. Mm-hmm.
[27:30] Parth Patil: And there are things that have not been done yet that are possible that people are going to build because AI told them it was possible.
[27:37] Sarah Enix: Yeah.
[27:37] Parth Patil: It was within their skillset, right? Like-
[27:39] Sarah Enix: But that, that catalyst kind of moment when you were saying, "Oh, I wanna know what it's like to clone my voice," that didn't come from a chatbot. That came from you.
[27:48] Parth Patil: Yeah.
[27:48] Sarah Enix: So I think that's kinda the stuff that excites me of like, oh, what are the problems that we're seeing in the world? And then like getting curious about them, having some time-
[27:56] Parth Patil: Exactly.
[27:56] Sarah Enix: ... to reflect on them.
[27:57] Parth Patil: Yeah.
[27:57] Sarah Enix: And then the possibilities are endless with a chatbot, with ... I mean, I know you love OpenAI.
[28:02] Parth Patil: Yeah.
[28:02] Sarah Enix: I love Anthropic products, but it's okay.
[28:04] Parth Patil: No, I have, I have a, I have a Claude key. I, I, you know, I'm-
[28:07] Sarah Enix: Okay.
[28:07] Parth Patil: ... I'm benchmarking Claude versus GPT-4. So-
[28:09] Sarah Enix: Okay.
[28:10] Parth Patil: I'm, and then I, I'm ... This weekend I'll probably pick up Gemini as well, just-
[28:15] Sarah Enix: Oh, okay. Nice.
[28:15] Parth Patil: I, my thing is like I'm not, I'm not attached to one platform. I'm like, more like you can con- you can expect me to use state-of-the-art language models wherever I ... Just 'cause if you're gonna talk ... I think the difference between 3.5 and 4 is, for me, very obvious.
[28:31] Sarah Enix: Yeah.
[28:32] Parth Patil: But it's like, it's important enough to me that ... Basically, GPT-4 is able to use tools, is able to like handle larger context. That's more important to me than simply autocomplete. Like if you, uh, I think the, the difference is almost like between a high school, 3.5 being closer to a high school student and GPT-4 being like a grad student with a laptop.
[28:56] Sarah Enix: Yeah.
[28:56] Parth Patil: Like-
[28:56] Sarah Enix: I mean, that's what I, that's how I-
[28:58] Parth Patil: Yeah.
[28:58] Sarah Enix: Like that was exactly my journey was I wasn't paying for any of them, and then I was like, "Okay, I'm gonna start paying for ChatGPT-4." And I was like, "Okay, this is great. I love this." But like-
[29:06] Parth Patil: Yeah. It's inv- it's, it's the best investment you can make in your own intelligence.
[29:10] Sarah Enix: Yeah.
[29:10] Parth Patil: Like you'll, like people are like, "Oh, I don't know if it's worth it at $20." Then I just put them on my family plan. Like I've ha- I have a teams plan, and I just have friends on it.
[29:18] Sarah Enix: Mm-hmm.
[29:18] Parth Patil: Because I would rather them discover like what a language model that can use tools is capable of than what they think is possible with 3.5. There is a learning curve to using these tools, I think. You know, like the, getting the most out of it requires you to communicate pretty effectively-
[29:36] Sarah Enix: Mm-hmm.
[29:36] Parth Patil: ... like your own strengths, but also like your hopes, what you hope is possible-
[29:40] Sarah Enix: Mm-hmm.
[29:41] Parth Patil: ... what you want to learn. And that requires effort. Like every button press is actually effort.
[29:46] Sarah Enix: Mm-hmm.
[29:46] Parth Patil: So everything, how you craft your prompt is effort. And it's not ... If you don't see why you need something like this or why some technology like this could be useful, like you may never discover what it can do for you.
[29:57] Sarah Enix: Mm-hmm.
[29:57] Parth Patil: Right? Like so there is a learning curve here, and I think this is why I'm so interested in this. Partly because I thought it was gonna be even more obvious to people, but maybe it hasn't been. And I think in some ways, like some people pick up the technology faster than others, and that interests me. Like why is it that kids are so good at using ChatGPT? And then when you go, I go to UCLA and I talk to a bunch of kids, and I realize it's 'cause they need it.
[30:21] Sarah Enix: Yeah.
[30:21] Parth Patil: Like they, you know, they have too much homework. And then they're like, "Well, if you assign me six essays, I'll probably generate-"... three or four of them. (laughs)
[30:30] Sarah Enix: You're like-
[30:30] Parth Patil: Which, and then the teacher's like-
[30:32] Sarah Enix: ... breeds the most creative solutions. (laughs)
[30:33] Parth Patil: And then the teacher's like using GPT to help correct stuff. So it's, it's really like on every level, education gets transformed. And I genuinely think that like, this is good for education. Um, you know, i- if you have a teacher in your pocket that e- doesn't judge you, that meets you at your level, like, there are so many topics that are outside of my knowledge, like, outside of my aptitude, where I am today, like, what I currently know. But if I go and say, "Hey, just explain this to me like I'm five." "Okay, cool." So then it explains it like ... "Okay, now explain, like, I'm six." And then I just do every single age above that.
[31:07] Sarah Enix: Yeah.
[31:07] Parth Patil: And then like, and then I'll be like, "Okay, explain it like I'm ..." And then I'll literally put my description in. I'll be like, "I'm a data analyst looking to pick this stuff up. I'm self-taught." And then I'll just put in a paragraph of like, "I love games and music." And then it'll like, explain it within that. And then it'll ... and then the cool thing is, language model will try to make that-
[31:24] Sarah Enix: Yeah.
