Building Reid AI: From Vibe Coding to Production Agents

Transcript

[00:00] Parth Patil: I realized a lot of stuff is a lot easier now. Now you can do in three hours what used to take three years. And that means that we have to be more ambitious, more creative, and more optimistic too. We have to make the future that we want to live in. And so, like, you have a chance to be a part of making that future.
[00:22] Kazuki Nakayashiki: Hi, everyone. Welcome back to another episode of Grasp Talk. Today, we are excited to have Parth Patel with us. Parth is an AI engineer and innovator, passionate about building conversational software and generative AI tools. He currently works with the Office of Reid Hoffman, leading AI agent development, LLMOps, and his special projects in the gen AI space. He also serves as a technical advisor at Blitzscaling Ventures and previously founded Creatia.AI, where he built advanced AI solutions such as chatbots, research automation systems, and generative coding assistance. As a Coursera instructor, Parth has taught thousands of people worldwide on AI and programming, creating hands-on courses to help developers harness cutting edge AI tools. Today, we will dive into his journey, his work shaping the future of AI agents, and his vision for making advanced AI tools more accessible to everyone. Thank you for joining us today.
[01:29] Parth Patil: Thanks. Thanks, Keisuke. Awesome.
[01:31] Kazuki Nakayashiki: Yeah.
[01:31] Parth Patil: Uh, thanks. Thanks, Kei. Great to be here.
[01:33] Kazuki Nakayashiki: Thanks. Yeah. Thanks so much. So, now you are AI engineer at the Office of Reid Hoffman. I remember in the previous video you said, you know, your, your official title is AI wizard at the company.
[01:45] Parth Patil: Yeah. Yeah.
[01:46] Kazuki Nakayashiki: But, yeah, so, but could you tell us what you do at the Office of Reid Hoffman and are there any AI projects you are currently working on?
[01:54] Parth Patil: Yeah. I mean, so many. Um, I guess I'll give a little bit of context. Uh, so when, uh, my last, like, normal job was, uh, like two years ago when I, I worked at a startup called Clubhouse, where we were doing audio conversations around the... Like, we were, we had an app where we were, uh, we created aud- audio conversations on the internet, kind of like Discord, but a little bit less gamer, more f- more for everyone. And it was very popular during the pandemic. It was one of the fastest growing apps of all time. And I was working there as one of the first data scientists. So my job was to figure out, like, why do people even use this? What is working? What's not working? How can we try to learn from that and get better at it? And while we were working on Clubhouse, November of 2022, ChatGPT came out, and it became the most popular topic across every single language all over the world. Everyone was talking about it. And I was talking to people on the app, and I was just like, "Oh, how are you using it?" And you just get so many different use cases, and, like, there were farmers from India that were using it for crop cycle planning. There's just, like, musicians using it to study music theory. There's all these different applications of general intelligence. And for me it was like, "Oh, my God, this is like a new computer." Like, this is, like, a hundred years early. Like, I didn't imagine we would see this in our lifetime that, like, the computer could speak every language and also write every programming language. And, um, and this was GPT-3.5, right? So initial ChatGPT launch. And then in March of 20- March 14th, 2023, GPT-4 came out. And one of my friends on the app... We, we used to be programmers. We were just programming, you know, chatbots, trying to figure out how to make the language model useful. And one of my friends, he was a, he programs every single day and just makes projects and then teaches people. His name is Echo Hive. He's on YouTube. He has a great successful YouTube channel now, and a Patreon. But, uh, he was saying, "Parth, you gotta be using GPT-4, because this is better than many of the engineers that we know. And you can just ask it to teach you how to program and you can make more interesting things." And in my career I had been avoiding programming because I was more of a data analyst, data scientist. So I was using SQL, I'd used Python, but not a full stack engineer at the time. And then, um... So then I said, "Okay, I'll give it a shot." So I sit down with GPT-4 on a Sunday and I say, "Uh, teach me how to clone my voice," and then it just writes the program, right? It writes the program. It uses an open source library, Tortoise Text-to-Speech. And then, uh, and then one hour later, I was, I was like, "Oh, teach me how to run this program." Then one hour later, it's on my computer. I'm talking to... I have a program that can, that sounds like me. And then I was like, "What? This is only one hour?" Like, this... I thought this was supposed to take forever (laughs). Like, this is supposed to take a team of people a few months. And then I was like, "Okay, teach me how to build a GPT powered chatbot." And then it wrote the first version of that program. Like, you know, it's a, it's like 25 lines of code, just text-based script, runs on your computer. And then I hooked it up together and I was talking to a program that had my voice, right? And I was like, it was like a voice assistant. And, and that was my first day of programming with language models. And for me it was like, oh my god, this, this is... now I need to become a programmer. Now I need to learn, because it's never been easier. Um, you have an expert. The mo- the, the model will teach you these things, and then you can just, you have to still, you know, put it all together. But whatever you have in your mind might be possible, and it might be inside the model, and then you just have to be programming with the model and then you'll see what's possible in concert with, with the, the model. And so I was working on that, and at the same time, Yohe, Yohe Nakajima puts out Baby AGI. And it goes viral, like mega viral. And I look at the code and I was like, "Oh my god, this is like, not... this is very straightforward." Like, it's a fairly simple program, but it's like... but it was so popular because I think at the time it was like the first, like, the promise of the idea that you have a program that can just start doing meaningful work for you and, and just, like, operates in a loop towards a, a goal that you give it, was a very alluring problem. The, like, the, the idea of this agent, right? The LLM is good-Generating text is great, but can it actually be useful and do things? And Baby AGI was this, like, spark of imagination that, like, went through the community of, like, what if we made these tools, um, you know, act on our behalf in a useful way? So, I looked at his code and I forked it, and I started working on this, like, these, these agent loops. And my friend who had introduced me, he was doing the same thing, so we were doing coding automation, like, "Oh, go build 100 ide- uh, come up with a list of 100 ideas. Go try to make those." And then I go, and I'd sit by the pool, and then the program will just be writing code. And then I come back, and then I go grab the code, and I go to the pool, and I'm like, "Wow, I didn't know you could do that. This doesn't even work, but that's interesting." And so I was like, "Oh, my God, Python." Like, that was my main programming language, but then I saw, because of the language model, just how powerful Python was, and then I was like, "Oh, wow. Like, language model is useful, but now when you connect it to the ability to code, you are able to reach into, you know, every single programming language." So data science, you can do, you can do analytics through language models if you let them write SQL queries and write Python code. Um, you can automate a lot of stuff like Excel, pre- presentations. Um, and so that was, like, interesting to see how, how much automation you could get from, from language models. And as I was doing... Then we had layoffs at the company, and I got laid off. And I was... On my first day of layoffs, I was like, "Oh, this is good. Now I can just do this all day long until I run out of money."
[07:30] Kazuki Nakayashiki: (laughs)
[07:30] Parth Patil: "And I'm just gonna do this until I run out of money, and then I'll figure out what to do after that." Like, I was willing to... And then, and basically I was like, my entire savings, I just put it all into OpenAI API calls (laughs) every single day. And, uh, my friends were like, "Wait, don't you want to get a job?" Like, I said, "It's tricky, because this, I think, is more interesting than getting a job." And a lot of the companies, a lot of the people I was talking to just didn't understand it. They were, like, afraid, and they were banning GPT, and I was kind of like, "I can't work with people who don't see the value of this kind of technology. And I would rather just study this independently myself and just make things for fun than work for people that don't let me use this technology." So, that was, uh... So I did that for a b- and I had a, I had four months of severance, so I was like, "Okay, I have a little bit of buffer room here where I can just focus without having to recruit." And then after the severance was over... Because I was like, "If I just sit here 14 hours a day, Saturday and Sunday, and I program with language models, I'll figure something out." Like, probably. And, and then worst case, even if I run out of money, like, then I'll just go get a job after, like... And maybe use the skill set, right? Like, I might learn something, and, and then use the skill set. And then at the end of my severance, I was like, "Oh, I'm not done. I'm gonna keep doing this." So then I just kept, I just kept burning my own money, um, building tools on top of language models. Chatbots, retrieval systems, data analysis tools, data visualization. Every single capability that I was, like, interested in. Even just, like, music. Like, I spent one month talking to language models about music and music videos and music theory and then music production, and I was like, "Wow, this is so useful outside of work." Like, this is just, like, the biggest... It's general knowledge, right? It's not just for work. It's actually just very generally applicable, um, super smart assistant. And then, uh, and so I was like, "Okay, I'll just keep figuring these things out." And, uh, then I started getting contracting work. I started getting consulting work for pro- th- related to this skill set. They were like, "Oh, could you build a chatbot that retrieves context from a bunch of podcasts? Could you, uh..." You know, like, like, "What kind of things can we automate now in the business?" And because I had worked in startups my whole career, and because I was a data analyst, I was very like, "Okay, this is, this is at least a huge amplification of analysis." Like, that's obvious, because you can automate SQL queries, right? Natural language. You can say, "Who's our largest customer?" Well, now the language model can write the SQL query. I don't have to write the SQL query. So, now I can ask the next question. I can ask the next question. Th- the model writes the SQL query. Now we have, we have high speed data analysis now. And that's just, like, one part. Then it's like, you can also use structured outputs. You can figure out more. You can, like, create... All the, the messy data can now be structured using language models. So, I was like, "Okay, this is probably a skill set that is useful." Even though I was like... So I started getting contracting work, and then so I was like, "Okay, cool." Because once you get that first, once you get that first job, like, the first contract, you're like, "Okay, there is a market demand for this skill set." So then yeah, I was like, "Okay, I'm gonna be safe. I'll be, I'll be, I'll be fine.
[10:42] Kazuki Nakayashiki: Yeah.
[10:42] Parth Patil: And then, uh... So that, that was nice, and then I kind of did that. I was... And then I, I did the Coursera. Coursera reached out, and I created some courses on code generation and data analysis using language models. And that was great, because I was just like, "Wow." Like, um, I mean, being recognized, and then our c- my course was, uh, showcased at Davos, and I was like, "Okay, cool. They are taking seriously that we need to update our curriculum." And even though I'm not, like, a full stack engineer, I do think that, like, the ability to generate code is an important message to send. It's not cheating. It's actually the future of how we build software, right? So, this was vibe coding. We were vibe coding before vibe coding was coined, right? Vibe coding became a thing this year. But we were vibe coding with GPT-3.5 when you have to copy-paste from ChatGPT. Copy-paste the error back into ChatGPT. It's like, "Oh, fix this. Oh, fix this. Okay, why does this work?" And so that was a, that was a, an interesting era. And so I, I, I love code generation. That's my, one of my favorite things about language models. And then, um, then I... In, about two years ago, I met, um, I, I was introduced by a mutual friend to Reid Hoffman. And, uh, we were working on a project called Reid.AI. So, we built, like, an AI to represent Reid Hoffman. Reid Hoffman, the co-founder of LinkedIn and one of the earliest investors in OpenAI. And so we met, we talked for four hours, and then the next day, um, he, he was like, "Oh, you should come work for me." And I was like, "Yeah, let's do it."And so, uh, I've been working with him. And so, the kind of projects, it's a lot of the same stuff, but now I'm not alone, right? And I have, like, people I can s- I can bounce the ideas off of. I have teammates, and I'm like, "Oh, what if we wanted to," like, we're, w- we're, you know, a lot of different stuff like translation, right? Translation's very useful, but now you can translate a speech into every single language. You can translate a podcast into every single language, and then you can use avatars to get the lip sync right, right? So, um, there's a lot of stuff that I was like, the ke- the tools that I was playing with now became useful. And then, um, now I have people I can, like, build the tools for, right? So, building internal tools, vibe coding, using Replit with a teams plan is great, because now I can make tools for my f- for my teammates, and then they give me feedback, and then I'm like, "Okay, cool. Now we can..." We, i- it's, it's, uh, and it's like lower risk than starting a company, like starting a startup, because you have, like, uh, you're building internal tools for an organization that already exists, and people are, like, uh, embracing AI. So, you get a very good feedback loop and com- sense of community and teamwork when you, when you make things. So, that's great. Yeah.
[13:13] Kazuki Nakayashiki: It's really fascinating.
[13:15] Parth Patil: Yeah.
[13:15] Kazuki Nakayashiki: And th- the end goal of, like, AI projects in, uh, the office of REED Hong Kong is for internal use? Or help you or your team thought about launching or releasing it to other external organizations and sell it to them? Or...
[13:30] Parth Patil: That's a, that's a great question. I think we're not, we're not opposed to it. I think we kind of just view it right now as, like, let's, let's just learn and experiment. Let's see what, you know... Some stuff does end up more publicly facing, but I think, um, a- and, like, REED AI, for example, is like, it's an internal tool, but we are exploring what would it take, like, what would it be, what would it take? What do we need to build to make it something that we feel comfortable putting in your pocket, right? So, releasing more widely as like a, maybe a companion experience. Um, so we're open to it. I think it's mostly like, it's easier to move fast and experiment when you keep things internal. I think this is the thing about vibe coding I, I recommend. Because a lot of people are now getting into code generation and vibe coding. And, and I always say like, "Okay, you can get excited, it's good, but like, start with things that are like, safer and manageable." And if you know everyone who's using what you're making, then you can, you can solve the problems before someone breaks it. And like, you know, you don't put it, you don't wanna, you don't wanna put it out there, have other people hack it, and then it starts becoming a nightmare to deal with. But if you're building internal tools, you have a higher bar, and you can kind of like internally break things, and that's fine, because you're, you know who's using it, and you get that feedback. So, I think that even when you're building code generation, like using code generation, there's this like lower risk thing, which is build things for your organization, and then take those learnings, and then you make the next, like maybe your 15th idea, you might be like, "Okay, let's make this for people, right? Make this more externally facing." And, and it might not even be an application that reaches the other people. It might be like the output of gener- like it might be generative media, right? It might be a video. It might be, yeah, it might be a video, it might be like a speech translated into 25 languages. So, the output of the model can still be, be made externally, um, available. And I think that you learn, you, you want to learn things, and you don't necessarily have to build and launch an app to learn things. You can also learn things by just making, making tools for, like your first customers are just like your own teammates and you, right? And then, uh, and then if it's useful, you might find that there are other teams that you can, you can partner with. So, I, I do like branch out into the broader network and talk to some of the, the firms in our network, and it's like, "Here's how we're using the tools." Um, and then I, and then, uh, my big recommendation is everyone should have like a vibe coder in house. (laughs) You know, like-
[15:50] Kazuki Nakayashiki: Yeah.
[15:50] Parth Patil: ... because we're kind of getting to a place, I think, that it is easier and faster to buy s- to build software, than to buy it. And I met Amjad in, uh, uh, from Replit in, uh, in San Francisco, and I mentioned this to him. I said, "I have a feeling that buy versus build has flipped." It, it used to be traditionally is like, if it's not core to your business, just buy the software and then move on. But now it's like, you know, you can tell AI to make the version of the, the application that you need, and you don't get all the extra features you don't use, and you're not paying for a two year subscription to some overly, like, one size fits all solution. Now, you can just make custom software. And maybe that's just like my, my bias, because I love generating code. But I love this, this world we're getting into, where I can just sit down with Claude code or Cursor and just be like, "Hey, let's make this." And then two hours later, I have it, and I don't have to talk to a sales rep. I don't have to negotiate a two year contract. This is something that's very interesting, and I think it's going to be very disruptive, the, the like personal software, uh, uh, era, you know?
