Siqi Chen, founder of Runway (finance software) and serial Silicon Valley founder, joins Sam and Shaan to share wild stories from his days at Zynga under Mark Pincus, discuss three AI business ideas he’s excited about, and recount an extraordinary Christmas crypto saga around his daughter’s rare brain tumor diagnosis. He also breaks down the Stanford “Touchy Feely” class on interpersonal dynamics that he calls the most impactful thing he’s ever done in his life.
Speakers: Siqi Chen (guest, founder of Runway, ex-Zynga, ex-Postmates), Shaan Puri (host), Sam Parr (host)
Cold Open: The Silicon Valley Hall of Fame Meeting [00:00:00]
Siqi: They said, “You shouldn’t confuse revenue for success.” So I said, “Well, you guys shouldn’t confuse a lack of revenue for success either.” And then they got kind of upset. Dude, this meeting — it goes in like the Silicon Valley Autistic Hall of Fame.
Introducing Siqi Chen [00:00:15]
Shaan: What’s up — we got our friend Siqi here. Founder of Runway. Back in the day, built a company, sold it to Zynga. Built another company, sold it to Postmates. Has gone viral many times. There are a lot of people in Silicon Valley who — it’s almost like a film director. It’s like, “Oh, they’re working on a new project, you really want to know what they’re doing.” That’s you, because you do things with taste. So excited to have you here.
Shaan: Do you have any good stories from the early days of Zynga? Because you sold a company to Zynga back when Zynga was the shit. Did you work with Mark Pincus? What’s he like? Give me a good Zynga story.
Siqi: Okay, I have a great story. How I came to meet Mark Pincus is actually a great story. I love so many good Mark stories. So yeah — Zynga acquired my first company and I joined as head of their production studio.
Sam: Wait, can you give the background on Mark? He’s like a Silicon Valley OG. Did he help fund Facebook to get off the ground? Was that his first big hit?
Siqi: So he and Reid Hoffman co-owned and bought the Six Degrees of Separation patent from a company called Six Degrees. And he angel invested in Facebook — but also licensed the patent. I believe for more stock into Facebook. That patent was basically the social networking patent, right? Like, how we’re connected six degrees of Kevin Bacon away from each other.
Siqi: And I think Reid and him realized that if this patent got in the hands of Microsoft or some big company, they would be able to squish innovation by a startup by holding this patent over their head. So they bought the patent and just decided: we’re not going to use it to stop anybody. Then they parlayed it into getting extra shares in Facebook. Amazing. I think that’s what happened.
The Zynga Stories: Ninja Dojo and the Head of Product Job [00:02:30]
Siqi: So when I joined, I was a director of products. And my girlfriend — now my wife — who I had recently met, moved to China for a work thing. And I was like three months into Zynga and I was like, I’m not really feeling it, it’s not fun. So I told them I was going to resign and move to China.
Siqi: And they said, “Hey, why don’t we just give you this new job — you can report to Eric Schiermeyer, one of the co-founders, and basically be head of product for the company.” I said that sounds fun. So I did that. I reported to Eric Schiermeyer, who was a co-founder of Zynga.
Siqi: What happened is, a month into the job, Eric Schiermeyer stopped showing up to work. Like, wouldn’t return emails, wouldn’t come to work. And I was like, what? And basically that’s when I reported directly to Mark and I was the head of products.
Siqi: And the punchline of the story is — the reason he stopped showing up is, I later found out, he decided to become a ninja.
Sam: That is not what I thought was coming.
Siqi: He literally wanted to start a ninja dojo and wanted to go do ninja training.
Sam: What in the Napoleon Dynamite is this story? How’s his ninja career now?
Siqi: I think he started another social games company.
Zynga’s Data Culture and Mark Pincus on Moats [00:05:00]
Sam: At the time, what was Zynga like? Was it on top? King of the world? Or was it already on the decline?
Siqi: What’s interesting about Zynga is they hired a bunch of people who were used to being investment bankers to be product managers, because it was all about the numbers going up. Highly analytical.