[31:24] Parth Patil: ... difficult topic make sense.
[31:25] Sarah Enix: Yeah.
[31:26] Parth Patil: Try to make it fun and interesting. You could be like, "Explain it to me like I like Spider-Man." And then it'll, you know, try to reason some analogy of, like, Spider-Man into the story. And that, that's really interesting. I think a lot of ... and this goes back to video games and I'm gonna bring it back to video games 'cause that's what I do.
[31:42] Sarah Enix: 'Cause why not?
[31:44] Parth Patil: S- if you think about video games, like, I was talking to my dad and he was like, "You know, I could never have imagined you would just pick up a game and that game would teach you business concepts."
[31:54] Sarah Enix: Mm-hmm.
[31:54] Parth Patil: But that's what RollerCoaster Tycoon is.
[31:56] Sarah Enix: Mm-hmm.
[31:57] Parth Patil: Right? Like, I'm not telling you how to think about a, like, your finances of your park, but you see a number, a green number. You need to turn the profit number up. Like, this, this ride is bleeding you out. But why would a seven-year-old wanna learn something like business? You don't, you don't, like, those things aren't like ... so what I've realized is that, that video games are actually a really powerful way to teach people, and that we take those concepts we learn in video games and we apply them elsewhere. In the same way that you read a book, and then hopefully you take some, some substance out of that book and then you apply it in your life.
[32:32] Sarah Enix: Mm-hmm.
[32:33] Parth Patil: Video games allow you to, like, experiment in this sandbox. I think board games do this too, which is like, in Settlers of Catan, you, you are pr- practicing making deals with people.
[32:41] Sarah Enix: Yeah.
[32:41] Parth Patil: Same thing in Monopoly, right? Like, "Oh, three sheep for one brick." And like, you're like, "No deal," 'cause you need this brick and I have plenty of sheep.
[32:49] Sarah Enix: But you have to remind-
[32:49] Parth Patil: Turns out sheep-
[32:50] Sarah Enix: ... someone else and you're like, "Okay, I have to say yes to everything that you say." So, I mean-
[32:53] Parth Patil: Yeah. And then you realize-
[32:54] Sarah Enix: ...)
[32:54] Parth Patil: ... you're on the receiving end of, like, a deal where, like, you're not getting ... so, so all that comes, like ... d- you play 100 games of Catan, I guarantee you, you learn something about making deals with people and what a good deal looks like and when to make deals. But it's not like that's ... that's not like the, it's not obvious to people from the, like ... same thing with video games. Get, they sometimes get a bad rap, but my dad, he, he, he pointed this out, like, last year, last time.
[33:17] Sarah Enix: Did he eventually buy you the PlayStation or no?
[33:19] Parth Patil: No, he didn't buy me ... we, I mean, we built a computer that would've cost twice as much of a PlayStation, so I feel like I won out. But really, he won out 'cause he was like, "Well, now I get to teach you about, like, chips." Right? Like, "What is RAM? What is a CPU? What is a GPU?"
[33:33] Sarah Enix: Mm-hmm. Yep.
[33:33] Parth Patil: And for me, like, he loved computers. My m- my mom works in, they both work in tech. So for me, I ended up getting even more in love with computers because I had that experience of, like, building my own. My brother built his own computer. All of our friends build our own computers. So we, we're all very similar in that way, like, once we have ... yeah. So we have, like, a tribe of, like, PC gamers.
[33:55] Sarah Enix: Mm-hmm.
[33:55] Parth Patil: And maybe, like, in the fu- I think that might be a dying thing. Like, I don't know. Like, actually, let me see if I can show you.
[34:01] Sarah Enix: Are you showing me your setup?
[34:03] Parth Patil: Yeah. Well, it's probably better for ... you can't really see from this angle, but it's, like, RGB. It's, like, tons of, like, it's just rainbow colored. I can control all the lights too. So for me, it's like a project I've been working, eh, I've been working on this since, I mean, this is a new, like, some of the par- one of the parts came from the original computer I built. But, like, you're always upgrading it. You're thinking about-
[34:22] Sarah Enix: Yeah.
[34:22] Parth Patil: ... what you need. It's kinda like a car. Like, I'm-
[34:25] Sarah Enix: Yeah.
[34:25] Parth Patil: ... I don't have a car, but I imagine how people, car people are. When I talk to car people, I'm like, "Oh, you're feeling what ... like, your, your relationship with your car is very similar to my relationship with my computer." Like, your car takes you on a freeway. Mine takes me on the, like, the internet freeways, right? So.
[34:39] Sarah Enix: (laughs) Yeah. Uh, the longest journey of them all.
[34:42] Parth Patil: Yeah.
[34:42] Sarah Enix: Um, no, I love it. And I love hearing all of the origin story, is amazing, because I know we're already, you know, 30 minutes in and we haven't even gotten to, what are you working on now?
[34:51] Parth Patil: Yeah.
[34:52] Sarah Enix: We've made the people wait. Um, but do you wanna tell us-
[34:55] Parth Patil: We can edit it down, but yeah. (laughs)
[34:56] Sarah Enix: No, we're not doing that. We're gonna keep-
[34:58] Parth Patil: All right.
[34:58] Sarah Enix: ... all these. But tell us, tell me what you're working on now.