[16:59] Kazuki Nakayashiki: Yeah. I was thinking actually the same. I mean, people will stop buying software outside, because they can internally, you know, vibe code to-
[17:07] Parth Patil: Yeah.
[17:07] Kazuki Nakayashiki: And also, if you hire a vibe coder, and it's much cheaper, like, than paying, uh, millions of bucks every year to external software, yeah.
[17:15] Parth Patil: Exactly.
[17:15] Kazuki Nakayashiki: Yeah.
[17:16] Parth Patil: Exactly. It's very disruptive. And there are some, I think... The other thing, there's like lasting effects. I think there are some businesses from the old world, where you look at it, and you're like, "I have a feeling that AI could build that." (laughs) You know? Like, there are some gigantic, uh, you know, like built businesses that have kind of like relied, now they have the network effect. But you wonder, it's like, would I buy that if I knew that GPT-5 could build it in three prompts? Probably not, you know? And, and that's an interesting, like, they're gonna have to adapt. And then, we have our own, you know, new options of, uh, ways to play the game. And it's a lot cheaper, right? You can move faster, it's cheaper. There's definitely downsides, right? Security vulnerabilities, um, you know, people complain that the applications that you vibe code don't have, um, they're not, they don't scale, th- people complain. I say, "Look...Everything breaks once too many people start using it. But that's, that's a good problem. You solve that problem, right? Like, like, when, when m- millions of people showed up at Clubhouse, everything was on fire and stopped working. But now, it's you have to try to solve that problem. You don't solve that problem when you have no one, right? And when you have no users, you're not trying to solve this, like, scale problem. You're trying to get it working and try to get something that people even want in the first place. So, I think that, and also, the models keep getting better, and I think that that's the other-
[18:33] Kazuki Nakayashiki: And cheaper, right?
[18:33] Parth Patil: Yeah. Cheaper, so you can run more calls. Uh, the agent wrappers are getting very good. And then, you know, even, even on the cybersecurity side of things, I think there's another way to think about it, is just, yeah, maybe, like, vibe-coded apps may be vulnerable, but then you can also imagine a world where you have an agent and it's like, "Hack, just hack my app and then tell me, and then my coding agent will patch it." And so, kind of, we have that in, in human world, right? White hat hackers. And instead of ransoming your data back to you, they just tell you, and then you pay them, you pay them to hack you, and then, and then you patch the vulnerability, and then that becomes ... That's a good kind of hack, right? And so I think we're gonna have AI that does that. I think Replit is starting to do that. They have ... Before you deploy an application on Replit, you can click run a security scan on my app. Um, last week, Claude code added the same feature, which is like, let's explore my application for possible vulnerabilities, and then let's go patch them. The, I think we're still gonna see, like, people make mistakes. I think there's gonna be a lot of, uh, entry level vibe coders that don't know, right? Frankly, don't know better, because they didn't go to an engineering ... Like, they don't have an engineering college experience. I, I didn't, right? So, um, so we're gonna see these errors. Um, but hope- I think AI is a huge part of the solution. And so we just need the models to get better. We're gonna get better at, um, using them, you know?
[19:56] Kazuki Nakayashiki: Yeah. Definitely. Yes. And I don't know if you d- if you involved in, like, building Dide.ai at time, but-
[20:04] Parth Patil: Yeah.
[20:04] Kazuki Nakayashiki: ... I mean, I remember, I just remember the last year, last April, the Dide.ai and Reid Hoffman, you know-
[20:10] Parth Patil: Yeah.
[20:10] Kazuki Nakayashiki: ... launched the conversation.
[20:11] Parth Patil: The, the video we put.
[20:12] Kazuki Nakayashiki: Yeah.
[20:12] Parth Patil: Yeah.
[20:12] Kazuki Nakayashiki: That was really amazing. But did you, did you work with the team?
[20:17] Parth Patil: That was me. Yeah. Yeah.
[20:18] Kazuki Nakayashiki: Oh, cool.
[20:18] Parth Patil: That was me.
[20:19] Kazuki Nakayashiki: Yeah. ？？？
[20:19] Parth Patil: That was actually the, the project that led me to working with Reid, which was, uh, one of, uh, uh ... M- my teammate on the project. I w- I was just working out of a beach house in Los Angeles. Um, and m- one of my friends at the house, Ben, he works for Reid, and he was like, "Parith, do you think you could build a chatbot on top of, like, a guy who has 30 years of podcasts and books and everything?" And I was like, "Yeah, I think so. Give me..." Because I think this was ... Oh, yeah. This was custom GPTs, right? From OpenAI. Remember custom GPTs?
[20:50] Kazuki Nakayashiki: Mm-hmm.
[20:50] Parth Patil: So, you could take ChatGPT, and then you could give it a, a separate personality, and then, um, upload a bunch of files and then say ... So, we were like, I was like, "Okay, yeah. Give me a couple hours." So, in three hours, I went, I got his books, I got a bunch of podcast transcripts, and I uploaded them into a custom GPT. I said, "You are ... Pretend you are Reid. You are Reid AI. Uh, you know, digital avatar rep- de- you know, trained to represent the, the body of work of Reid Hoffman." And so there's a little bit of prompt engineering and RAG, uh, and custom GPT. So, this is no code, right? So, this is, like, a very simple, it was like proof of concept. And, um, and then we built that, and then I showed it to him, and then he showed it to his teammates, and he was like ... And then, and then, uh, that was when he was like, "Oh, yeah. We should ... W- like, we, like, you should come meet Reid." And that was the beginning, which was like a very simple prototype kind of proof of concept. But then when, when I met Reid, it was like, "No, I build a lot of systems like this," actually more, you know, on the, with Python, like, closer to the baby AGI, more agentic, not just like retrieval systems, but, like, tools that can act. And so it was interesting, because we both have this, like, theory that, like, these agents are just very useful. Even if they're not, like, uh, fully autonomously end-to-end, they're still very useful copilots. And I think the next 10 years, we're gonna see an explosion of these copilot-type systems for everything, right?
[22:12] Kazuki Nakayashiki: Yes.
[22:12] Parth Patil: And, uh, you know, all business, your medical, like, i- i- it might be ChatGPT, it might be a specialized application. You know, there's gonna be therapy. There are, eh, there's an AI agent for almost every single possible use case. And we, we kind of were like, "Okay, like, Reid AI is cool, but, like, what if we, like, we have this chatbot, but w- what if we, we use one of our portfolio companies." Uh, at the time, it was called Hour One, and they, they were doing avatars. And so we were like, "Well, what if we had ..." And I was, uh, I don't know. Have you guys seen the movie Tron?
[22:46] Kazuki Nakayashiki: Yup.
[22:46] Parth Patil: You know Tron? Like, 19, uh, I think 1982?
[22:49] Kazuki Nakayashiki: Yeah.
[22:49] Parth Patil: Uh, Tron, Tron is this movie where, uh, there's the old Tron and then there's Tron, uh, 2012, they made a r- uh, Tron 2, and then this year, end of the year, there's a third Tron movie, Tron Ares coming out. But Tron is one of my favorite movies. And in Tron, Jeff Bridges is the main character, and he gets stuck inside of a video game, and he's talking to programs, and he's in this, like, digital matrix, and ... And my dad watched it when my dad was in engineering college, and he was like, "We didn't know what the computer looked like. This was awesome." And then, uh, uh, and then he showed us the movie, me and my brother, and then when we were working on Reid.AI, I was like, "Oh, you know," we were thinking, "How do we show this? Like, how do we demonstrate this chatbot to the world," right? Like, w- if we wanted to, like, showcase this kind of experience. And I had the idea, I was like, "What if you could just talk to yourself," right? Like, kind of like a mirror, you know, you have a copy of you, and you can talk to it. And so the first video we put out, the one that went viral that, I think that in April, was, um, Reid talking to his digital clone. And so under the hood was the custom GPT that was powering the voice and, uh, powering the, the, the text generation, right? So, 20 years of his knowledge and retrieval. And then you have the video avatar, which was by Hour One. So, it was, it was, uh, that was the, like, we scanned Reid and then now you have this avatar. And then the voice was Elevenlabs.And so, we were able to, like, you know, clone his voice, clone, clone his image, and then we have this, like, chatbot that pretends to be his mind. And then he, he... we had this video edited together of them talking to each other. The interesting thing is, at the time, there was no real-time avatar technology that was this good. So, it was all pre-generated and edited together, with pre-generated, uh, footage. But people were so excited by that, that companies, startups started coming to us, and they were like, "We would love for you to try our real-time avatar technology." And so, like, then, then we started... you know, when you build something and you put it out there, uh, especially if you're early, right? So Yohei, right? Yohei built Baby AGI, and now that went viral all around the world. And, um, and it inspired so many people, me included, right? So, you see something is possible and you're like, "Whoa. I didn't know that was possible." Like, "I wanna, I wanna try this." So, this, the builders also realize that. The startups see that, and they're like, "Oh, they're working on this. We're working on this. It would be cool if we teamed up." So, a lot of startups come to us, and we're like... and they're like, "Oh, we're building an avatar technology. We'd love for you to try it out." So, that's how we, we kind of like used... Read AI is, like, kind of this magnet for people who are interested in avatars, people who are interested in agents, people who are interested in the concept of digital twins. And for me, it was just like, I like Tron, I like science fiction, and it'd be very cool for us to, like, showcase this in a way that, like, reminds us of the movies and the video games, but actually is possible today with the technology that we have. And it's off-the-shelf technology. Like, it's not a, like, it's not, it's not, uh, you know... The original Read AI could, could be made with, like, consumer grade applications, state... you know, like off the shelf s- regular subscriptions, right? So, it was like, here's what we can do with ChatGPT, Hour One, and Eleven Labs. Now... and then, we've also done more advanced avatars, where we actually go in and, and we partner with, um, more, you know, like, heavier grade avatar technology. But it's, it's really more like, uh, a bunch of related projects that are just like, "Oh, let's do another experiment. Oh, let's try this, Younger Read, interview..." like, he interviews his younger self. And so, we use Hedra, or we do more of a real-time thing with, uh, with HeyGen. We're using HeyGen's real-time avatar right now. And that's a lot of fun, because, um, you know, every time we... every month or every two months, like, some startup is like, "Oh, what if we teamed up and we tried this new thing?" And there's so many more capabilities of AI agents that we have yet to even tap into, like Vision, for example, or, like, doing more of a hologram, life-sized... We've done a life-sized hologram of him. Um, so there's a lot of... I think it's mostly, for me, it's like, it's like, how do you show people what agents are? Because, you know, agents are very, like... What are they? They're, like, invisible employees that just kind of, like, work? No, no, no. Th- they can be, also be characters, right? And so, um, we just like to u- we like to use Read AI as a way to demonstrate the capabilities of language models in a way that is more in- inter- interactive and intuitive, and maybe, like, uh, gets people inspired, yeah
[27:20] Kazuki Nakayashiki: Yeah. Yeah. Definitely many startups are inspired by Read AI, for sure. And, you know, yeah. Collaborate and come up with any, uh, new ideas. But, have you thought about-
[27:30] Parth Patil: It's, it's... One thing I never expected was that, like, how m- I never expected it to be this popular. Like, um, and, and, like, it, it... you know, it's been on the news, it's been on many podcasts, and now we have real time. So, it actually talks to people in real time when we, we showcase it. Um, it's gone on a speaking tour, and I get in- I get, I get dee-... uh, I get messages on LinkedIn all the time. They're like, "Hey, uh, it would be awesome if Read AI could come and, uh, judge our hackathon." And I'm like-
[27:59] Kazuki Nakayashiki: Yeah.
[28:01] Parth Patil: ... "I, I can judge your hackathon." Like, what? (laughs) Am I the agent?
[28:04] Kazuki Nakayashiki: Yes.
[28:04] Parth Patil: What happened? (laughs)
[28:05] Kazuki Nakayashiki: (laughs)
[28:06] Parth Patil: So like, th- he gets a lot of invites to interesting events, and I'm like, "Well, this is... What happened?" Like, I thought, I thought I built an agent, but I feel like the agent now. (laughs) But, it's fun.
[28:15] Kazuki Nakayashiki: Yeah. Yeah. By the way, we are talking to the other you, right? Not Past AI?
[28:20] Parth Patil: Yeah. Oh, yeah.
[28:21] Kazuki Nakayashiki: Not... Oh, okay. (laughs)
[28:22] Parth Patil: Past AI. Yeah.
[28:23] Kazuki Nakayashiki: Sorry. (laughs)
[28:23] Parth Patil: This is the year, this is the year I'm gonna clone myself. I'm gonna do something-
[28:26] Kazuki Nakayashiki: Oh.
[28:26] Parth Patil: ... uh, similar to that. I'm working on it. I think it's a, it's a lot of fun. And my, my only theory... you know, I kn- I did it... w- we did Read AI for fun, and I was... because I was like, "Science fiction, this is cool." But now I'm like, "Wait a minute." Someone once asked me, she was like, "Oh, can I hire you?" And I was like, "I have a job." And then she's like, "Well, you cloned Read. Like, what if you clone yourself? I'll pay to talk to your clone." And then I was like, "Wait a minute. If my clone can pay the bills, then I have to find out..." Like, that, that would be amazing, right? Like, imagine, I can just go to the beach and then play video games and, like, have fun. And then the clone is just here doing all the work and thinking really hard. Um, so, I'm gonna try it. Maybe it'll work, maybe it won't. But, um, it'll be fun. Yeah.
[29:08] Kazuki Nakayashiki: But what if your clone go to beach and start gaming?
[29:12] Parth Patil: (laughs)
[29:12] Kazuki Nakayashiki: But you... So you have to do work, right? (laughs)
[29:15] Parth Patil: It might, actually. I was, uh, I was talking to the... so, I'm working on, like, its knowledge base right now. And so, we really... I have Claude code, and I talk to Claude code, and then it uses a GraphRAG system under the hood. And I say, "You ask me questions, and then I'll talk, and then you construct my knowledge graph." And, uh, and it was like, "Oh, what games do you like?" And I was like, "Hang on." I go to my Steam, I copy-pasted all of my play history, every single hour, every single game, and I pasted in, "These are the games that I play." (laughs) And now it's like, "Oh, I understand." (laughs).
[29:45] Kazuki Nakayashiki: (laughs)
[29:45] Parth Patil: So, it gets, it gets a sense of my taste. And when it asks me questions, it's very... it's much more like... 'cause it uses knowledge graphs to make the question much better. 'Cause it's like, "Oh, you like, you like Age of Empires, you like StarCraft. Like, tell me more about that." And then I have this conversation with this AI, and it's kind of like downloading some p- something like my philosophy, right? I hope, I think. And I don't think it's me. I don't think you can actually... like, the more I offload to this, the more I'm gonna have fun. And the more it can just wear the suit and be this, like, machine, I think. It'll be fun. We'll see.