Siqi: So I went in and just the amount of knowledge they had about growth — one of the things that really blew my mind that I think about a lot is when I first had a conversation with Mark Pincus early on. I asked him, “Hey, what do you think about this industry? It just seems like it’s hard to have a competitive moat, because it’s so easy to enter — you build a new game, it expands. How defensible is it?”
Siqi: And I think about his answer quite a bit, because his answer was, “No, this is great. I want more new entrants into the space because it’s free R&D for me.” And I was like, okay, that blows my mind. That is next level. Because he was just so confident in his ability to execute that anyone who’s going to come in with some new idea, they can just fast-follow it and do a much better job of growing it. Which is exactly what they did. FarmVille wasn’t the first farm game. Poker wasn’t the first poker game.
Sam: Wait, did he have like Genghis Khan energy where he was just conquering?
Siqi: He did. We were working like 80, 90, 100 hours a week at Zynga. That was kind of the norm.
Sam: It seems like a waste of talent, though, to have that conqueror energy and do it in the lamest way possible. Like, FarmVille. That’s what he’s trying to do with a few billion dollars?
Siqi: There’s a quote by Max Levchin at the time — Max Levchin, who created PayPal, is by all accounts like a genius computer engineer, single-handedly fighting fraud at PayPal with a one-man army against all of these financial scammers around the world. The guy’s brilliant. And then his next startup, I think, was Slide. And it was making like widgets for MySpace and Facebook — slideshows, or Super Poke, where you throw a chicken at your friend.
Siqi: And he goes: “Be really careful what you choose to work on, because everything can be optimized endlessly.” So once they got into that, they could just sit there and optimize and get the number of chicken slaps per day up. So he’s like, choose wisely. Should I be optimizing this chicken slapping, or should I be doing something else?
Siqi: I’ve always thought about that, because I found myself falling into the same trap. No matter what I’m doing — if I’m selling little widgets, my life becomes about selling widgets, and I can optimize that to infinity. And people do. If you go look at how any company works, that’s what they’ve done.
Sam: Dude, when my wife worked at Facebook years ago, I remember talking to all of her co-workers at a party. I was like, “Oh, you guys are amazing, you’re so smart — what are you working on?” And they explained it to me, but it boiled down to: they were trying to convince Brazilians to put more stickers on their photos. There was like this one dog that had a long tongue — do you guys remember that dog? They were trying to get Brazilians to use that dog sticker. I felt like a cartoon. I was like, what are you guys doing?
Max Levchin and Slide: The Meeting That Became Legend [00:09:00]
Siqi: I also have a Max Levchin and Slide story. This goes back 17 years, so I don’t think he’ll mind anymore.
Siqi: When I first moved to the city, I had this app on Facebook blowing up. And some guy who worked with Keith Rabois — Keith Rabois was, I think, the CEO of Slide at the time — said, “Hey, we like your app, you want to meet the team?” And I said, “Oh my God, this is amazing.” I mean, I’m out of college, about to meet Max Levchin.
Siqi: So I go in, we had a meeting at noon. I go to the Slide office and no one is there. Everyone’s out for lunch. I talk to the receptionist, I say, “I’m here to meet Max and Keith.” She’s like, “They’re at lunch, they’ll be back in half an hour.” I sit there for half an hour, and then they walk in and they act like they kind of forgot there was a meeting. But we go to the meeting room and we start talking about what I wanted to do with the app — build it independently or maybe join Slide.
Siqi: They asked how much I wanted for it. I said, “Well, I have a co-founder, so $2 million — a million each.” And they kind of laughed at the number. I was like, what? I looked at our app data and the amount of money we’re generating — it’s not a huge amount but it’s fine. It’s not too much.
Siqi: And they said, verbatim: “You know, we bought Super Poke, and he’s easily the 18th or 19th most important person in the company now” — out of a company of like 60 people. I said, “Okay, that’s a compelling offer.” And then I said, “Yeah, this is probably not gonna happen.” I said, “We’re profitable, we’re making a bunch of money from ads, so it’s cool if you don’t want to buy it.”