[35:01] Parth Patil: Yeah. So once ChatGPT came out and I started writing programs again, started writing Python, it started teaching me Python. Um, and I started doing that, like, every single day, like, Saturday, Sunday. Every day, I'd sit down for a couple hours, talk to ChatGPT and have it teach me some, some concept in programming. Because for me, I was like, "Okay. Learning this has never been easier." And then also, because I can use this newfound knowledge of programming to build tools like ChatGPT. So, like, if you actually sit there and ask ChatGPT to help you build a GPT-3 powered chatbot, which I did this once. So I was just like, "If I can clone my voice in three hours, how long does it take to build a chatbot?" So it turns out, you ask, you sit there and you say, "Teach me how to build a GPT-powered chatbot." It'll write the first version of that script. And then you'll be surprised because it's, like, 15, 20 lines of code. It's not gonna look, it's not a full feature, rich. It's not ... Chat- ChatGPT has, like, a couple hundred people working on it. But you can have a chatbot in a loop, sitting on, you know, running on GPT-4 in, like, 15, 20 lines of code. That's what OpenAI has done. They've, they've made it possible to leverage AI without having to train the underlying model. So you can use their model, which is really good at generating text. They also have DALL-E, generative, great at generating pictures. So I don't have to train an image model. I use DALL-E. Um, they even have a vision model. So you can put an image in or a video in, and then it can tell you ... it can, it can try to reason with you on what's happening in that image or in that video. So if I'm working on music.You can take a screenshot of your screen, send it to GPT and be like, "Help, I'm stuck. Try- I'm trying to, like, arrange my song. How do I, like, f- flesh out this song?"
[36:46] Sarah Enix: Mm-hmm.
[36:46] Parth Patil: And it'll be like, "Well, maybe you move this section to the intro."
[36:49] Sarah Enix: Yeah.
[36:49] Parth Patil: So, vision gives the language model even more, like, of a clear picture of, like, what your problem is, and it's very information-depth. So, like, I think the kind of stuff that I do these days is, like, well, what kind of context do I need to give the language model to solve my problem? It might be, like, watch my screen, or it'd be like... Or, uh, or I might be like, I want to come up with a plan but I want to go through a bunch of different options. So before I start my plan, I'll have a language model come out with, like, five or six strategies, and then it'll go through each strategy, evaluate some pros and cons, and I want it to take into a bunch of different considerations. Like, here's my skillset, or I'm alone, so, like, don't- don't give me a solution that requires 50 people. Like, these are things that are interesting that you can, like, you can start... Like, I think saying, like, uh, I spent one month on GPT just, like, working on music projects, just having it teach me music in a way that's, like, um, fun, right? So, like, turns out, turns out the language model, having absorbed so much knowledge of the internet, it's got some understanding of music theory. It's got some understanding of, like, the tools you use. So if you say, "I'm working in Ableton," uh, like, it- it- it- it's, like, absorbed the Ableton manual, right?
[38:05] Sarah Enix: Yeah.
[38:05] Parth Patil: Or you can even put a synthesizer manual behind a language model and then have it pretend and just, like, teach you that, how to use that tech as if it's, like, the engineer that built it and wrote that documentation. So I think that's awesome. I think, I think language models are a faster way to learn about the world than even YouTube. And I thought YouTube was-
[38:22] Sarah Enix: And you're gonna release your music video soon, right?
[38:24] Parth Patil: Well, there is a video. Um, if you, if you go to my YouTube channel, there's, there is a music video, and I've made it... So I wrote the song myself, but then I used Midjourney, GPT-4, and... So I basically went to GPT-4 and I was like, "Hey, I have a tool called Midjourney. It generates images. I also have a tool called Runway. It takes images and turns 'em into video four seconds long." The reason you have to give this context is because the language model is, only knows something, know, like, has a knowledge cutoff. So it doesn't actually know what these tools are yet. Maybe GPT-5 will. But when you say you have acce- like, "Here are the tools I have. How do I make a music video that's maybe three minutes long for a bass music song?" Like, if I can t- generate images, and I can, and they can be pretty much whatever we imagine, and then I can turn those images into four-second clips of video. And then the language model is like, "Okay. Well, actually, if you add a editing tool, you can also, like, extend those four-second videos."
[39:17] Sarah Enix: Do this, do this, and then do this, and then do this.
[39:18] Parth Patil: It's just like, "These- these are your options." It's like, "We can generate, like, 50 different images. The ones you like, we'll feed 'em into the video model, and then, you know, generate a couple hundred videos, and then the clips that you like, we'll put 'em into the, into the, the, like, editing track, and then we'll, we'll..." It'll be like, "Oh, and then you can beatmatch it," right? So, so like, "If your BPM is this, then- then actually, like, you wanna switch the, like, you wanna switch the shot every X number of se- seconds." And I, all of this is just, like, you know, I'm not... Du- I'm not, I'm not, you know... We're- we're in LA, but I'm not from Hollywood. So for me it's like, "All right."
[39:50] Sarah Enix: That's not bad.