[30:20] Kazuki Nakayashiki: And I'm always curious, like, in a creating AI clone like Reid Hoffman is very diff- you know, o- obvious, like, you know, has social value, also business value. But so for people not like, you know, just ordinary people-
[30:35] Parth Patil: Yeah.
[30:35] Kazuki Nakayashiki: ... so what is the, like, social and the business value and their use case in the future?
[30:43] Parth Patil: You know, it's funny. Peop- a lot of people ask me, like, "What's the business? Like, what's the business?" And I'm like, "We were doing this because we are just having fun and exploring what's possible."
[30:53] Kazuki Nakayashiki: Yeah.
[30:53] Parth Patil: Now I'm starting to suspect there wi- there is a business. It's not validated yet. And actually, I don't think necc- it's not clear necessarily that the video avatar is necessary. Um, but I think that there is, uh... Like Yohei, right? Oh, okay, it's like, Yohei. You have Yohei on, and he has his unique perspective from his career and his, uh, his, his programming, his like... Ev- everything that he does that's not inside GPT-5. Now, some people will say GPT-5 is just more important than his perspective. I would ag- I would disagree. I think that, like, especially with the knowledge cutoff of the language models, they have this, like, they're kind of frozen in time a couple of months behind humanity. And so people like us that are just making and living in the real world, we are learning and, and we have informa- we do have insights from the real world that the models do not have, and that's your personal knowledge. I think that your personal knowledge is actually valuable, and, um, I also think that it's not neces- even if it's not necessarily clearly valuable to other people, which I think it is. Like, I think if the, if my friends could tap into my knowledge without me necessarily being there, I would love to offer that to them, right? Like, people are always asking me, "Oh, wh- w-" They're like... Nowadays it's like, "Parth, I'm having a hard time figuring out why GPT-5 is better than GPT-4." Now, I can answer that question, and I do that, right? Like, I'll a- like, my best friends, I'll answer that question. But it, I get that question a lot, and I wonder, like... Um, and I don't have a blog. I mean, a blog is not fully fleshed out, but it would be nice if, like, they could get that answer without me having to be there, and if, if it could be, be at my level or even better than my level, that would be nice. And I think also for me, my long-term memory is not as good as I wish it was. But, you know, you can, you can imagine your AI clone has perfect retrieval across 20 years of work, right? Th- this is not hard to, hard to en- engineer. So, you're kind of like, okay, there... I think there is value even for yourself. Um, I think the, the most important value is for yourself. Like, I talk to a, I talk to my own knowledge, right? I put a conversational assistant on top of my knowledge graph, and then I talk to it, and I feel like I'm bouncing ideas off of, um, like, like a ghost, in my, in my, like, like a version of me, like a spirit, right? And I think eventually, maybe, this is just a joke, but probably is gonna happen. Imagine, like, 100, 200 years from now. My great-great-grandkids are, like, about to make, about to make a very stupid decision. And then my digital twin, my ghost, comes back and is like, "Do not bring dishonor to the family." (laughs (laughs)
[33:40] Parth Patil: It's like, it's like in Mulan, right? They sent... The ancestor spirits send Mushu to protect Mulan. I would like to do that if my, you know, if my great-great-grandkids need help. They, it's like, "Oh, why don't you talk to your, you know, your, the clone of your great-great-grandfather? He might have some, uh, he might have some perspectives that people don't have," right? So not everyone is... Like, we're all different. And actually, I think that's why it's valuable. Like, if you're different from the rest of the people in the language model, you're definitely different from GPT-5. Um, you have to... I think that's valuable. I think that's valuable. Like, you have a perspective, right? You have a unique life. That's valuable. That's, by itself, is valuable. The question is, do we make that perspective available to people? And I, I would like to make that available to my friends, for sure, and maybe to strangers. But most likely, at least my friends should be able to access my, my perspective, even if I'm not there or maybe, like, maybe one day I get injured, and I'm unable to, like, recall some of this. There's so many... I think that the, it is valuable. I think right now it's weird. But I think in the future, it will be normal, actually.
[34:46] Kazuki Nakayashiki: Yeah, definitely. Yeah. That's (laughs) exciting future and, and, and use case. So, but when you were building Dide.ai, I mean, this is, like, a benchmark. Is... Having benchmark is, uh, like a, like the, how to say? The always... I know a lot of startups are struggling with having benchmark and what metrics they should follow. So, did you have any metrics or numbers you-
[35:11] Parth Patil: Like evaluations.
[35:12] Kazuki Nakayashiki: ... evaluation process-
[35:12] Parth Patil: Yeah, yeah.
[35:13] Kazuki Nakayashiki: ... to say, "Oh, this is really similar." You know, the response is, maybe it would say or, you know, similar to this.
[35:20] Parth Patil: Yeah. So, this is, this is a continuous kind of thing. Um, first of all, I think you're right. Every... Okay, you cannot have an AI system, uh, you cannot reliably deploy it for anything serious unless you are evaluating it. Like, unless you are giving it a grade on the things you want it to be good at, you can't, like, you cannot improve what you don't measure. That's the age-old quote. "If you're, if you're not measuring it, impossible to improve it." So, now we have to measure conversational, like, accuracy. Um, but that, what does that mean, right? So, um, sometimes it's like retrieval. Okay, you ask a question, and it is not in the context window of the model. It's in the knowledge base. So, did the AI even retrieve the fact? Okay, that's one eval. Then the second eval is like, did it interpret the fact well enough to be, you know, a- as we would ex- as, as we would expect Reid to do so? And on one hand, you have, like, all the que- so it's like, how do you create these, these datasets? I think, one, it's like you have all these podcasts. You have questions people are asking already to the real Reid.And then you have, you ask those same questions to the AI Reid, and then you see, okay, it's like 60% there. Okay, w- well, now we need to know, we need to improve prompt engineering. Or actually, it's not even in the knowledge base, we need to make the knowledge base more richer. And this is a, we're in the very beginning of this. Like, this is, this is when I say, like, internal tool is a good place to start, because you have to build these evaluation criteria before you roll it out widely. And, um, and also, like, there's a lot of human testing. So, there, my, me and my teammates, we literally talked to this, and we were like, "Okay." Then there's evaluating the voice, right? We spent a lot of time just getting the voice to be like him, and then you do, like, then all of a sudden, you try the voice in Japanese, and you're like, "Wait a minute. Now, we gotta start all over." Like, you, you, it's like, we want it to be good in different languages. We want it to be good. Uh, we want it to sound like him, so then we have to, like, make sure we have the right raw data that we're using to train the clone. And then we want it to say things that are in, in the vein of what he would, he might say. And then also, and this is character design. I think of this as, this entire thing as character design, because, um... And, and I think it's up to the creator, right? In, in ReidAI's case, you know, Reid, uh, you know, Reid is like, it shouldn't say, "I'm Reid Hoffman." It should just say, "I'm ReidAI." It should be transparent, upfront. It's not like trying to deceive you, right? It's like, this is a digital twin of Reid, and it's not going to say, "I'm Reid Hoffman." Or if it does, then I have to go and fix something, because like, that's an evaluation criteria, right? So, when I say... One of the questions that we ask it 1,000 times is like, "What are you?" And then it's like, you gr- y- it's like 99% of the time, it says, "I am ReidAI." M- Good. Like, that (laughs) you're not gonna pretend to be Reid. Uh, but... So, we have a bank, and it keeps growing, the questions you're asking. And then we also, what, what I started doing is I have, like, other, I mean, other LLMs are talking to him, and those other LLMs are like personalities, and so then they generate questions that that personality might ask. So, it's like, "Okay, what might a Fortune 500 CEO, um, ask Reid?" And then generate 15 questions, and then you, you plug that into the evaluation criteria. You see the outputs, and then you also have LLMs, you know, judging that, and then you also have humans judging that. And we're in the very early phases. But thanks to tools like Cloud Code and, um, GPT-5, you can basically be like, "Okay, let's build an evaluation suite. Like, this is the criteria we wanna m- uh, to, to measure. These are the types of things." And then it just starts building this, like, tool for you to inspect the quality of the, of the system. And I think that, like, you have to build these evaluation things. Otherwise, you won't be confident. You can't... And if you're not confident, you're not gonna put it in any serious application. And, uh, but it, it's getting easier to do, like, build some of these, like, systems that allow us to build more trust, definitely
[39:20] Kazuki Nakayashiki: Yeah. So, do you remember the very first response of Reid on ReidAI? What did... Was he satisfied with the response, or was he surprised? Was he ?
[39:31] Parth Patil: Oh.
[39:32] Kazuki Nakayashiki: Did he complain?
[39:32] Parth Patil: We once... Okay, I think it was like, we asked it, we asked it, I think, 10 questions, and we asked Reid 10 quest- the same 10 questions. And-
[39:41] Kazuki Nakayashiki: Interesting.
[39:41] Parth Patil: ... I think it was like five out of 10 were acceptable to him. And his main feedback was like, "Uh, it's too much buzzword bingo." It just uses too much of the like, you know, it sounds a little too ChatGPT. But then I'm also like, "Well, Reid also uses a lot of, uh, you know, buzzwords, for sure." Um, which is my feedback. But it was, it was very, uh, eye-opening, 'cause it's like, now you have a lot of the GPT, the biases of like, the underlying language model show up in the character. And then it's like, you can try to prompt engineer some of that. You can try to do like, few shot for like, style. Um, and one thing I realized that was kind of helpful was, in the prompt, having examples from... So, I think of like, the different sources of data. So you have, like, podcasts, you have books, um, you have speeches, you have, um, tweets, et cetera, like LinkedIn posts, blog posts. And actually, they all have a different kind of purpose. I think the podcast, on top of the fact that you get this, like, high quality voice audio that you can train a voice clone on, which is great, right? Studio quality, no, no noise. You know, I can feed this into ElevenLabs and clone myself at a pretty good, like, at least the voice level. But also, what podcasts do, which is different from books and blog posts and writing, is conversational style. So, how you say something is often very different from how you write it. Because when you're writing, you're thinking about, you know, making it very structured, and you're really compressing your, your idea into a form that's accessible for... It's like, more publishing. It's more editorial. You're spending a lot of time writing. But in, in conversation, it's more like you're s- it's closer to you in a stream of consciousness kind of way. So, I noticed that, which was like, okay, if I have examples from the podcasts in the prompt, I can kind of like, get, you know, iron out some of these like... I can make it more conversationally s- like, mimic his style. Um, I'm sure if you, you want to go one step further, you could even fine-tune the model to get his voice. Like, how he says something. As opposed to, like, RAG, I think of more as, like, what he says, like, what did he... Like, what is the piece of information? And then, um, I think about fine-tuning and like, the prompt engineering with examples as like, how he s- how he speaks. Right. So, a little bit like, how, it's like, how he sounds, the voice, uh, voice clone. How he, uh, how he presents information, which is more like style. And then there's what he says, which is more like, the fact retrieval, information retrieval. And, uh, there's a whole extra layer, like reasoning, we can add, right? And I, I think about this like, well, what if... Like, it's, it feels not impossible that you could make a version of this that is, like, better at some of this stuff than even ReidImagine, like, something he, an idea he had 25 years ago that he kind of forgot, but the AI can retrieve it more quickly. So, this is something that is interesting to me as we get more, you know, deeper into it. And it's applicable outside of, outside of this. I think this is just one example, but you can imagine, like, a lot of characters like this and similar in different video games, a lot of different possible... I think video games are gonna see intelligent NPCs that are similar to this. Um, I'm surprised we haven't seen it yet, but I think it makes sense, because it might be the unit economics are not quite there. But I w- I do believe that if you look at the games that people play, the games that I play, we love our, you know, the NPC characters in these games. Imagine if they felt real, like, very real. Even when you're playing Pokemon, like even if it doesn't speak English, like you get attached to this creature, and it has this memory. I mean, it doesn't even have memory, but imagine you give it memory, and then it, like, it ha- it- it'll, it'll feel more real, right? And I think that that's gonna happen, um, once these systems end up more inside entertainment and media, um, traditional entertainment and media.
[43:43] Kazuki Nakayashiki: I see. And then I was curious of the data set, because, uh, you know, you said you use podcast, book and, you know, other, like public speakers and so on, so-
[43:51] Parth Patil: Yeah.
[43:51] Kazuki Nakayashiki: But sometimes people change their perspective over time, right?
[43:54] Parth Patil: Yeah.
[43:54] Kazuki Nakayashiki: Let's say if someone asks, "Oh, what's your thoughts on AI?" In early days, "Oh, I'm skept- skeptical about AI." But later, they realize the value, so they change, "Oh, AI is the future." So, in that case, they have two opposite, you know, assaults on a certain question. So-
[44:09] Parth Patil: Yeah.
[44:09] Kazuki Nakayashiki: ... how did you... I'm not sure it happened to lead, lead AI, but if so, you know, how did you make sure the data is correct with his current thoughts?
[44:19] Parth Patil: So, I don't think it's even completely solved yet. I have my own... See, I, because I'm like a lonely developer, it's very much like I try to solve, like, as much as I can, and then I understand that some things are not solved. But eventually, it will, we may have time, then we can get better at it. But for this, my current interesting solution here is like, um, like your perspective on something over time, there's the, the latest perspective, which is probably more relevant, but the evolution of your perspective over time is a little bit beyond regular RAG. Like, it's not, like if you look at traditional, like, like na- I call it naïve RAG, like top-K similarity. Let's go find the seven paragraphs that are most similar to the question that the user is asking and then assume the answer is in those seven paragraphs. Which is just, like, ridic- it's naive. And that's fine for, like, low-stakes kind of stuff, but then, um... And this is something I've been using for, like, a year and a half, uh, is knowledge graphs. Specifically, uh, GraphRAG. So, Microsoft GraphRAG is the, is my favorite framework for this. But it basically is, like, like, some questions require multiple calls to the knowledge base, and those questions req- they sometimes, like, uh, for example, you take Lord of the Rings, like the book, Lord of the Rings. And if you say you have naïve RAG, you put the book, you index the book, and then you use a chatbot that has naïve RAG, so, like, custom GPTs. And you ask, "What is the largest creature in this book?" Like, what happens is that the LLM is looking for the paragraphs that have text that is similar to the phrase "largest creature." And, um, and it'll only look at top-K. So, it might look at seven, it might look at 20, depending on how many, um, how many paragraphs you say it should look for. But if there are a thousand creatures in this universe, how can you be confident in the answer to the question, "What is the largest creature," when it only looks at seven and it's, like, guessing? Like, it's guessing, basically. It's like, "Oh, I think it's probably this one," 'cause there's a mountain as a comparison size. And then you're like, "Okay, that's not robust." That's not, that's just like you're lucky if it gets the answer right. And because actually that question requires, uh, like whole dataset reasoning. You, you, you need to know all the creatures to, or at least, like, you need to know the, all the, you need to know all the creatures to be able to ask that, answer that question reliably. So, I like GraphRAG, because you can, you can basically take all this data and then construct essentially a Wikipedia of entities and relationships on top of that data. And then when you ask a question, that question gets split up into many queries, and then it scans the graph, and then you're like, "Okay, this is the largest creature." And so you get a more reliable answer to any question that requires whole dataset reasoning. It's not bulletproof, but it's the best idea I have so far for this kind of thing. Because it's like then you can see, oh, here's this perspective. Like, here's my perspective on this topic, on, on, favorite, favorite video game, but that changes over time. Like, well, back in the day it was Age of Empires, then it was like Pokemon, then it was like RollerCoaster Tycoon, Mario Brothers Now it's like Cyberpunk 2077. But it needs to kind of get that big whole picture to get a sense of, like, that evolution in preference over time. So, I like GraphRAG for this. Um, and there's a trade-off. Y- takes like 20 seconds and, like, sometimes a dollar to answer a question. (laughs) And you're like... But I think then you can take that answer and you can put that into the naïve RAG solution, right? So, you can use GraphRAG to create a richer, um, fast retrieval system
[48:03] Kazuki Nakayashiki: I see, yeah.