Siqi: And they said, “You shouldn’t confuse revenue for success.” At this point I was just really upset. So I said, “Well, you guys shouldn’t confuse a lack of revenue for success either.” And then they got kind of upset.
Siqi: But I forgot to mention — at the start of the meeting, they were saying something like, “Hey, there’s a recession coming,” like talking to each other. And then one of them said, “I can’t wait to buy all these shitty companies for cheap” — and we’re in the room. Like, what the hell?
Siqi: Dude, this meeting goes in the Silicon Valley Autistic Hall of Fame. And then at the end, Keith was like, “Hey Max, what was your body fat percentage in the PayPal days?” And Max was like, “8%.” And it’s just like… I’m not the fittest person but I’m not selling you my company because you’re fitter than me. Like, what is happening?
Siqi: But it all made sense the next week when I was talking to my friend John, who also had a company building Facebook apps. He comes up to me at some party and goes, “Siqi, you won’t believe this — I just had a meeting with Slide last week.” And I’m like, “Really?” He goes, “Yeah, it was so weird. They started the meeting talking about how there’s a recession coming and they’re going to buy all these cheap companies.” And I was like, oh my God. It was a script.
Shaan: It was in episode one of this podcast I think. My buddy Sully told almost the same story. He gets invited to Slide, and what they said was, “We want to hire you.” And he’s like, “I don’t really want a job. I thought you wanted to buy it or something?” And then one of them opened the door and goes, “You see all those guys out there? They’re going to build your app in six days if you don’t take this offer.” And he was like, okay, I’m not interested in this even more now.
Siqi: His app at the time was called Superlatives — like, you would name your friend “most likely to go to jail” or something. And he’s like: “You’re threatening me that you’re going to take my app? I think my app is stupid. The fact that you think my app is cool makes me think you’re stupid.” That’s what he thought in his head.
Who Became Someone? Early Silicon Valley Encounters [00:13:30]
Sam: Siqi, I love talking to people who have been around a bunch of these folks before they kind of “made it.” Who is a tycoon or big shot now that you’re shocked by — because of when you knew them when you were both younger?
Siqi: I don’t know that I met a lot of people who weren’t great who later became huge.
Sam: Who’s someone you knew early — and why did you know they were going to be something?
Siqi: Drew Houston — Dropbox — it was 2007, early 2008, and we were just talking about growth. They had this college referral plan. Anish Acharya, who’s a General Partner at Andreessen now — he was just a founder building apps on mobile phones, right? He would come to my office and we’d talk about it.
Siqi: Probably the best one is Chris Wanstrath, who’s the co-founder of GitHub. He was a contractor at Posterous — this job I took when I first moved to the city. He was actually a contractor for my first company, Serious Business. I had him build a translation layer between Facebook Markup Language and my code. Just a contractor. And he was really good at it.
Siqi: I distinctly remember — we were at 21st Amendment, or actually the Anchor brewery, doesn’t matter — and I gave him a full-time employment offer in early 2008. He was considering it. And the same night, while we’re talking about it, he said, “You know, DHH just put Rails on GitHub, so I think I might work on GitHub full-time.” And I was thinking in my head: the smart thing for me to do is actually shut down my business and go work for him. But I didn’t say that.
Sam: Did you have any massive angel investment wins? You’re around the hoop for so long and so early.
Siqi: Yeah. The best one by far is Amplitude. I was introduced through Matt Akko, who runs Data Collective now — he introduced me as the very first angel after I sold to Zynga. And that was like 10x in about 12 months, so I was playing with house money after that.
Siqi: Then he introduced me to this team that we ended up being the fourth customer of for their analytics platform — it’s now Amplitude. We used it, we thought: this is great, this is the best thing we’ve seen since the Zynga internal tools. And I managed to get into the seed round. That IPO — we ended up with a 400x return. Pretty great.
AI Business Idea #1: Smart Email Management [00:17:30]
Shaan: Let’s jump in with ideas. You’re an idea guy — what do you think are some cool ideas or opportunities you’d want to be working on right now?