[39:50] Parth Patil: "Like, let's try it." And then, uh, I basically pulled two all-nighters, and I just was generating, generating, generating, and I had a GPT that could, that would... So GPT-4 would write the prompt for Midjourney and it would generate 24 images at a time. So it was just generating, generating. I'd be like, "I like this one. I like that one." And so there's a lot of curation involved. But in two days I had a music video for my song. And I was thinking, like, "Wow, like, I didn't have to ask anyone permission to do this. Like, I didn't have to hit an arbitrary..." I don't have any Spotify listeners and I don't have a, like, public profile as a musician. So, but for me to have a music video, I was very proud. I was like, "Whoa," like, "I wro- I wrote this song but now I have a music video." And then my friend who's a DJ, he was like, "Whoa, do you wanna do the visuals for my, for my set?" And then I was like, "Whoa," like, "that's not even, like, my job, but I'm..." And he was like-
[40:43] Sarah Enix: That's dude.
[40:43] Parth Patil: "... I'll pay you for it." And I was like, I was like, "But this is awesome. Like, I get to do this."
[40:46] Sarah Enix: Do you remember the Apple screen, the one where it would, like, the, it would have, like, the crazy backgrounds-
[40:50] Parth Patil: Yeah.
[40:50] Sarah Enix: ... and, like-
[40:51] Parth Patil: Exactly. Exactly. It- that's exactly the kind of stuff. I was like, "You just come up with, like, trippy visuals," and then some DJ is like, "Well, I want those trippy visuals." "Cool, put my logo on it." Yeah. It's true. So- so for me it was like, oh, I can now do these things that I used to think were, like, another, like... I think about multiverse Parth. Like, I think about, "Well, well, am I having the most fun? Am I actually doing the right thing?"
[41:13] Sarah Enix: Yeah.
[41:13] Parth Patil: Or is there a different version of Parth in a different universe that's doing, like... What about the version of me that's making video games?
[41:19] Sarah Enix: Mm-hmm.
[41:19] Parth Patil: Like, is that guy having more fun than me?
[41:21] Sarah Enix: (laughs)
[41:21] Parth Patil: Or is the version of me that's, like, only making music instead of, like, working on tech? But now because of AI you can kind of do multiple things, because the amount of leverage you get out of these tools means that you can reach into these other skillsets. You can, you can, you can use tools that would previously require a whole team, right?
[41:40] Sarah Enix: (clears throat)
[41:40] Parth Patil: So even, within data science, the fact that I can use vision models, the fact that I can generate text, but I am not the one training the model, like, on a... Like, I'm not training GPT-4.
[41:50] Sarah Enix: No.
[41:50] Parth Patil: But I'm building, like-
[41:50] Sarah Enix: Yeah.
[41:50] Parth Patil: ... applications around GPT-4 that leverage it. Same thing, when you build a custom GPT in ChatGPT, it doesn't require programming skills. It requires, like, you're- you're designing an assistant. You're saying, "Here's, like, your role. Here's how you present information." Like, you like to make charts.
[42:07] Sarah Enix: Yep.
[42:07] Parth Patil: Or you like to write programs. Like, uh, generally you have a bi- Like, so you can create-
[42:13] Sarah Enix: Yeah.
[42:13] Parth Patil: ... like, a character.
[42:13] Sarah Enix: You as, like, a boss have a vision-
[42:15] Parth Patil: Yeah.
[42:15] Sarah Enix: ... and you know where you want to go, and you just have to, like, "Hey, tell me how to get there," or, like, "Organize this data," or what have you.
[42:20] Parth Patil: Yeah. And you can design that assistant to be, like, what you need, right? You could be like-
[42:23] Sarah Enix: And is that what you're building now, a bunch of assistants?
[42:26] Parth Patil: Yeah. My- my main focus is, like, I build assistants on GPT-4.
[42:30] Sarah Enix: Okay.
[42:30] Parth Patil: So, like, all kinds of assistants. Everything from custom GPTs to, like, programs that run on my computer that, like, control my mouse, that can, like, see my screen when I work, um, to stuff that, like, you know-... something that goes through your emails, reads all 800 emails, and then is like, "Hey, these are the five things you need to actually respond to. Everything else, I've already categorized and moved away." So I think, uh, building, building chatbots is fun for me, because I think it's not clear where, like, where they actually go. But I do know where they are going right now. It's like, they initially, like, it was like, ChatGPT can generate text so it can come up with strategies and ideas. So I built, like, a strategy generator, and then I was like, "Well, what if you gave it tools? Can it actually do the job instead of telling me I gotta do the job?" Like, like, how far does this go? So it's like, turns out if you connect it to the internet, it can answer questions about what's happening in the real world. Or if you give it access to your email, it can actually, like, label the emails. It can draft a reply. Um, if you give it access to Notion, it can, you know, like, you can have it manage a Notion page, populate a page, or, like, run a database. Or things like a data analyst used to do would be, like, write SQL queries, and you'd write a SQL query to generate a chart. But actually, ChatGPT can both write the SQL query and gen- write the code to generate the car- chart. So data analysis means you're no longer writing things by hand and getting stuck at the code layer, but the data analyst is more valuable asking intelligent questions to a language model. The language model should do the, the nitty-gritty of, like-
[44:05] Sarah Enix: Mm-hmm.
[44:05] Parth Patil: ... scripting that's, like, build the chart. And if I don't like the chart, you just tell it, "Oh, actually, I want to explore this new trend, this other aspect." And the language model will get to work rewriting the analysis, running it on the fly, and then it... That feedback loop is really, is really powerful.
[44:21] Sarah Enix: Mm-hmm.
[44:21] Parth Patil: And-
[44:21] Sarah Enix: And you have some copycats already, right?