[48:03] Parth Patil: Because you get this, like, new synthetic data that represents the answer, but, like, it's like, "Oh, we get this question a lot. Well, let's ask the full knowledge base." Even though that's not good for conversational speed, we can still take that and give it to the, the fast retrieval system that we use for conversational speed. So, that's my, I mean, this is just like my, this is just like my, me hacking through the... This is vibe coding. Like, this is me vibe coding. And so, like, I, I would love, if people have better answers to this, like temporal change across large datasets, like, I would, I would love to, um, hear about that maybe, like, in the comments or whatever. (laughs)
[48:39] Kazuki Nakayashiki: Yeah. Uh, I'd, we would love to learn that as well.
[48:42] Parth Patil: Right. Yeah. That's a great question. That's a great question.
[48:44] Kazuki Nakayashiki: Oh, yeah. Thank you. And, um, so i- is Reid's character in public is different from, like, his character in private office? And if the, you know, public, you know, if, you know, when, when make, even though making like, you know, digital clone of Reid, so with public data, but so if you fine-tune with internally, so maybe the character will be different.
[49:09] Parth Patil: Yeah. Definitely.
[49:10] Kazuki Nakayashiki: So, it's not a problem? Or... Okay. So how do you manage? Yeah.
[49:14] Parth Patil: Yeah. I mean, it, the, the first version that I built was off of public data and the custom GPT, right?
[49:18] Kazuki Nakayashiki: Okay.
[49:18] Parth Patil: I, I had his books, and then his podcasts were available. But it gets a lot better... This is why I think, like, where it's like, who's gonna clone you? The, the person that can clone you best is gonna be you, because you have the primary sources, the best quality primary sources. And also, like, sometimes that information isn't even out there, and then you just, you just like, oh, like, like you asked me 10 questions, and I'm like, "Oh, here are my answers." Well, now we have a new piece of data that doesn't exist on the internet. We put it into this character, and now all of a sudden, that is actually a unique ve- like a unique proposition of this character, is that its memory and knowledge is, is actually like based on some private data, or like non-public data, which might be more important in some ways than, than everything that's publicly out there. Like, and I think, um, I think it's actually in my voice clone, like my agent that I'm working on, I have my preferences for a technology stack are in there, 'cause I had it look at a bunch of projects I've been working on for the last two years. I was like, "Just, like, read all my code and, like, think about these technologies and put it in the knowledge graph." Like, like what, this is the kind of stuff that I like to play with, and yes, it's biased. It's my bias. That's actually the point, right? Like, it's my, my perspective is biased, and I want that to be in, in this thing. And while, like, ChatGPT will give you one answer, this thing is gonna give you a totally different answer, because it is grounded in your perspective.
[50:42] Kazuki Nakayashiki: Mm-hmm.
[50:43] Parth Patil: And I think that, like, um... Yeah. That's why I think, like, you should, you should make your own... The o- no one else is gonna replace you. You're gonna replace yourself if you want to.
[50:53] Kazuki Nakayashiki: Mm-hmm.
[50:53] Parth Patil: And actually, even if you tried to do that, you would realize that you are much more than what you thought. Like, you, you have this, like, expanding sense of self when you try to replace yourself. You're like, "Oh, I'm not just a data analyst. Actually, I have so many other things about me that I'm interested in that I like to think about as part of my identity." So, it's interesting. It's very... It's, it's like kind of a, um, an expansive sense of self that happens when you try to do these things. Yeah.
[51:20] Kazuki Nakayashiki: Yeah. And I'm a little bit afraid of the future where people asking, "Hey, AI agent," you know, "Plan my life, successful life, and do it for me," or something like that. (laughs)
[51:33] Parth Patil: It won't happen.
[51:33] Kazuki Nakayashiki: So, it-
[51:34] Parth Patil: I mean, you, you-
[51:34] Kazuki Nakayashiki: It won't happen?
[51:35] Parth Patil: Well, you can choose to... Why would you give the fun things away?
[51:38] Kazuki Nakayashiki: (laughs)
[51:38] Parth Patil: Like, you, you go, you could give the things that you don't want to do away, and I think that's fine. Um, but sometimes I'm like, "Damn, if this thing could just do my job." But it just doesn't, right? Like, we're not there yet. And I'm kinda just like, "Well, I'm gonna have a job for a while." (laughs)
[51:54] Kazuki Nakayashiki: Yeah.
[51:55] Parth Patil: Um, but also, I think it's like you... I like to automate the things I don't want to do. And then the things I do want to do, or if I have a very strong opinion and it's like, "Oh, this is a quality. This is my opinion what high quality looks like." And I don't think that these systems yet have that quality bar. They just don't... They're not good at, like s- like, we have that quality bar, because we are in the real world. So, I listen to a lot of music. I go to a lot of concerts. So, I have my opinion on what high-quality music is. And, uh, that's my opinion. And I would not give that to an AI. Like, I don't think an AI can do that, because it's like, can an AI get my opinions on this better than everyone else? Like, nah. Maybe it'll get, like, maybe. I like to see it. Um, but it's also like, why would I give that up? Like, it's like my taste in food, right? Like, this is not something that I'm so eager to give away. Like, I, I think I'm gonna keep all the most fun things for myself.
[52:52] Kazuki Nakayashiki: Yeah. For sure, yeah. Definitely. And now you have worked with Reid Hoffman closely-
[52:58] Parth Patil: Yeah.
[52:58] Kazuki Nakayashiki: ... and I'm curious did your impression change. I mean, before you meet Reid and work with Reid, and you had an impression, I think, on him-
[53:05] Parth Patil: Yeah.
[53:06] Kazuki Nakayashiki: ... but now you have worked with him.
[53:08] Parth Patil: Yeah.
[53:08] Kazuki Nakayashiki: Did your impression on him changed? And also, what's the biggest lesson you learned from working with Reid? If you could share.
[53:16] Parth Patil: Yeah. Yeah. I read his books growing up. And, uh, um, I listened to Masters of Scale, the podcast, uh, when I was in, in college. No, when I was in, yeah, when I was interning at my first startup. And for me, it was like, gave me the confidence to go into startup. Because he's very much like a, he's, he loves games, board games, and strategy kind of games. And I also, you know, growing up loved strategy games. So, when he was like, "Oh, the theory of the game," like how do... He, he was kind of like using analogies to games as a way to make startup uncertainty easier to navigate, right? So, I, and I think for me, it's like, I love games. And I think one time... One way to kind of like... Games are fun because you get to be competitive, but it's not gonna, it's not like, it's not like, uh, life or death, right? So, we get to explore our, our personalities, how we work well with each other, how we compete with each other, what our strengths and weaknesses are through games. And then the, the... It also just gives us... Sometimes it's like, we're... Let's say I introduce us to a new board game. All of us are learning the new game at the same time. So, some people learn more quickly. Sometimes you think, "Oh, this reminds me of that other game. This reminds me of that mechanic." And so you kind of are trying to pattern match to learn how to get good at a new game. That actually is very relevant in startups. Because I think in startup, it's very much, it's all new games. Everything is a new game. And, uh, you know, you might be the first person or first company or first team to be attacking a problem, and, and you're kind of like, you have to form a theory of the game that you're in. And that's what, that's a common Reid phrase is like, what's their theory of the game? Like, how are they thinking about this competition? How are they thinking about this ecosystem of problems? And...So, I mean, it was pretty incredible meeting him in person. And I was like, "Oh, wow, he's, that's him." Like, that is, this is how he kind of is. Like, the most, uh, like, it's very much the same guy that you, you see. And I was like, "Okay, that's, that's, that's crazy." Um, biggest thing I've learned, I think, is, um, like, this experimental, like, mindset of, like, we should just be figuring things out. Like, he loves to experiment, um... This Read AI, right? Like, I remember after it went viral, I was meeting, I was talking to him, I was like, "Man, I saw Read AI on CNBC yesterday. What is happening?" And he's like, "Hey, you know, when you're, when you're f-" Like, eventually, maybe everyone will have something like this. But when you're the first to do something, you kind of get a dis- you get a lot of attention, right? Because now, it's... Same thing. Yohei Nakajima, Baby AGI. Perfect example of this. Everyone's building LLM wrappers now. But Baby AGI was like, it got all of this attention and momentum because he was, he put it out there first, even though it was not complete. Like, it wasn't like... It didn't work end-to-end. I remember I was using it, I was like, "Wait, it can't do everything." I was like, "Obviously, it can't do everything." But that's crazy, like, people love it, even though it's not perfect yet, because of what it inspires them is possible. And so, uh, that was a huge lesson, was like, you know, you, you know, if you- if you're first to something, like, there are advantages to being first. Sympt- similar to, like, this philosophy of blitzscaling, right? If you think there's winner-take-all market, like, you have a different way to play the game. You should play the game differently if you think there's a chance you could win it all. So that, these kinds of, like, these kinds of, like, philosophies are... And, and that's the other piece that I think I totally underestimated my whole career. Read is a philosopher. And I'm more of just, like, a hacker, kind of programmer, like, you know, tech, tech guy. But I realize now the value of philosophy much more deeply. And I meant the very beginning of this, but he's always quoting philosophers, Wittgenstein, just like people, I'm like, "Who are these people? Like, these, they died 200 years ago. Why are we talking about them?" But actually, like, it's very relevant, because, um, things like language and the importance of language, and language and cognition, how you think, how you speak, how you're write, all of this is actually very important now. And a lot of the people who are thinking about this, the great thinkers of the past, kind of were running into these ideas w- before they were relevant to technology, especially now that we have programs that can emulate thought, right? How you think, and then if I think about philosophy, it's like, do you under- like, how well do you understand people? Because technology is not just like, it's not just, we don't just build machines for the sake of building machines. We build machines to solve problems that people have. So, you have to actually understand people and, like, have deep empathy, try to understand, like, what they're, you know, what do people actually care about? What do they want? What are their desires? And that gives you a better lens with which to kind of build technology, I think. Yeah.
[57:56] Kazuki Nakayashiki: Yeah. Definitely. Making something people want, and that's, uh-
[58:00] Parth Patil: Yeah.
[58:00] Kazuki Nakayashiki: ... why say always say, and-
[58:01] Parth Patil: That's right.
[58:02] Kazuki Nakayashiki: Yeah. But you said, you know, like, first to be the market is important, but do you have any idea something it's unexplored yet, but could have a huge potential? I mean, could be AI agent-
[58:16] Parth Patil: I think there's so many.
[58:17] Kazuki Nakayashiki: Yeah, yeah.
[58:17] Parth Patil: There's so many now, right? Like, I feel like I have these ideas every day, honestly. Um, the question is, what do you actually spend your time on? But I think if you look at, like, just look at any new capability, like, let's take language model. Like, this is my bread and butter, is language models. Like, okay, now we have language models. Okay. What, like, make a list of all the things that you could not do two years, like, three years ago that language models let you do.
[58:44] Kazuki Nakayashiki: Yes.
[58:45] Parth Patil: And then it's like, "Well, three plus years ago, anyone that had messy data could not do data analysis." But one of the most powerful capabilities of language models is, uh, the structured outputs. The ability to take unstructured data and turn that into structured data, right? Because when you take unstructured data and turn it into structured data, you're able to connect it to traditional software, which ex- traditional software expects structure. Language models, they seem like this, like, random kind of black box, but actually, you take, like, the, like, uh, like, if you take my bio, and then you say, um, "Extract every single company he's worked at," and then, like, extract the title, like, his title as he self-defines it. Now, you can put that into a CRM, right? But that might be unstructured at first. It might be the output of a deep research report. But then the LLM allows you to extract structure from that. I think this is the, one of the most powerful capabilities of language models, and I think people, uh, I think pe- a lot of people are just ignoring it. Um, and that was one of the first things that, actually, I was working with, uh, before Agents, was structured outputs. Just trying to get the l- model to do JSON. Like, that, that, I spent a month working with Microsoft Research, some library. I was like, "Microsoft Guidance, I like it. It's, e- it allows me to go from a blob of text into, like, fixed JSON." And, and I think people were just like, "Whatever." I'm like, "No, no, this is actually probably one of the most important things of all time." But, um, you know, you kind of have to be a little, like, in the weeds to, to realize it. But the implications, like, okay, what's the second order effect here? Well, one, it means that, um, everyone that complains about not having clean data can go v- use LLMs to create a clean dataset that represents business logic. That's awesome. So, now you can do, like, data analysis even though you have messy data. And then the second thing is, uh, like, people do this with, like, transcripts, right? Like, they have a meeting transcript and they feed it into the LLM, and then it's like, "Well, these are the action items." That, that's powerful, right? Now we can just quickly move to the next thing. Same thing, when you ask a question to the model, it uses structured outputs to write the search query for the search engine to process. So, it's happening under the hood, how the model connects to the rest of the software. And then I think the one thing that this does that is still completely ignored by most people is generative UI.Like, being able to... And, and people talk about this, like, "Oh, well, you don't want to, like... You shouldn't..." So, people complain about vibe coding, it's like, they're like, th- at its limit, generative UI is unfamiliar, and then people won't build, uh, familiarity and patterns. But I think there's, like, an, a middle ground here, which is, like, what if you're talking to Reid.AI, and, um, you know, it's like, "Oh, I had this conversation last week." Like, Reid had this conversation last week with, um, Satya Nadella. And then, the video comes up. That, that's, like, an application of generative UI, right? So, within the constraints of the structured output, the UI that we give it, it is able to fill it in with the generative piece. So, I have a feeling that this is, like, largely untapped, very important to me. Um, like, u- user interfaces that are more, um, just in time. Like, the right user interface at the moment that you need, is actually very interesting to me, and, like, kind of not really explored. Um, I really like GraphRAG. I love the advanced retrieval stuff, but, um... And I think the applications are more, uh, like, I think the applications of some of those systems is, m- far, further reaching than people f- believe. But I think you have other people that are just so scale-pilled that are just like, "Uh, we don't need this, because eventually GPT-7 infinite token context," whatever. And I'm like, "Yeah, okay, fine." But, like, until we have that, like, this is kinda cool. Like, this allows me to create new artifacts, um, in the meantime. And so, I, I like, I like, uh, yeah, I think... Because memory will probably get... I mean, will, memory will get solved, but it's not solved. Knowledge, structured knowledge is valuable, in my opinion, um, because there's not a Wikipedia page for everything on the planet. Like, there's plenty of things that don't have a Wiki- I don't have a Wikipedia page. Like, so, but, but, like, LLMs could make that. Now, I would, it wouldn't be called Wikipedia, because now it's not human review. It's not like humans wrote it, humans reviewed it. It's a different thing. But, um, that synthetic thing might actually be very useful. Um, so I think there's a lot of artifacts that these models can create that will be valuable to us, um, in, in ways that we can't expect. Let's see-
[01:03:18] Kazuki Nakayashiki: Like art.
[01:03:19] Parth Patil: Yeah. I think generative games. Generative games. That's, that's an area that I'm, like, very excited, and it's, hasn't yet happened.
[01:03:27] Kazuki Nakayashiki: It's like a Genie-3 view?
[01:03:29] Parth Patil: I think there's two versions of this that I'm very excited about. So, one is the, the Genie-3 kind of etched, uh, that they did this with Minecraft too, which is, like, the model renders the whole game on the fly, and then you interact with it. But it's, like, it's, like, diffusion model, just, like, uh, the game is just totally, like, dreamed on the fly. I think that's exciting. Uh, I also am more interested in, like, um, like, th- that's cool, but I think there's, like, generative games with more, like, where, uh, like, closer to Dungeons and Dragons. I don't know if you've ever played Dun- if, if you're familiar with Dungeons and Dragons. But it's like a, it's like a game where you speak... You're kind of like talking, you're pretending to be a character and you talk, and then the, someone is conducting the, uh, the game, and then the, the game unfolds around you in conversation. But it's really more like improv. It's more like a conversation with friends. You're pretending to be characters. I think there's something there, because there's, that even can be automated using LLMs. And so, like, if, if that experience can be made easier, then it can be made more fun. And also it can be made more accessible, so it doesn't have to just be fantasy. It doesn't have to be just dragons and, and witches, and. It can, it can be anything. And so there's... And then I think that applies to a, a lot of genres of games. So, it's like, so maybe the diffusion games are, like, where it ends up, but there are other things that we can do until then that, like, like if you think about Pokemon and Yu-Gi-Oh!, what does the generative version of Yu-Gi-Oh! look like? I, I think about Yu-Gi-Oh! a lot, because it's like, in the TV show, I don't know if you guys remember Yu-Gi-Oh!. But, like, in the TV show, everyone has their own deck. And, like, Kaiba has Blue-Eyes White Dragon, and then Yugi has Dark Magician.
[01:05:13] Kazuki Nakayashiki: Yes.
[01:05:13] Parth Patil: And it's like, "This is my signature card." And I'm like, "Man, what if we had a card game where everyone had their own signature deck?" And it was just, you had your deck, I have my deck. And maybe this is what NFTs was supposed to be, but, like, now it's possible, right? Like, generative AI should make that possible. So, I'm kind of like, "This feels obvious. Is it big idea? Who knows." So, I'm kind of like, "Well, let's just go hack some version of it. And then, like, if people are having fun, maybe we see what, we see where it goes."
[01:05:41] Kazuki Nakayashiki: Yeah.
[01:05:41] Parth Patil: But, yeah.
[01:05:41] Kazuki Nakayashiki: Very interesting. Yeah. Yeah. But, so, uh, uh, uh, so when you mention about Wikipedia, I, I came up with a random question. But do you think in the future, do we need Wikipedia? Because, because we have access to LLM. I was just curious about that.
[01:06:02] Parth Patil: I think we're gonna need, um... We're gonna want some things where it's, like, humans, um, I think there's gonna be humans and AIs. There's gonna be a version of Wikipedia that's probably way bigger, which is made by humans with AIs. And I think that the, the feel, the... I'm a little bit, like, I think we need something, because wh- what's gonna happen is, it's already happening, which is, like, there's gonna be a lot of synthetic stuff out there, and then it's, like, how do you tell what's real? What's, what do we believe that's, like, fact? And this is a hard problem. And, um, some people think that crypto is part of the solution. Maybe. I don't know. Um, but if the internet ends up filled with a bunch of fake information, fake, fake data, we're going to have, like, a lot of side effects. I think one of the side effects is that, um, if social media platforms keep getting filled with AI, we will have our group chats, which are purely human, maybe. Like, we're gonna have places we go where we only want to be with people, and maybe in the real world, right? Because that's where it's easier to be... It's obviously, like, yeah, you're real, I'm real, right? But, like, if, if it's Twitter, it's too easy...... for Twitter to be filled with bots. It's too easy for that to dilute the experience. And then, uh, then people lose trust in their information. And if the trust in the information goes down, that's gonna be really, like, it's one of the biggest problems that is not solved. And then I think, like, Wikipedia, is Wikipedia this answer? I don't know. Is it like, a version of Wikipedia that's built with AI, but, like, there's some kind of, like, verification system? Uh, Balaji talks about crypto as being a possible solution here. I'm like, "Hey, we need to try these things, and like, figure it out, because the problem is only growing." But I think, like, every problem that AI creates also will create billion-dollar solutions. Like, like, ev- like, AI is part of the solution in many of these, in these problems. It's a problem of we need a system with the right incentive structure to preserve the values we, we care about. So, I think of it as, like, okay, if that's the chaotic world we are headed towards, how would we get a network of people where we actually agree on the values and we orient ourselves towards, like, playing a certain way so that we get better information streams? Maybe that's group chats. Maybe that's, like, a return to, like, the physical world, right? So, um, totally, like, I don't, I don't think I have the answer, but I'm, I'm optimistic that we will figure it out. I think technology is a huge part of the solution in most of these cases as well. Same thing with cybersecurity, right? If vibe-coded apps are vulnerable, then also AI is gonna be a part of the solution. Like, AI hacking your website and then telling you before, ins- instead of telling, like, a bad actor, that's gonna be... I, I would pay for an a- I would pay for AI to hack my stuff and tell me and not tell anyone. (laughs) You know? Yeah.
[01:09:01] Kazuki Nakayashiki: Yeah. Definitely. Yes. So yeah, that reminded me of the early days of Wikipedia. So, when I was a college student, high school student, I still remember teachers always say, oh, you know, the trust in information on Wikipedia was way lower than I see in the current situation today.
[01:09:18] Parth Patil: Yeah.
[01:09:18] Kazuki Nakayashiki: As we have today, so. But now, people say, "Oh, AI hallucinates, so we can't trust AI." But-
[01:09:24] Parth Patil: Yeah.
[01:09:24] Kazuki Nakayashiki: ... people hallucinate.
[01:09:25] Parth Patil: Yeah. Yeah. Yeah.
[01:09:25] Kazuki Nakayashiki: And I think, yeah, that's... I see the interest in, how to say, cyclical change over time-
[01:09:32] Parth Patil: Yeah.
[01:09:32] Kazuki Nakayashiki: ... of information we should trust, you know? Yeah. Sorry .
[01:09:35] Parth Patil: I, I feel the same way. I feel the same way. And then I think, man, there's so many topics I care about, and then there's no Wikipedia page. And I'm like, okay, I would rather there be an AI Wikipedia page, at least. And then, um, and then the people who care, we can go be like, "Whoa, whoa, whoa, whoa, whoa." Like, maybe it's human review, maybe it's definitely gonna be AI part of the review process. But, like, I think it's better than there not being a page. But maybe Wikipedia maybe won't do this. And, like, maybe, like, you and I, like, we might just make these, like, synthetic wikis. I do this a lot of times with GraphRAG, actually. I just make a bunch of wikis for my own favorite topics that do not have, do not have pages. And I'm like, "Okay, like, it's fine. I'll make my own. Me and my LLMs will make my own." For the games that I play, for example, right? Like, they may not have a wiki for that game. And then I'm like, "Oh, we'll just, I'll read the docs, and then we'll have GraphRAG go and, like, index everything." At least now I have a wiki. Is it accurate? Well, the alternative is it doesn't exist. And I can see that, the primary sources. So like, same thing with Wikipedia. You can cite Wikipedia, but, like, you better look at the source that it's linking to, 'cause even that, like, people who write, they, they put stuff on there that's not even, like, like they just... They're making the source and then they're putting it on there. And there's the... So, there's abuse factors even in human Wikipedia. So, um, yeah, it's gonna be an interesting one. But I think we're gonna have some synthetic version of this. Yeah.
[01:11:03] Kazuki Nakayashiki: Yeah. And I remember someone said, you know, if we show the citation, the answer, then s- people, without checking it, most people trust it. But when we check, you know, the, the, that information is generated by AI or someone, you know?
[01:11:17] Parth Patil: Yeah.
[01:11:17] Kazuki Nakayashiki: It's not correct. And they cite another information, and citing, citing, citing the wrong information all the time, so-
[01:11:24] Parth Patil: Yeah.
[01:11:24] Kazuki Nakayashiki: ... it's kind of crazy.
[01:11:25] Parth Patil: It's garbage in, garbage out, and then multiple layers of garbage. And then you're like, "What are we doing?" Um, that's a very-
[01:11:31] Kazuki Nakayashiki: But if you have cite-
[01:11:32] Parth Patil: Yeah?
[01:11:33] Kazuki Nakayashiki: But if you have a citation in answer, we-
[01:11:36] Parth Patil: Assume it lends credibility, yeah.
[01:11:38] Kazuki Nakayashiki: Yeah.
[01:11:39] Parth Patil: Yeah.
[01:11:40] Kazuki Nakayashiki: That's, that's an issue, yeah.
[01:11:41] Parth Patil: It is an issue. It is an issue.
[01:11:43] Kazuki Nakayashiki: Yeah.
[01:11:44] Parth Patil: Um...
[01:11:45] Kazuki Nakayashiki: So, yeah.
[01:11:45] Parth Patil: Maybe reasoning helps with this. Maybe reasoning helps with this. But it depends on how many, how many layers deep does it go, right? Like, if it's all-
[01:11:53] Kazuki Nakayashiki: Yeah.
[01:11:53] Parth Patil: ... AI-generated, citing AI-generated, then, uh, where, what is real? This is a qu- huge question, what is real? And then we also have this problem in video. The video models are so good that you can't... it's not a, it's not a... Two years ago, it's like, oh yeah, six fingers, it's clearly AI. Now it's like, the bunnies are hopping on a trampoline, and I'm like, "Uh, it looks real to me." (laughs) You know? And then a week later, Google is like, "That was Google Gemini." And I'm like... My thing is like, on video, I'm like, why don't we just generate things that are obviously not real first, instead of recreating reality, which is impressive, but, like, deceptive. Or, like, don't do it for deceptive reasons. Like, make some other u- You can generate any universe. The video model is like an advanced simulation. You can generate a candy universe. No one's gonna look at a candy universe and be like, "That is real." And actually, maybe that's... My preferred way to go is like, oh, let's do animated. Let's try some other style where, like, I'm not trying to deceive you here with, like, something that is pretending to be real. Of course, I worked on Read AI, but he says he's a Read AI, right? Like, this is part of the design choice is, like, even if it would deceive you, if you... Like, it's like, it's gonna trick your grandma two years ago. Now it's gonna trick you, because there's no skill in identifying whether it's AI. You're kind of just like, you scrolled to the comments and you're like, "Hopefully this is real." And then it's like, "This is AI. This is AI." And then you're just like, "How do you trust it, even if the comment is written by an AI," right? (laughs) So, it's, uh, these are, these are the... I, I, I knew we would get here. I didn't think we would get here so quickly, is for sure.
[01:13:35] Kazuki Nakayashiki: Yeah, it's a crazy days we live in.
[01:13:36] Parth Patil: Crazy days.
[01:13:37] Kazuki Nakayashiki: And I was thinking about this idea, I mean, to train humans to distinguish if it's AI or not. So, let's say, in early days, we are kind of easier to get tricked by spams, you know. Like-
[01:13:50] Parth Patil: Yeah.
[01:13:50] Kazuki Nakayashiki: ... like, "Oh, you were chosen. You have the right to get $1 million." You know, from, I don't know, like, king or someone.
[01:13:56] Parth Patil: Yeah. Yeah, yeah, yeah, yeah.
[01:13:56] Kazuki Nakayashiki: And so, this is just totally example. But now-
[01:13:58] Parth Patil: Yeah, yeah. The email, the email you get. The email spam.
[01:13:59] Kazuki Nakayashiki: Yeah, email, email. But now, you right away distinguish, "Oh, this is spam." But early days, some people got tricked, right? So- but they say-
[01:14:07] Parth Patil: No, but now, the spam can be generated by the LLM and you can't tell.
[01:14:11] Kazuki Nakayashiki: Mm-hmm. Y- yeah, yeah. Tha- that's right.
[01:14:12] Parth Patil: (laughs)
[01:14:12] Kazuki Nakayashiki: But if we have a platform that teach people, oh, hey show, let's say, two videos or two pictures-
[01:14:19] Parth Patil: Yeah.
[01:14:19] Kazuki Nakayashiki: ... two, uh, texts or something. Then if they say, "Oh, this is spam," or, "This is AI generated," or, "This is human created," then, then we can train, "Oh, actually this is generated by AI," so, um, then so that we can how say, get used to the things that generated by AI or I don't know.
[01:14:37] Parth Patil: I don't think this is gonna be possible. And it's just like-
[01:14:39] Kazuki Nakayashiki: Oh.
[01:14:40] Parth Patil: I don't think it's possible. I mean, look at VO3, right? Um, I can generate something in VO3 and, and, like n- I mean, I, I did this with, uh, Runway Gen-
[01:14:54] Kazuki Nakayashiki: Mm-hmm.
[01:14:54] Parth Patil: ... 3 Runway, and I made a Flux. I, I created a LoRA of myself, so I had an image model Flux 1, and I fine-tuned a LoRA on my face, and then I made a music video. This isn't... I did not put it publicly on my YouTube channel. It's a private video. Maybe I'll put it public after this now 'cause it's like now it's, it's like obviou- okay, I'll just say it's totally AI. But I didn't put it up because I was like, "Oh, my God," 'cause I showed it to my brother, and I was like, "It's me in 15 different universes. In the 1920s, me g- going to space." And my brother is like, "The only reason I can tell that this isn't you is because I know you've never been to the moon." And I was like, "But otherwise, it looks like I'm on the moon." And I'm like, "Okay, there's no skill in discerning this," right? For the person that doesn't know me, it's like, "Oh, you must have been in a movie about the moon." And I'm like, "I, n- it's, I mean, this is a movie that I made. It's like a little two-minute clip of me in different universes." But I didn't make it public 'cause I was like, "Ah, actually, like, I, I, I, like, I'm not sure, like, how I wanna deal with this yet. Like, because it's, looks so real." And my brother was like, "No one can tell. Like, it looks like you. I, I grew up with you, it looks like you. So everyone else, for all intents and purposes, will think it is you unless you tell them it's not you." And so, I think that, um... And, and then it's like, that's a lot of work, like training, and also the models are just getting so good. The models are already at a point where I don't think you could train me to tell the difference. If someone wanted to make something that you could not tell was real, they're gonna be able to do it. So, then we need systems of, like, um, the word is providence. Systems of providence, like, systems of, like, this... We need new systems where we can s- ascertain, like, collectively, like, this is a network of, of, of content that is real. And then actually everything else, you have to assume it's fake. And unless we can prove that it's real somehow because of the new system we create, which doesn't exist yet, but people are trying to make it, you will just deny everything. Like, you're just gonna have to, like, it's like, "Oh, don't believe everything you read on the internet. Don't believe everything that you see is real on the internet."