[44:24] Parth Patil: Oh, yeah, yeah, yeah. When I created my... I have created a bunch of GPTs, and on the day that the GPTer- GPT store released, there were six copycats of my favorite. The one that, like... My most popular GPT had six copycats on day one. And I know they're copycats, because the other people who copied, uh, it used the same Midjourney image that I generated. And you know how image generators are. Like, you can't generate two of the same thing. So if six people name their project the same thing as you, and they copy your image, it's like... And your image was generated by you, then it's like, you know you have copycats. And I think there's, there is this aspect of, like, um... Yeah, it's like a new Wild West. So there are people and creators. There are people that don't want to be creative, and I guess they're taking notes. I don't know. (laughs)
[45:10] Sarah Enix: They're like you.
[45:11] Parth Patil: Yeah.
[45:11] Sarah Enix: I have a question, like a tangible thing.
[45:13] Parth Patil: Yeah.
[45:13] Sarah Enix: I know we've talked about this, so, you know, answer this how you want, but-
[45:17] Parth Patil: Mm-hmm.
[45:17] Sarah Enix: ... for example, like, if I want to write a bunch of LinkedIn posts in my voice of saying, you know, "Hey, I'm... This is who I am. This is my writing," I train it, and then say, "I want to write a LinkedIn post about cars and how cars are my favorite thing from childhood, and give me a business example, and make a LinkedIn post." Like, how would you start that, uh, with the agents that you've built so far?
[45:43] Parth Patil: Y- so I would... Uh, I think, like, the thing... I have an opinion on AI-generated content on LinkedIn. I think there's a lot of GPT out there, and there's a lot of, like, people re- auto-replying with GPT. I don't think that's a good thing.
[45:58] Sarah Enix: Okay.
[45:58] Parth Patil: Because the more people spend time with language models, the more you start de- Like, you will detect some of the patterns that you see in how a language model presents information. Like, it uses similes and metaphors that we, that aren't accessible.
[46:12] Sarah Enix: Yeah.
[46:12] Parth Patil: It, it's like, where it'll be, like, very, like, corny in some ways. And so I, I kind of have the opposite approach to LinkedIn, which is, I bring my actual personality into the post in LinkedIn. If I use AI, it'll be to generate an image.
[46:25] Sarah Enix: Yeah.
[46:25] Parth Patil: But if I'm generating an image, it's not something that's real. It's something that's not real. Like I'm-
[46:29] Sarah Enix: You use LinkedIn as Twitter. You just have, like, one account that you just, like-
[46:31] Parth Patil: Yeah, yeah, yeah. I j- I just tweet on LinkedIn. That's, yeah, that, like, I tweet on LinkedIn. And I don't really even... It's not like I'm, like, known on Twitter, but I, I'm very just like, I'll put that short quip, and I might put an image that represents it. But I think, um... Yeah, I, I, I don't think that using GPT-4 to craft copy for social media is the best way necessarily to use GPT-4. Yeah, or, or, or language- or agents. And I think, I think if you're building, like, chatbots, like, you should help them do... Like, help them actually get... amplify your message instead of, like, try to, try to replace it. I think h- the more, like, what, what people are gonna... I guess it's like, you could, you could have GPT generate the copy for your post, but if it's not grounded in some interesting opinion that you have based on some project or something that you've made, it's gonna come off boilerplate. And then it's like, are you better off not having posted? Like, maybe you should have just been yourself, right? Like, especially if other people are increasingly using the same-
[47:39] Sarah Enix: Less authentic.
[47:39] Parth Patil: Right. If everyone else is, is, like, using GPT, then you kind of stand out by being, like, casual.
[47:45] Sarah Enix: Yeah.
[47:46] Parth Patil: Right? Like, GPT is-
[47:47] Sarah Enix: Maybe it's a good thing. Like, hey, take and, like, write your article and then say, "Hey, ChatGPT, for whatever agent this is, give me a social media marketing strategy so I can make 15 pieces of content out of this one."
[47:59] Parth Patil: Yep. That, that's a lot better, I think.
[48:02] Sarah Enix: Okay.
[48:02] Parth Patil: Yeah, yeah, yeah. Helping you come up with strategies, helping you think about cross-platform, that's really where the language model s- shines. I think it's like... Transforming con- text copy is, like, very, uh, uh, m- people will, they'll jump to that. But w- when you say, like, "Oh, I w- I want people to find me. I want people to know that I have a skill." Is that coming ac- Is that, is that how I come across when I post something like this? Like, in my case, I might, I might say something that's anti-management consulting, because I'm very... I like to make jokes about management consultants. My friends are management consultants. I just do that because it, it's fun.
[48:39] Sarah Enix: 'Cause you love 'em. (laughs)
[48:40] Parth Patil: But I'll take, I'll take what I'm about to post, but if we go to GPT and be like, "Does this offend anyone?" And it'll be like, "This could be... This could make management consultants kind of annoyed." And I was like, "Okay, good. That's exactly what I'm going for." Right?
[48:51] Sarah Enix: "Perfect. That's exactly what I wanted to happen." (laughs)
[48:52] Parth Patil: Yeah. So, so, like, maybe having it, like-... see if you come off a certain way or having it like, you know-
[48:59] Sarah Enix: Doing like a sentiment analysis almost-
[49:01] Parth Patil: Dou-
[49:01] Sarah Enix: ... of that.
[49:01] Parth Patil: ... yeah. That, that, that, like basically I think like you should use the language model to reason with you. Like, it's like having like a, like m- for me it's like I tell it to act like a PR.