[01:17:07] Kazuki Nakayashiki: Isn't it sad? Isn't it a kind of dystopia (laughs?) in the future?
[01:17:11] Parth Patil: It is. It is.
[01:17:13] Kazuki Nakayashiki: Mm-hmm.
[01:17:14] Parth Patil: As long as we don't solve it, it's gonna feel bleak.
[01:17:18] Kazuki Nakayashiki: I see. Yep.
[01:17:19] Parth Patil: Yeah, yeah.
[01:17:19] Kazuki Nakayashiki: Yes.
[01:17:20] Parth Patil: But on the flip side, whoever solves it is gonna become a billionaire.
[01:17:25] Kazuki Nakayashiki: Yeah. (laughs)
[01:17:26] Parth Patil: (laughs)
[01:17:26] Kazuki Nakayashiki: For sure. For sure, yes.
[01:17:27] Parth Patil: So, there's an opportunity, right? Yeah, like... (laughs)
[01:17:30] Kazuki Nakayashiki: Yep. (laughs)
[01:17:30] Parth Patil: Yeah, maybe.
[01:17:31] Kazuki Nakayashiki: Uh-huh.
[01:17:31] Parth Patil: Or, maybe they'll make it, like, they'll do it just for the good. I don't know. We'll see, but people are working on it, and I think it's a, it's a, um, it's probably one of the most important problems, um, that AI creates. AI creates some problems. It is gonna solve g- a lot of problems. It's gonna create these problems, and it's gonna be a part of the solution. And then you have even crypto people are like, "Oh, well, we should make it so that if it's real, it must be verifiably on this blockchain." And I'm like, "Hey, man. If you can figure out how to get everyone to agree to that system, maybe." You know?
[01:18:02] Kazuki Nakayashiki: Interesting. Yeah.
[01:18:03] Parth Patil: Yeah.
[01:18:04] Kazuki Nakayashiki: And so, so what's next, uh, at th- the Office of Lee-Rock One? Can you share? Oh, I don't know if you can share, but do you have any-
[01:18:12] Parth Patil: Yeah. Uh...
[01:18:12] Kazuki Nakayashiki: ... future AI project coming out, or?
[01:18:15] Parth Patil: Well, so I spent the last week-
[01:18:17] Kazuki Nakayashiki: ... you put in?
[01:18:17] Parth Patil: Last Wednesday, I spent the whole day talking to 20B GPT-OSS. Um, just, like, my personal project, right? Like, I spent the whole day talking to GPT-OSS, and then I was like, "Okay, wow, like, this model is very efficient in, in terms of, like, s- it runs on a Mac mini at 30 tokens per second, and it's completely local." And I'm like, "Wow, this is crazy." And so, then I have a new laptop coming, and I'm gonna be playing with the 120, uh, billion parameter variant. So, I'm gonna be working on some local agents just to understand, like, I think I've kind of been mostly in the, you know, using the API, and, like, using these, um, frontier models which you're kind of renting your intelligence. But, you know, a lot of other people, of course, have been in open source this whole time building local models, but I think the local models will be, um, agentic, and you're gonna have these- this kind of, like, like, local running agent smaller in scope but, like, useful for what it needs to be. My dad always tells me, he's like, "Parth, the fridge does not need super intelligence." So, he, he always says that. He's like, "We- we... There is a version of this that doesn't need to be, like, trained on the entire internet for it to be useful," and, and 'cause he comes from, like, the last era of computers, right? So, he thinks about, like, like, uh, uh, he thinks about, like, the cloud. He thinks about, like, on-prem. He thinks about a l- a lot of this stuff, and he's like, "You get every... You get a lot of different sizes..."... of, uh, of these things. And you're gonna get the biggest ones, yeah, it's gonna be on the internet. The biggest, smartest models are going to be the ones you rent. But also, like, you're gonna have useful small models. I think, I think, I think about my Roomba, and it's like, this Roomba is s- so behind. It's not even GPT-1, right? Like, it gets stuck in the corner, and then I'm like, "Oh, my God." And my cats are like, "This thing is not smart." And I'm like, "This is..." But imagine a world where, like, the devices are actually intelligent because they have intelligent, locally running, small models. That's gonna be interesting. Um, in terms of, like... So that was, like, an exploration from last week. And then last couple of days, I've spent most of, most of my time working with GPT-5. And that has made a lot of my, uh, a lot of our projects internally a lot faster. Like, there are things that you can make now in three hours that used to take me weeks. And, um, I think the combination of Claude code, GPT-5, and honestly, me just getting better at using them. Like, I think the huge thing is, like, adapting yourself, um, switching the way you work to being more agent-friendly, how you orchestrate these tools. Um, you know, you start by using one. When you get good at using one, can you use multiple? And then when you use multiple, can they work together on things that are even more advanced? I think we're getting there. It's very human-in-loop. It's not like... End-to-end automation, it's getting better at long-running tasks, better at end-to-end automation on fir- you know, better at zero-shot. But the, the speed at which you can build tools is accelerating. And I think we have some tools that, um, uh, that I'm working on that are made possible by GPT-5. Like, yesterday, I was like, "Wow, GPT-5 built this, like, pipeline in, in, in two hours," that I was gonna spend a lot of time on, but now I'm like, "Wait, we can go much bigger on this in a shorter timeframe." And, uh, I'll have more to share. We'll, you know, we're, you know, we always try to share how we do things. Like, we, we try to create content, "Oh, here's how we made Read AI." So, I think we're gonna have some, some more stuff like that, that people can kind of, like, get the playbook. But really, it comes down to just, like, play with the tools, and then, like, use them and get really good at using them. And then, and then, um, share what you're learning with people. And then that kind of is, like, a good feedback cycle. Yeah.
[01:21:56] Kazuki Nakayashiki: Yeah, definitely. And before, you know, started, we starting this recording, so we talked about email capture. That's the idea we hacked, you know-
[01:22:04] Parth Patil: Yeah.
[01:22:04] Kazuki Nakayashiki: ... during the last week, weekend. And actually, I used GPT-5, and, and it worked in a single try.
[01:22:10] Parth Patil: Yeah.
[01:22:10] Kazuki Nakayashiki: But when I used, uh, Claude, uh, 4+4.1-
[01:22:15] Parth Patil: Yeah.
[01:22:15] Kazuki Nakayashiki: ... or something like-
[01:22:16] Parth Patil: 4.1.
[01:22:16] Kazuki Nakayashiki: It didn't work in a single try. So, I was surprised by the, the capability of GPT-5 and that's really interesting.
[01:22:24] Parth Patil: This is how I'm feeling right now, man. Like, I'm like, "Oh, my God." 'Cause I call it, like, one-shot, like, oh, did it one-shot the problem? I'm like, "Oh, my God," 'cause if it can one-shot the problem, then you're like, "We gotta think bigger." Like, "We gotta do way more stuff, we gotta try more things," because, like, turns out, like, this is not gonna take a week. It's g- the first version might only take two or three prompts. And if that's the case, then, like, we need to be more ambitious. We need to be more creative, because the models are really getting to this level of, like, uh, code generation has never been this good. And now we have the CLI tools that are getting better, the, the wrappers around them that are making them even faster and more reliable. The tools, MCP is great. MCP is also a little bit, you gotta be careful, but it is very good. And then, uh, and I agree. I think, what can the model one-shot? I think I'm like, I look forward, for me, um, my AGI test, when it's like, "Is it AGI?" My test is, I'm gonna just go to the model and say, "Clone Clubhouse," and then that's it.
[01:23:25] Kazuki Nakayashiki: Mm-hmm.
[01:23:25] Parth Patil: "Can it build Clubhouse in one shot?" (laughs) You know, it's like-
[01:23:28] Kazuki Nakayashiki: Ouch.
[01:23:28] Parth Patil: ... take your time. Maybe take two hours. I'm gonna go for a walk, and I come back, and there better be an application that I can play with. But that mindset of experimentation is important, because I think, like, you think about big companies, and they sit around, and it's like, "We should make this." So then they write some document on, like, what the feature is gonna be, and then, like, six months of, like, planning and, like, building. And I'm like, "No, no, no, no, no, no. We're gonna sit here, Replit is gonna be here, and we're gonna tell it to make this thing, and then we're gonna see how far we get in this meeting." And then that updates our own, uh, priors. We're like, "Okay, cool, turns out we can get this far in two prompts, in three prompts." It's, like, really, like, there should be a hackathon where it's like, you get 10 prompts, and then whatever you submit at the end of that, that's the submission. Like, how far do you go in 10 prompts?
[01:24:13] Kazuki Nakayashiki: That's funny.
[01:24:13] Parth Patil: Like, well, how far do you go in one prompt? Like, that should be a hackathon category, right? Because then you're gon- you start thinking about, um, uh, Dexter, uh, Dexter Horthwee, he's, he said, he's like, "Context engineering," right? Like, the new phrase? Which is true. It's like, are you bringing the model? Are you bringing the right context into the context window? And then you can be much more efficient with what you get on that first try. And I think it's, it's awesome to hear that you're doing that, and that that is a great story of GPT-5 one-shot. Because I think there's so many more of these. Like, we don't even know what it can do that quickly unless we try it. So, that's why I say, like, don't judge a model until you've talked to it, like, unless, until you've prompted it a hundred times. Then, maybe you can judge the model. But there are so many things we don't know yet that we will figure out just by playing with this and talking to each other and sharing what's working, you know?
[01:25:02] Kazuki Nakayashiki: Yeah. I'm looking forward to joining Promptathon. (sniffs) Someday.
[01:25:06] Parth Patil: (laughs) Yeah, yeah, yes, Promptathon. (laughs)
[01:25:09] Kazuki Nakayashiki: Yeah. Yeah, because, because at Hackathon, I know some people bring their startup project, you know, or pre-built project, you know, to the Hackathon.
[01:25:15] Parth Patil: Yeah.
[01:25:15] Kazuki Nakayashiki: And they win the prize, and it's kinda unfair for-
[01:25:18] Parth Patil: I agree.
[01:25:18] Kazuki Nakayashiki: ... some people who rarely build, you know, that, that project during the Hackathon, right? Something like that.
[01:25:23] Parth Patil: Yeah.
[01:25:23] Kazuki Nakayashiki: Yeah.
[01:25:23] Parth Patil: Yeah.
[01:25:23] Kazuki Nakayashiki: But Promptathon, you can't cheat, right?
[01:25:25] Parth Patil: Yeah, yeah, yeah.
[01:25:25] Kazuki Nakayashiki: You need to bring your prompt.
[01:25:27] Parth Patil: You got one box, one prompt-
[01:25:29] Kazuki Nakayashiki: Yep.
[01:25:29] Parth Patil: ... one attempt. (laughs) No thinking allowed, no planning around it. (laughs) Yeah.
[01:25:35] Kazuki Nakayashiki: Yeah.
[01:25:35] Parth Patil: Yeah.
[01:25:35] Kazuki Nakayashiki: That's funny. So...
[01:25:36] Parth Patil: It's fun. Yeah.
[01:25:37] Kazuki Nakayashiki: Yeah. Anyway, yeah.
[01:25:38] Parth Patil: It's fun, but it's actually an important exercise. Even if you don't ship the application, it's an important exercise. Because when someone tells me, "Oh, it's gonna take six months," I'm like, "What?"... why is it gonna take six months? Like if- is it because you need to meet actual people? Like, is there- what is the bottling? If it's actually just software, we need to move way faster. If it's like hardware, if it's in the real world, if it's like brick and mortar stores, I get it. There's more work to do. Relationships, human relationships. But if it's
[01:26:04] Kei Watanabe: just like pure software, we should be going very fast, and we should be learning very quickly. Yeah. Yeah. Um, um, this is, uh, just a general question, but- Sure. ...
[01:26:16] Kazuki Nakayashiki: so where is the source of information and ideas? So what kind of source, like on, like X, LinkedIn, YouTube? What do you see and how do you collect ideas?
[01:26:27] Parth Patil: Yeah. Um, so I, I, I have like nine, no, I have ten messaging apps. And, and, uh, I- so for me it's like ten messaging apps. So, like, people are always DMing me. They're like, "Oh, did you see this? Did you see that?" And I don't try everything, but I'm like, "Oh, interesting." But you have to have a, like your own network. You have to have a lot of people that are doing different things and- in AI, uh, like, or, or in technology, or just creative, right? People who are creative and different. And then they- you have to bounce ideas off of them. I think the, the network intelligence is the most important thing. Um, Twitter is good. I think that there is a, the problem... There are problems with Twitter. I think it can be kind of a time sink. And also, I think people tend to, they tend to pay attention to some of the wrong things, and they go deep in some of the wrong directions, in my opinion. Um, but it's, it's good 'cause you get a good feed. But I think that like, I have a, you know, I have like my own, all my favorite hackers, I put them all into a Discord. And then we have our own like, "Oh, here's what I'm making. Here's what I'm trying." Like, "Oh." And then it's like, "Oh, like, someone might be hiring." And it's like, "Oh, looking for an engi-..." So there's opportunities. And I think once you start networking the people who are, uh, you know, really playing with the technology, then the network intelligence benefits the whole group. And that's important. And it goes back to, like, what is real, what is false. Like, you gotta create these, like, spaces where you have higher trust, and you curate people so that it's like, "Oh, these are good people. They are helpful. They're generous with their informat- with their, with their advice and, and time." And then they're also learning and by making things, right? So I think for me personally, it's like, I don't want to talk to people that don't make things. That's just like, I'm kind of getting biased like this now. It's a- maybe it's a bad thing, but like, I prefer spending time talking to builders. Like, it's like, "Oh, you made this yesterday?" I was like, "That's the kind of person I want to talk to." Because, like, it, th- that's the experimentation we need. And also, there's just too much, there's too much happening, right? So, no one person can keep up with all of this progress in AI, um, alone. I, I think it's singularity. There's- for me, the definition of singularity is when information starts, like, technology starts advancing so quickly that you cannot keep up unless you're using AI and friends to keep up, right? Like, you have people that are covering different topic areas. So I have friends that are just like, really good at working with open source models, and I just like, learn from them. I just let, you know, ask them dumb questions. And you need to ask dumb questions, like non-judgmentally just, uh, ask questions. No one is an expert in everything, but everyone can get very, you know, deep into a few things. And then when you share that, then the whole group benefits, and then you get the interesting intersections, right? So I'll be like, "Oh, the voice, you know, the ElevenLabs v3 is actually very good for emotional control." And then my friend, who's more of like a creative artist designer, and he's like, "We should do like a Dungeons and Dragons kind of experiment where we have LLMs." And then I'm like, "Oh, we can use the OpenAI Agent SDK." And he's like, "And we can use these image models," right? So, the combination of two people that are actually very different, because they're focused on different things, but if they agree that there's a cool, interesting idea, that is actually the magic. The magic is the intersection of your intelligence and my intelligence. Um, so I don't think th- and I think also... Yeah, I mean, I do think I generate a lot of ideas. Um, but the, the, the, the, the consuming a lot of like, y- having friends that are different is, for me, the most important thing. 'Cause then I'm like, "Oh, what's your favorite LLM?" Like today, like my friend's like, "Parth, you're the only person I know that likes GPT-5." And I'm like, "Wow." Then I have to like explain why I like GPT-5. Because otherwise, he's gonna assume that like maybe it's completely useless, based on everyone else he's talking to. Maybe I'm wrong. Maybe I'm like, maybe I'm in the- like, maybe I am wrong. But, but, um, that's the kind of thing, right? It's like, if I don't ask him what... It's like, "Well, what model are you using? Okay. When? Why? What's your use case? Okay, interesting. Well, that makes sense then. Like, why it's better for your use case, worse for mine." But unless you're talking to people, and I mean like, talking to people, like not Twitter. Twitter is not social. Twitter is like people kind of just blast stuff out there. But a dialogue is actually very powerful. So I also use Clubhouse still. Um, not a lot of people are on there, but for me-
[01:30:45] Kazuki Nakayashiki: Mm-hmm.