[49:12] Sarah Enix: Yeah.
[49:12] Parth Patil: But in the same way that like, you know, I'm okay with disagreeing with my PR person, right? When I have a strong opinion, this is how I want to present myself.
[49:19] Sarah Enix: Mm-hmm.
[49:20] Parth Patil: You have to actually have an opinion on how to present yourself.
[49:22] Sarah Enix: Yeah.
[49:22] Parth Patil: Otherwise the language model will come up with something bland. And, and then like you might get engagement but it's not like, is that how you want people to remember you as like GPT?
[49:31] Sarah Enix: Yeah.
[49:31] Parth Patil: No, you want people to remember you for like the thing that gets you, the thing that's diff- like yeah, maybe it's like your style.
[49:37] Sarah Enix: Yeah.
[49:37] Parth Patil: Maybe it's like the vibe. Maybe it's like you want them to realize you think about this thing a lot.
[49:43] Sarah Enix: Yeah. What are your favorite prompts?
[49:44] Parth Patil: So advising how you come off... I, I think that the whole strategy thing is perfect. Like-
[49:48] Sarah Enix: Okay.
[49:49] Parth Patil: ... you know, like how do I make s- how do I present this idea on YouTube? And then it might come up with like a new like idea for content.
[49:57] Sarah Enix: Mm-hmm.
[49:57] Parth Patil: Like a format.
[49:58] Sarah Enix: (clears throat)
[49:59] Parth Patil: But I think that's, that's, that's pretty-
[50:01] Sarah Enix: Yeah.
[50:01] Parth Patil: ... keep it, keep it grounded in like your unique perspective is important.
[50:05] Sarah Enix: Do you, do you have like a blanket advice of someone who's like trying to get better at prompting of saying, "Hey, make sure you, you format your structure like this," or, "This is how to make your prompts better." Like what's kind of your blanket advice there?
[50:19] Parth Patil: So I have two slightly different approaches to this, and they may seem like at odds with each other, but I think they make sense. So I think if, I think like, uh, one, I, I meet people, some people I meet, they'll spend 20 minutes crafting a prompt, and then they hit send.
[50:38] Sarah Enix: Cool.
[50:39] Parth Patil: And I'm like, maybe some of your instructions require that amount of like, like sometimes there's some things that are important and we should be very detailed and we're like, this nuance matters, so I'm going to be, I'm going to spend time on how I communicate this. But the video gamer in me is like, we have hot keys. Like, we should be moving faster than this. Like it's actually, like if, if GPT is like, "Oh, I can't do that," sometimes I'll just be like, "Question mark?" I'll just per- question mark, and it'll be like, "Oh, I'm sorry. I'll get back to work." So sometimes the language model like, it's like you just want to go quickly to the next step. And in that case like, I'll just be like, "Help. Fix." So I'm look- sometimes I'm just looking for the one word that takes me from where I am today to like where, like can the language model... So I'll, I'll screenshot... If my program runs into an error, I'll screenshot it. Paste it into ChatGPT, and I'll be like, "Fix this." And, and I'm trying to get that across in as few button press, as few letters, as few words as possible.
[51:38] Sarah Enix: You don't say please? Isn't it like a statistic that chatbots are better when you're more polite?
[51:46] Parth Patil: That's, even, even if that's like... I know there's a lot of that out there like of like tip the model. You will never see me ri-, like, no, I think those, those things like will, I- get ironed out over time, like, like when the models get better they're not gonna have these like... Maybe they will. Maybe this is what you get when you train language models on the internet. But I, I don't think that's like reasonable to expect regular people to understand.
[52:13] Sarah Enix: Yeah.
[52:13] Parth Patil: And so I would not advise a tactic that is not intuitive.
[52:18] Sarah Enix: Okay. That makes sense.
[52:19] Parth Patil: If that makes sense, because like I, I think it's like the things that are not intuitive, that's on the company to iron that out-
[52:25] Sarah Enix: Yeah.
[52:26] Parth Patil: ... and like, you know, improve the model. If it's not, if, if it required some like arcane specific knowledge, then we're basically in the old world of programming and gatekeeping. Where really it's like, I think there's two modes. One, move fast, in which case try to think about the one word that would allow the model to like take you to the next step of your problem. And then provide the context, so like it's not just one word. If I'm providing an image, there's a lot of information in that image. Right? If I screenshot my, this song that's in, in progress, I'd be like, "Look, I have three minutes of, of music. How do I, how do I expand this to, to five minutes?" Like that image is actually where the context is, and so, and then like the prompt can be less communicative. But if you're lazy, another way to do it is to literally speak. So because speaking is faster than typing.
[53:10] Sarah Enix: Oh yeah. I use mine all the time.
[53:12] Parth Patil: Yeah.
[53:12] Sarah Enix: And I just
[53:12] Parth Patil: use my thumb. On mobile, on mobile, yeah, especially if you're using ChatGPT on mobile. This is where voice is so powerful. Like I see people sometimes do demos of chatbots, and they'll actually like type with their thumbs. I'm like, "What are you doing? Just, just talk to it."
[53:23] Sarah Enix: Use your thumb. Yeah.
[53:24] Parth Patil: Yeah. Do speech to text. You can, and then you can, you can rant, right?
[53:27] Sarah Enix: Yeah.
[53:27] Parth Patil: So it's like you want to rant, if you want to give a lot of context as text, you can rant. So I know people, like I got friends that, um, you know, when they're like stuck in traffic in Los Angeles, right?
[53:37] Sarah Enix: They'll just ramble.
[53:38] Parth Patil: They'll just ramble to ChatGPT.
[53:40] Sarah Enix: Yeah.
[53:40] Parth Patil: And then-
[53:40] Sarah Enix: And honestly this is turning into a bit of a ramble, and I do love that. I do want to wrap it up and-
[53:45] Parth Patil: Yeah.
[53:45] Sarah Enix: ... ask you a couple last questions, um, so we can keep this under-
[53:49] Parth Patil: Sure.
[53:49] Sarah Enix: ... an hour. Um (laughs)-
[53:52] Parth Patil: Yeah. What should we cover that I haven't already?
[53:53] Sarah Enix: I know. Well, I know you've talked, you've covered what you're working on, which is agents.
[53:58] Parth Patil: Yeah.
[53:58] Sarah Enix: But how can we support you? What are we, what are you looking for? How can you, you know best market your skill set? And then maybe talk about some things that are exciting in your life, um, any new projects, anything like that.
[54:12] Parth Patil: Yeah. Okay. So I guess I'll, I'll, I'll see, say where I think agents are right now, and then that's like, that'll give people a sense of like what I'm hoping we figure out. Um, and if people in the audience, you guys have better ideas or tools that I haven't covered, I'd love to see them. So basically like agents are like chatbots. They're pretty good at planning. They can come up with a plan. If you give them the right context, they can actually execute parts of that plan. If you give them the right tools, they can figure out how to use those tools to attack your problem. And, you know, if you schedule them to work on the background, like you can actually have them running parts of your business or parts of your life on semi-auto mo-, like somewhat autonomously behind the scenes, which is awesome. That's like a lot of leverage, and you, I think people are underestimating that, that, like the actual value there.Then, where we're going next, where it is kind of like nebulous in two ways is, I don't think that the future is going to be chatbots on a website that you type to. I think, I think it might be like the entire user interface. If these chatbots can act, then theoretically they should be able to move anything on the screen, take any action on your phone. Anything that you use a mouse and keyboard for, some AI system of the future likely will be able to do it and you'll just have to talk to it, right? "Computer, do the thing," and then the computer starts doing the thing. Like, that's where we're going. So maybe like the user interface also is natural language. So like maybe w- the, the actual screens we interact with, like do they look more like video games?
[55:43] Sarah Enix: Mm-hmm.
[55:43] Parth Patil: Like, how do we get a more of a real-time interaction? If you have one useful chatbot, does it make... Like, are you gonna have many? And I think, I think we may actually have many of these assistants working for us behind the scenes. So then if, if you have many, like I don't want to have 100 tabs open. But like, you play video games, and you see like 200 units on the screen doing your job, like doing what you want. There's something about, like, units working on your behalf, and then maybe the chatbots of the future look more like video games because you want to get it more real time. Or maybe they, they can interrupt each other, or maybe they can proactively ... Like, maybe they text you in the morning, they're like, "Ah, I already checked 800 emails. You only need to look at three, these three."
[56:24] Sarah Enix: Mm-hmm.
[56:24] Parth Patil: Like, that is really interesting to me. And then can they collaborate with each other?
[56:28] Sarah Enix: Mm-hmm.
[56:28] Parth Patil: Like, do ... Can I build like three or four of them? Can they work together on a problem that's harder? So if, if people are like interested in, in tapping into this frontier, and it's, it's a frontier. Most of this is like largely fresh, unsolved green, uh, greenfield opportunities. So I think that's, like, where I'm interested in, and I think if people want to be, you know, if you want to catch up to where I am, a lot of my code is just out there. I'm putting more code out there. Yesterday I put two more code examples out on my Replit.
[56:57] Sarah Enix: On?
[56:57] Parth Patil: Replit.
[56:58] Sarah Enix: Okay.
[56:58] Parth Patil: It, so like, it's ... Replit's like a ... Yeah. If you look me up on Replit, you'll find a bunch of just basic code examples for how to get started building programs like this. And I think they're ... I'm trying to make them accessible at a high school level.
[57:10] Sarah Enix: Mm-hmm.
[57:11] Parth Patil: So like, I don't ... In the same way that the Rubik's Cube, if, if I'm gonna teach a six-year-old a Rubik's Cube, I'm gonna teach a high schooler how to build AI agents.
[57:18] Sarah Enix: Mm-hmm.
[57:19] Parth Patil: Because it's not ... There's no textbook, right? So we're in this, we have this opportunity to kind of just like explore the frontier of natural language together. And, um, and I think that's, that's really interesting. I think the job market can be whatever it is, but if you're out here studying something, like you might get a job from the future. You never know. Like, you might, you might ... The, the jobs of the future, I guarantee, are more interesting-
[57:42] Sarah Enix: Yeah.
[57:42] Parth Patil: ... and probably more satisfying than the jobs we have today. And I think if you start working with these systems, you can figure out ... You get a clearer picture of where the, the future is, um-
[57:51] Sarah Enix: And how to be a part of that.
[57:52] Parth Patil: Yeah, and how to be part of it. Yeah.
[57:54] Sarah Enix: Yeah. Okay, last question.
[57:56] Parth Patil: Mm-hmm.