[01:30:45] Parth Patil: ... a lot of my friends are there and, and we just have conversations. Once a week, we're like, "Oh, this is very interesting. This is very useful. This is how I'm using it." And, you know, for three hours, a conversation like that ends up being very helpful to stay up-to-date with things, and also to just get more ideas in the mix. And you have to have a lot of different perspectives. Like, I go to meet my friends in different cities, because I wanna s- I want to think differently. I live in LA instead of San Francisco because I like all the creative people here. And, uh, they're just totally... They're not in technology, but for them, they're like, they're so interesting and creative. So when I ask them, "Oh," like what, like learn more about what they're building, I get so many, so much inspiration. So you have to have people that are very different from you. I think this is... Network intelligence is key. AI is gonna help, but like network intelligence, curate your network, choose interesting, you know, hardworking, creative, generous people. Yeah.
[01:31:40] Kazuki Nakayashiki: Yeah. I like the word network intelligence and that's a great word.
[01:31:44] Parth Patil: Yeah.
[01:31:44] Kazuki Nakayashiki: And also, that's why I'm going to-
[01:31:45] Parth Patil: It's also a Reid concept. (laughs)
[01:31:47] Kazuki Nakayashiki: Oh. Oh, yeah. (laughs)
[01:31:48] Parth Patil: It's, it's one of Reid's like... (laughs)
[01:31:49] Kazuki Nakayashiki: Yeah. (laughs)
[01:31:49] Parth Patil: I mean, he's the networks guy, but now it's more important than ever, because, you know, you have AI. I have AI. So like, both of us are amplified. But now, when we come together, we-... or even more amplified.
[01:31:58] Kazuki Nakayashiki: Yeah. We can come up with new ideas and, and-
[01:32:01] Parth Patil: Yeah.
[01:32:01] Kazuki Nakayashiki: ... see things from different perspectives.
[01:32:02] Parth Patil: Yeah. Otherwise-
[01:32:03] Kazuki Nakayashiki: That's what-
[01:32:03] Parth Patil: ... you know, you can-
[01:32:04] Kazuki Nakayashiki: Yeah.
[01:32:04] Parth Patil: It feels lonely and you can feel like you're going a little crazy. But you-
[01:32:07] Kazuki Nakayashiki: Mm-hmm.
[01:32:07] Parth Patil: That's why you need, you need people and conversations, yeah.
[01:32:11] Kazuki Nakayashiki: Yeah. And that's why I love going to hackathons because I can meet a lot of great, interesting, creative developers trying-
[01:32:18] Parth Patil: Me too. Me too.
[01:32:18] Kazuki Nakayashiki: ... or tackling some interesting ideas from different approaches and, yeah.
[01:32:22] Parth Patil: Yeah. Hackathons are... I, I recommend that for... A lot of people are like, "How do you get started?" I was like, "Go to a hackathon." And then you're gonna be like-
[01:32:28] Kazuki Nakayashiki: Yes.
[01:32:28] Parth Patil: You're gonna realize like, we can go fast, you can make things for fun, you're gonna learn quickly, and that people are using the state-of-the-art tools, and you're like, "Wow." And, and there's no experts, right? Like, everyone's just learning. So, it's good to go to that environment where it's okay to be like ha- just good- it's good to... I went to a hackathon like two years ago, and my mindset isn't like... Now I just, I never left the hackathon. I'm still in the hackathon mode every single day now, right? So, it's a very like... I think it's a good, uh, exploratory, uh, set of- like way of thinking and, and making, yeah.
[01:33:01] Kazuki Nakayashiki: Yeah. And I have a designer friend and he- I used to be asking him, "Hey, let's go to, you know, a hackathon together." But he hesitated because he's designer. He thought he couldn't contribute to coding or something.
[01:33:12] Parth Patil: Yeah. Yeah.
[01:33:12] Kazuki Nakayashiki: But nowadays, he can use AI tools to code.
[01:33:14] Parth Patil: That's right.
[01:33:15] Kazuki Nakayashiki: So, now he asks me, "Hey, next time to go- you, you go to hackathon, please ask me. So, I'm a designer but I can-"
[01:33:21] Parth Patil: (laughs)
[01:33:21] Kazuki Nakayashiki: "... use AI tools now. So, I, I can-"
[01:33:23] Parth Patil: Yeah.
[01:33:23] Kazuki Nakayashiki: "... code and build something."
[01:33:25] Parth Patil: Yeah, yeah, yeah. We- I did the same thing. I brought my designers to hackathons. In LA, we have hackathons, but our hackathons are like video model, creative hackathons. It's like, you're gonna go, you get paired with three random people, and then it's like, "Make a music video in three hours. Whatever tool you want." And so, for us, the hackathons here are more like in the media side of things. "Okay, what kind of story can we tell? This tool lets you do like, you know, style transfer. This tool, like VO3 plus like Midjourney plus..." So, you, you- people have- they all have their own favorite tools and then you come together and you're like, "Wow, we can now make this new thing." So, here our hack- we have technical hackathons, too. But I think our creative cre- even creative hackathons, very useful, right? T- storytelling and just like learning how to use these tools. 'Cause there's no like textbook, there's no experts. So, the people who are the experts are those people. Right.
[01:34:16] Kazuki Nakayashiki: Mm-hmm.
[01:34:17] Parth Patil: Yeah.
[01:34:18] Kazuki Nakayashiki: Yeah. Definitely. Yes. Yeah. We are living in the kind of golden age.
[01:34:22] Parth Patil: Renaissance.
[01:34:23] Kazuki Nakayashiki: E- era.
[01:34:24] Parth Patil: Renaissance.
[01:34:24] Kazuki Nakayashiki: Yeah, renaissance. Yeah.
[01:34:28] Parth Patil: Yeah, yeah.
[01:34:28] Kazuki Nakayashiki: Yeah. So-
[01:34:28] Parth Patil: Yeah, I think so.
[01:34:28] Kazuki Nakayashiki: A- also-
[01:34:29] Parth Patil: Oh, yeah.
[01:34:30] Kazuki Nakayashiki: Oh, yeah. One more question. So, when you make, you know, your AI clone or AI, you know, digital clones-
[01:34:35] Parth Patil: Yeah.
[01:34:35] Kazuki Nakayashiki: So, you need to have like your dataset, right?
[01:34:38] Parth Patil: Yeah.
[01:34:38] Kazuki Nakayashiki: But, yeah. According to our conversation, you talk to people on Clubhouse which doesn't have transcript, or you talk to people on Discord, which usa- usually, you know, data will be deleted or removed eventually.
[01:34:50] Parth Patil: Mm-hmm.
[01:34:51] Kazuki Nakayashiki: So, how do you keep, you know, your ideas? Also, do you have any tips to, you know, save your realtime data for your AI agent?
[01:34:59] Parth Patil: I don't do that level of, um, collecting data.
[01:35:03] Kazuki Nakayashiki: Okay. Not yet.
[01:35:03] Parth Patil: I think, um-
[01:35:05] Kazuki Nakayashiki: But-
[01:35:05] Parth Patil: And I think it's not gonna be perfect copy because of that, but-
[01:35:08] Kazuki Nakayashiki: Mm-hmm.
[01:35:08] Parth Patil: I think the goal is a little bit more like I don't need it to be a perfect copy of me. I'm, I'm me.
[01:35:14] Kazuki Nakayashiki: Okay.
[01:35:14] Parth Patil: I need it to be like, uh, you know, just like a, like a, like a different, like a, like a different kind of- like I don't care that it's not exactly like me. I care that it's helpful to me. And I think there are- actually it's good to have places where things are not recorded because then people feel free to speak. Not everyone-
[01:35:37] Kazuki Nakayashiki: Yes.
[01:35:37] Parth Patil: Like, we, we're recording this conversation, but I spend so much time just being authentically myself that I don't- I'm like, I'm pretty much the same-
[01:35:44] Kazuki Nakayashiki: Interesting. Yeah.
[01:35:45] Parth Patil: ... way, you know, record it or not. But-
[01:35:47] Kazuki Nakayashiki: Thanks. Thanks.
[01:35:47] Parth Patil: Not everyone's like that. Sometimes you put the recording button on and now people don't wanna speak.
[01:35:51] Kazuki Nakayashiki: Yeah.
[01:35:51] Parth Patil: And I think that's like, you want people to speak more-
[01:35:54] Kazuki Nakayashiki: Yeah.
[01:35:54] Parth Patil: It's more important that we speak than for it to be recorded. And, um, so I'm not a fan of like automa- always-on recording. Uh, I do use Granola sometimes like w- but that's like-
[01:36:06] Kazuki Nakayashiki: Yeah.
[01:36:06] Parth Patil: ... you know, double opt-in on the meeting transcript. And that's more for work. It's like, "Okay, did I make sure I get everything?" Um, so that's different. But, um, if we lived in a world where everything was recorded, I would not be happy.
[01:36:19] Kazuki Nakayashiki: Okay.
[01:36:20] Parth Patil: Yeah.
[01:36:20] Kazuki Nakayashiki: So, you don't like AI pin?
[01:36:22] Parth Patil: No.
[01:36:22] Kazuki Nakayashiki: I remember Humane, some AI-
[01:36:24] Parth Patil: No. I don't-
[01:36:24] Kazuki Nakayashiki: ... metaglasses. Yeah
[01:36:25] Parth Patil: I don't like it. I don't like it. The, the AI pin, yeah, th- this concept is like, um, it's just like, well, you didn't get my permission. And so, then now I'm like, "Well..." And, and even the meta glasses, sometimes I feel, you know, like people just- and I see the light, but then now I'm like, "Okay, now I'm on camera." Right? So, this is- this, this makes you ch- it changes the way you act for some people.
[01:36:47] Kazuki Nakayashiki: Mm-hmm.
[01:36:47] Parth Patil: I think, um... So, but that's gonna be an interesting social norm we have to figure out. And I think for me this is- it's more like I like talking to this system and be like, "Here's... Hey, here's what I'm thinking about." And then ask me a question and I talk to it. And then when I'm talking to it, it is collecting data. I don't need it to follow me around forever in the real world, because I think that like-
[01:37:09] Kazuki Nakayashiki: Sure.
[01:37:09] Parth Patil: That some people want that, and that's fine. Um, I'm not gonna judge them for it. Um, I think for me-
[01:37:14] Kazuki Nakayashiki: Okay.
[01:37:14] Parth Patil: ... it's more like, uh, that's not... I'm- I don't want that relationship with technology yet. And I understand, it might be-
[01:37:20] Kazuki Nakayashiki: Mm-hmm.
[01:37:21] Parth Patil: There may be very huge benefits that I'm ignoring. It's just that I think the social trade-off is serious. Um, it's pretty serious. Like if someone is- has the meta ray ban on, some of my friends have them, when they engage it now I'm like, "Okay, guys, we're on a show." Like (laughs) it's like-
[01:37:37] Kazuki Nakayashiki: (laughs)
[01:37:37] Parth Patil: ... we're no- now we're acting, right? And it's just like-
[01:37:39] Kazuki Nakayashiki: Yeah.
[01:37:39] Parth Patil: I prefer to be like, like, you know, chill, you know? So...
[01:37:46] Kazuki Nakayashiki: Mm-hmm. I see. Thanks. Yeah. I was, I was going to ask something but I forgot that.
[01:37:53] Parth Patil: No worries.
[01:37:54] Kazuki Nakayashiki: (laughs) Yeah, yeah.
[01:37:55] Parth Patil: We'll, we can-
[01:37:56] Kazuki Nakayashiki: Yeah, we can cut-
[01:37:57] Parth Patil: (laughs)
[01:37:57] Kazuki Nakayashiki: Yeah, we can cut those. Yeah, yeah. Yep, yep.
[01:38:01] Kei Watanabe: Maybe, uh, you already answered it, but, so you used GPT-5, also you used Cloud Code, and-
[01:38:09] Parth Patil: Yeah. Yeah.
[01:38:10] Kei Watanabe: ... the codeless context, so, are different. So, if you switch, so you need to have a context, right? Continuous context.
[01:38:16] Parth Patil: I'm so glad you asked this question.
[01:38:19] Kei Watanabe: Oh, yeah, thank you.
[01:38:20] Parth Patil: I'm so glad you asked this question.
[01:38:20] Kei Watanabe: I was curious.
[01:38:21] Parth Patil: Perfect.
[01:38:21] Kei Watanabe: MCP server, something, uh, context entering, how do you orchestrate or collect tokens?
[01:38:25] Parth Patil: Yeah. So, um, so, the... so, the first month I used Cloud Code, I was just learning it, and I was like, "Oh, my God." And I used it every single day, and I was like, I burned 560 million tokens. I was like, "How is this real? Like, this doesn't make any sense." (laughs) And they're like, you know, "Who?" I'm like-
[01:38:43] Kei Watanabe: Yeah.