[57:56] Sarah Enix: What is the mantra, vibe, sentence, quote, fill in the blank that you think about when the days get tough and you're doing the second all-nighter on your music video or you go outside and you're like, "Oh, wow, it's sunny"? What is that quote for you?
[58:13] Parth Patil: So I have a few of these, but-
[58:16] Sarah Enix: I'm not surprised.
[58:18] Parth Patil: You have to tell yourself crazy stuff if you want to believe, like if you, if you want to go down this path. I think, I think, uh, okay. I think if you think about that s- that phrase, like 1% every day, like if you can try to go for 1% improvement every day, that actually compounds a lot, right? So you try to learn a new concept, one new concept every single day. Learn a new topic, learn a new skill, a new technology. Because what happens is like you combine three or four skills and they give you a whole new superpower. Or like you start thinking about problems. Like, a year ago, I didn't know how to do any of this st- like, m- a lot of this stuff I didn't know how to do. And every week, I covered a couple new topics, and every week, I gained a ... U- that, that ends up adding up. In, on the day-to-day, you don't necessarily feel it. Sometimes you're just like putting time in, putting time in. You're just banking a bunch of time. But then a month later, you, like, your eyes open and you're like, "Oh, this part can be done like this. This is easy."
[59:18] Sarah Enix: Yeah.
[59:18] Parth Patil: And you start connecting dots. And that's really important. So I think it's like you, you just put in ... If you, especially if you want to work with AI, I think there's this, like, aspect of playing that's super important because, um ... And I think kids, it's like, it's like when you're playing or you're exploring a topic, that you're intrinsically motivated by it. This is a different, this is different from, like, my boss telling me I gotta learn AI.
[59:38] Sarah Enix: Yeah. Yeah, yeah, yeah.
[59:39] Parth Patil: That, that's a spiritually mindset.
[59:40] Sarah Enix: Where it's like you're naturally curious.
[59:41] Parth Patil: Yeah. There's, there's something that you're obsessed with that you're intrinsically motivated to do. And then you put AI on that problem. Then it's not like j- ... You don't nec- You know, don't put it on, don't throw it at work, because then that's for someone else, someone else is telling you how to apply it. Apply it to yourself where you are most motivated, and then you'll be ... Like, it won't feel like work, right? Two days of an all-nighter working on a music video, that doesn't, that's not, I'm not drained. I'm actually like ... My problem is I just have to sleep. Like, my problem is that I don't wanna sleep because I'm like obsessed with where I'm going, and then my body is like, "Hey, you gotta shut down." And I think that's where like, that's the opposite end of the spectrum. It's like when you attach it to your intrinsic motivation, you'll realize, you'll start feeling like there's not enough time for everything that you wish you could do. Like, I'm thinking about this music video, I'm like, "Man, I can't wait to plug this music video into some tool next year and then generate a video game." Because that, like if, if that's where we are, you know, you can generate images, videos, music, like why can't we just generate a whole video game, right? That, that, that's kind of where my mind starts going.
[01:00:45] Sarah Enix: Not a video game. An HBO series happens, and that HBO series happens, then a theme park is created, and now we're back to Rollercoaster Tycoon. (laughs)
[01:00:51] Parth Patil: Yeah. Oh, there we go, there we go. Yeah. Yeah.
[01:00:55] Sarah Enix: Digital storytelling.
[01:00:57] Parth Patil: Yeah.
[01:00:58] Sarah Enix: Parth-
[01:00:58] Parth Patil: Yeah. Kids should be playing with it. I think that's-
[01:01:00] Sarah Enix: (laughs) Exactly.
[01:01:02] Parth Patil: Kids should be playing with it. When would we do this for fun? Like, there, th- this is so much fun.
[01:01:05] Sarah Enix: I know.
[01:01:06] Parth Patil: And, yeah.
[01:01:07] Sarah Enix: I love it. I could, I could talk forever, but I will give, uh-
[01:01:11] Parth Patil: No, this was-
[01:01:11] Sarah Enix: ... the people their time to do what they want with. Um-
[01:01:15] Parth Patil: Yeah.
[01:01:15] Sarah Enix: Thanks so much for listening. Thanks so much for watching. Parth, you ca- Send over social media handles, LinkedIn, anything like that specific. @parthvader I think is a lot of your social handles?
[01:01:26] Parth Patil: Yeah. It's like I have a couple different han-. It's par- I probably should unify this, but if you just go to parth.club, that's my website. It links to a bunch of different projects. It also links to an AI that can link to m- more of my stuff. So like if you end up talking to Parth on ChatGPT, there's a GPT there, and it'll ... It can show you where my music video is and you can talk to it. It's not as funny as me, but it's, he's kind of like the butler. (laughs)
[01:01:50] Sarah Enix: It tries. It tries.
[01:01:52] Parth Patil: Yeah.
[01:01:53] Sarah Enix: Well, thank you for chatting with me, and, um, I'm really excited to, to watch your projects, watch and see how you evolve and change, and I can't wait to play the video game that you're gonna make. And yeah.
[01:02:04] Parth Patil: Yeah.
[01:02:04] Sarah Enix: Just thanks for chatting.
[01:02:06] Parth Patil: Thanks for having me. This was fun.
[01:02:07] Sarah Enix: (laughs) Alrighty. Talk to you later.
[01:02:09] Parth Patil: Thanks. Bye.
[01:02:09] Sarah Enix: Bye. And I'll just l- Oh, are you leaving? You left? (laughs) Oh my gosh. All right.