[01:38:43] Parth Patil: ... you know, like, I'm like, $10,000 worth of tokens, but I'm not paying $10,000, so, uh, it feels like it's very discounted. But then I, and I was building like, okay, but, you know, my laptop, my compute, my g- this is my gaming computer, so I, this is like my main, my desk, but, like, some of my work is on my laptop, my work is on... then I have a Mac Mini. And I was like, "Okay, well, what if Mac Mini had coding projects on it? And then I could just connect to that Mac Mini from my phone, from my laptop, my other devices, put it on a private network. Now, we have a single chat. And then, but I can use that from any device." So, that was one thing that I was doing, which is like, um, for, like, multiple clouds 24/7 on a, on a single device. Just that thing never sleeps. I can, you know, connect to it from any other device. Very cool. Um, then, uh, you know, GPT-5 came out, and Codex CLI is now good. Well, it's like, I think Cloud Code is better than Codex CLI. I think even the people at OpenAI would admit this, but... and Codex CLI will improve. They're going to work very hard on it, and, um, you know, they're going to vibe code it with GPT-5. They're going to tell GPT-5 to add features, but you can also fork it and you can modify it yourself. And I think that the... so, now it's like you have multiple CLI agents, and you want them to work on the same project. So, then I was... this is, like, last week, I've been thinking about this a lot. Like, I want to use both of them, um, and I want them to work on the same code base. How do I do that? And I think, uh, so there's this thing called Cloud Squad.
[01:40:14] Kei Watanabe: Nice.
[01:40:15] Parth Patil: And, uh, if you just look up, it's like a GitHub repo. It's, it's a way to interact with multiple CLI tools, and then they basically use their own work trees. So, now, I'm like, "Okay, well, now I need to get, I need to change the way I work with these tools." So, instead of doing, you know, normal GitHub branching, it's like, "Okay, work trees." Now, now, it's like, can you work, can they work on different copies of the same code? And then, when, then triaging and, and integrating the solutions becomes possible. So, Cloud Squad is the most recent kind of tool that I've been using, because it also allows me to go from Codex to Cloud Code in the same tool. And, uh, and you can fork it, and then you can tell GPT-5 to modify it and build a UI on top of it, right? So, remember, everything, everything can be modified with... if you have the code, you can modify it, and you can make it even more personal. So, I think we're at the very beginning of this. I think the CLI tools are a temporary thing. I think we will also want more traditional user interfaces than, uh, than CLI. But CLI is powerful, and you get raw access to the machine. And then, a lot of the tools have CLI, uh, support. So, it's a good starting point. I have 17 MCPs, uh, that I use a lot, and, uh, one of them is this knowledge-based system, the knowledge graphs. But it doesn't, it's not just one knowledge graph. It's, it's, it's like an MCP that allows it to create new knowledge graph, query them, list them, right? So, if, if one of the tools is the ability to create and query knowledge from long-term storage. So, there is a certain ev- I think there is a set of tools, like, that you want to give some of your agents. Not all your agents need every tool, and that can be risky, right? Um, but, you know, access to your email, your calendar, it becomes so much more useful when it's like, "Oh, look at both my calendars. Plan my next three months of travel." And I was using Cloud Code to do that. Same thing with, like, like, expense reporting. I get so many AI subscriptions, right? And they all have, like, Stripe, Stripe, Stripe. So many, like, emails of, like, receipts. And then, I was, like, behind on my expense reporting, and then I, and I got an ang- email, and they were like, "Parth, you gotta file your expenses." And then, I was like, "Cloud Code?" We're late on my expense reporting. Go into my email and get every single receipt, then use GPT-4 Vision to categorize every single thing by vendor. Get the, you know, the s- total of the, the amount. Use GPT-4 Vision, look at the receipt, extract all the structured outputs, extract that information, put it in the name of the file, put them all into folders for every single month. And in 30 minutes, it does, like, 400 expense report. Um, you know, it finds all of them. There's no way I'm gonna do this manually ever again. But the realization here is that, oh, wow, coding agents are very powerful for non-programming tasks that require structured logic, right? Just general automation. So, Cloud Code does my expense reporting. Which, a year ago, I was like, "I wish there was an agent that did expense reporting." But now I'm like, "Maybe that's too small." That's, like, a side effect of coding agents is that they can do this. So, like, we need to think even more deeper about these problems, because actually, a lot of that stuff is just, you know, GPT-5, CLI, Codex, Cloud Code might be able to do it. And, um, does that mean that there is no such thing as an experience reporting agent, that it'll just be a feature? Like, who knows? Like, maybe that's just a feature, not a product. But this is, like, an early realization we're having now, of, like, the coding copilots are very useful in non-programming tasks. So, then, actually, we're... I could see, and I think that, like, that's probably going to play out more, which is that, like, the, the agents that non-technical people use will actually be under-the-hood coding agents that are just, like, guardrailed heavily, and, like, these are the f- flows that they do really well, right? If an expense-reporting agent is, at its core...Like, it's just a, maybe it's just a Cloud Code wrapper, right? Like, that's what I'm using it as. But then the same agent is also, like, the one... Because of 17-MCPs, it's doing, like, web research and it's also, you know, helping me, like, file my own knowledge away. So, (smacks lips) I think it's funny that Cloud Code, for me, for like two months, is like an everything app. Um, and that's why I'm trying to do the Cloud Squad, which is like, that way I can keep using Cloud Code, but I also benefit from GPT-5 for a lot of programming stuff. So I'm just like, okay, I'll use both of them. And I use this when I need personal assistant, this I use for co-programming. And that way I can kind of like... It's, we keep adapting. I don't think... Three months from now, it's probably going to be different, but, um, it's, it's my current approach. Yeah.
[01:44:51] Kazuki Nakayashiki: Yeah. Interesting. Yeah. A couple years ago, actually, at the hackathon, someone told me about the Cloud Squad, and I was curious about that. Yeah. Today idea, I see.
[01:45:00] Parth Patil: Yeah. My buddy at OpenAI sent it to me. He's like, "Parth, Cloud Squad, worktrees." And then I was like, "Huh, I should learn WorkTrees." And now I'm like, "Okay, I need to learn WorkTrees because this is actually very interesting and it might be a good way for multiple agents to work on the same project without undoing each other's progress." So this, like, coordination problem, it might be a part of the solution. Yeah.
[01:45:22] Kazuki Nakayashiki: Yeah. Definitely. Yeah. And I now remember my question. So, sorry.
[01:45:26] Parth Patil: Yeah.
[01:45:27] Kazuki Nakayashiki: It's about cloning, cloning yourself. So, I think in the future, I don't know, if you really want to clone yourself, and I think we should put something, the brain, to monitor your brain activities-
[01:45:38] Parth Patil: (laughs)
[01:45:39] Kazuki Nakayashiki: ... with AI pin and... Because when you're exposed to some information, a certain information-
[01:45:45] Parth Patil: Yeah.
[01:45:45] Kazuki Nakayashiki: ... you, sometimes you don't pay attention to it. You don't care.
[01:45:47] Parth Patil: Sure. Sure.
[01:45:48] Kazuki Nakayashiki: But if you record everything, the AI might assume, "Oh, you learned about this." But, you know, your brain, if your brain activity is lower, mean is up, you didn't pay attention to it. So, but if you listen to something, your brain activity is high, so, okay, this person learns about... Or listen, at least resonates with this information, so we should use the information or weigh this information when we-
[01:46:08] Parth Patil: Maybe.
[01:46:09] Kazuki Nakayashiki: ... answer questions.
[01:46:09] Parth Patil: You might be right. You might be right.
[01:46:10] Kazuki Nakayashiki: Yeah.
[01:46:11] Parth Patil: I, I, I... Look, I love it because, like, I'm a cyberpunk kid. I love cyberpunk. I don't know if you've ever played Cyberpunk 2077, but that's one of my favorite games. And they have a lot of, like, human augmentation and, like, the whole game, you have a brai- you have brain chips. Like, brain chips are like... The g- I play the game and I'm, like, buying brain chips, you know? (laughs) But I think about, like, would I put a brain chip in? Yeah, maybe when I'm, like, s- 60 years old, maybe, if it's safe. But... And then run GPT-8 on it, like th- you know, maybe, you know. Um, but I think, like, more realistically, like, non-invasive, non-invasive techniques would be a good, like, mid- middle ground, um, as long as I don't... 'Cause I don't know, like, Vision Pro, very powerful technology, but no one uses it, you know? Like, so, it's so clunky and, like, unwieldy and isolating. So, I think it's like, if they can make it... And that's why Meta is winning on the... 'Cause they made it cool, right? So, if it's, like, cool and useful, then, um, that'll be interesting. I think there might... You might be some- onto something. Like, if my glasses were like, "Oh, yeah. Parth is paying attention to this. We should probably save this for later," and it's like, "Oh, would you like me to, like..." Like, you really like this person, like, "Do you want me to, like, you know, remind you in six months to reach out to them?" Like, that could be useful, right? Yeah.
[01:47:27] Kazuki Nakayashiki: Yeah. Yeah. Definitely. Yes. So yeah, I'm optimistic about the future and technology as well. Yeah.
[01:47:33] Parth Patil: Yeah.
[01:47:34] Kazuki Nakayashiki: Anyway, yeah. Anyway, thank you so much for sharing a lot of, you know, insights and lessons, you know, you learned.
[01:47:39] Parth Patil: Yeah, absolutely. This was fun.
[01:47:40] Kazuki Nakayashiki: Learning a lot of things. Yeah. Yeah. So what-
[01:47:41] Parth Patil: No, thanks for bringing such, like... This is the kind of conversation that I love more than anything, so I appreciate-
[01:47:47] Kazuki Nakayashiki: Yeah.
[01:47:48] Parth Patil: ... the, the... I appreciate the time.
[01:47:49] Kazuki Nakayashiki: But before we... B- before any... But do you have any advice to people who are new to AI-
[01:47:55] Parth Patil: Yeah.
[01:47:56] Kazuki Nakayashiki: ... or have never coded before, wrote any single code?
[01:48:00] Parth Patil: Yeah. Yeah, yeah, yeah, definitely. Um, I think that, like, this is a great... This is probably one of the most interesting moments in human history that we get to live through this transformation. Um, there are... And, and I think it's like, it's a huge leveling of the playing field. We now have, uh, we have copilot systems that are smarter than many of our friends, smarter than most of the people that we know, and they can help you learn any topic. I was not an engineer two years ago. Now, I think, I think I can safely say that I am an engineer. Um, you know, I'm not perfect in any ways, but because of language models, I'm able to teach myself almost any topic I want. That means that, like, if you want to get good at something, you can. You can't get good at everything. You have to pick a few lanes. But it's really important to play- use the technology, figure out what you want to do, like, like your own... You might have, you have to have a vision, you have to have an interesting life and a perspective. You have to, like, go live an interesting life so that you're like, "Oh, I want to do this. I want to make that. I want to..." So when you realize what are the things that you want to make, AI is going to help you make those things. And you would be surprised at how quickly you can get moving on your own ideas. And I think, also, don't only, uh, use it for work, because that's a... Like, what if your company fails? Like, what if the company isn't around a year from now and then you wasted, like... All your energy was just trying to turn this into money, when actually, maybe, you could have made your own projects, your own life better, your family, your friends, like, your passions? And then the things that you're passionate about, when you apply AI to that, then it doesn't feel like work, right? Now, it's like, "Oh, I'm playing." Like, "Now I get to do this. This is going to be fun." And then you go away further, and then you figure things out, and that's like... I think, apply it to your passions, use the technology. Like, it's, it's, it's not enough to talk about it. Like, that is not... Use the technology. Then that way, you're like, all the noise kind of fades away when you're like, "Oh, this is, this is what it actually is. This is the part where it's very good. This is where it's not good." And the only people who are figuring that out are the people using it. And there's, there's not really a textbook or expert, so don't look for certifications. The best certification is, like you said, hackathons. Like, go make something-And then just share that you're working on it. And you'll be surprised, people will come to you with, like, "Oh, yeah, I'm also working on this." Now you have collaborators, you have peers. So, you know, you want to make things, share and talk about them. Not everything is going to be a company, that's fine. A lot of things are just, like, you know, think of it as art. You know, I think of sometimes like, "Oh, yeah, this is, like, a beautiful thing that we just did," and for fun. And then, but you create a conversation and learn from who, the conversations you have, meet people. Then that's going to build your network. Then you get the network intelligence, that's going to give you the sense of like, okay, cool, we are going to be good. Like, the n- it's, it's uncertain. Everything is, everything is changing very quickly. So we have to adapt. The more adaptable you are, the faster you learn, the better off you're going to be. And s- sometimes you don't have energy or time to adapt, that's fine. You have friends, you can kind of do this, it's a team game. And I think that, um, yeah, that's, that's like... And, and do things, don't, don't just apply it to work. Like, you might find that the most interesting passions are now possible, things that you were putting off are now possible because of these tools. So you should, uh, you should apply it to the things that you care about. Yeah.
[01:51:17] Kazuki Nakayashiki: Yeah, totally, yeah, 100%. And sorry, this is the very last question.
[01:51:22] Parth Patil: Sure.
[01:51:22] Kazuki Nakayashiki: So since GRASP is where people share, you know, what they're reading, learning, as their digital legacy, and people can actually create an AI clone through learning process. But so we want to ask this question to you. So what legacy or impact do you want to leave behind for future generations?
[01:51:44] Parth Patil: Um...
[01:51:45] Kazuki Nakayashiki: Sorry, it's a tough question, at the end, but yes.
[01:51:48] Parth Patil: Uh, I hope that people who interact with, um, anything that I have done, uh, for if you, if I've, if you've even watched just a video that I have put out, um, and I don't put a lot out, but, like, if anything that, you know, something I made, I hope that you hope, I hope that you realize like, it's like, um, just you should, like, it's, it's like... Hopefully it inspires people to just make things and learn-
[01:52:12] Kazuki Nakayashiki: Yes.
[01:52:12] Parth Patil: ... and to challenge the notion. Like, there is no, like, we have so much to rebuild, we have so much to make, so much to create, and make things. I think, like, actually make things. That's very exciting to make things and to share those things. Um, and hopefully it makes people more creative and ambitious, ambitious as well. Because a lot of the stuff that I've, I've realized a lot of stuff is a lot easier now. I thought it was gonna take 10... Three years ago, a lot of these capabilities were never... I could not dream of these capabilities. Now you can do in three hours what used to take three years. And that means that we have to be more ambitious, more creative, and more optimistic too. Like, we have to actually, like, um, it's not a, it's not 100% obvious that this will end out, end well. We have to go make that, we have to make the future that we want to live in. And so, like, you have a chance to be a part of making that future. This isn't, it doesn't just happen to you. You, you are like, you are happening to the world around you. So, um, hopefully, hopefully, like, people realize that they are, they're all like, you're all the main characters, right? Like, you're, you're in the driver's seat. Like, this is, this is, like, your opportunity, right?
[01:53:24] Kazuki Nakayashiki: Yeah, definitely. And, and yeah, thank you for the beautiful answer. And yeah, thank you for joining today. And we-
[01:53:30] Parth Patil: Oh, thank you.
[01:53:30] Kazuki Nakayashiki: ... learned a lot and yeah, thank you so much.
[01:53:33] Parth Patil: Appreciate it, guys. Yeah.