Siqi: So I built a Zap, and it basically categorizes my emails with GPT-4 into different labels. I also use a product called Sanebox, which does something similar filtering emails. But there’s been no good version of email management software that I’m able to customize with prompts and train.
Siqi: I want to be able to say: “Here are things related to my kids.” There’s no keywords necessarily — you kind of have to read it. It’s like some email to my babysitter, or an email from school. If it’s related to my kids, I want you to put it in this folder. The idea is I want to have different inboxes for different contexts.
Siqi: People are doing different kinds of email categorization. What’s missing is there’s no way to train it. I think building some way for a product to understand all the context about your life and organize your stuff — starting with inbox — will be really powerful.
Siqi: I’ve seen at least 12 companies do this and no one has done a really great job of it.
Sam: I just tested this product — you guys heard of this? I can’t think of the name right now. But it was this thing where it recorded my screen for weeks at a time and would see how I’m typing, what I’m saying to people, and give me feedback on the productivity of my day and how I can improve it. Have you guys seen this? Started with an R, I think.
Shaan: Is it Rewind?
Sam: Rewind, yeah, yeah. Rewind.ai.
Siqi: They stopped working on it — now it’s called Limitless, and they’re working on a pendant now.
Sam: Yes. I was tinkering with Rewind, and the promise — it’s not there yet, but the promise was amazing. I was so into it. And it’s kind of describing what you were just explaining, because we’re having to use Zapier and OpenAI or ChatGPT to duct-tape this all together. Their premise was amazing. That is exactly what I’m looking for.
AI Business Idea #2: AI Fake Dating Apps [00:19:00]
Siqi: I’ve got a third one, which is a lot spicier. You know, like this is not a venture-backable thing, but — you know Character AI and Chai and all these chatbots that people are really using for, you know, sexy chats? These things print money, by the way. They immediately generate millions of dollars a month. I thought, if I were just in it to print money, the thing I would make is something kind of like Tinder but basically everything is AI-generated. So it’s not like, “Oh, you’re creating a robot.” It’s a fake dating app where everyone is attractive and super interesting. And then you can go off of your fake Tinder app and go on Instagram — there’s a real Instagram account owned by the person you met on your fake Tinder app.
Sam: Siqi, do you know who Tai Lopez is? He’s the guy who had the infomercial that was like, “Here in my garage…”
Siqi: I do know who Tai is.
Sam: So Tai — he came on MFM years ago. And an accusation that I learned about based off comments on YouTube is that years and years before he was whatever he became famous for, he owned dating apps. And the accusation was that all of the users were completely fake — guys in the Philippines running them — doing exactly what you’re describing.
Siqi: That’s true of the majority of dating apps. What do you mean? That’s just like standard operating procedure. Like Plenty of Fish, Seeking Arrangement — that is their business model. They have men acting as women. And there’s also the webcam industry. Up until OpenAI, the companies with the most advanced AI technology were probably these companies, because they are better at creating chatbots than anyone. They have the most advanced technology — I’m not kidding, I know people who work there or run it. They’ve just developed better chatbot technology than anyone.
Sam: That would be so funny if that’s where AGI starts. It’s not OpenAI, it’s not any of the labs — it’s whatever some webcam site developed.
Shaan: That’s not a crazy idea. Isn’t that how it typically has been, where the vice industries are the ones pushing the envelope?
Siqi: Yeah, right. Like, it was transformative for OnlyFans too. If you’re a big influencer on OnlyFans, you have an army of 100 people who are like typers — chatters. But they’re not just chatters flirting. They’re also basically salesmen. What they’re doing is chatting to upsell you: buy this video, buy this photo, buy the subscription. They’re not just customer support — they’re actually sales. But they need to come across like they’re the original person. Which is just hilarious.
Sam: What do you think that office environment is like?
Siqi: I actually know someone who works in that, and it’s just normal. It is the most boring cubicle, normal office environment you’ve ever seen.
Sam: That’s almost better, it’s weird.
Sam: The other thing with dating apps — I remember seeing a study about Match.com. If you were a guy on Match.com, you’d basically send out 30 messages and get one back. And what they realized is that a lot of guys’ accounts would go inactive because they’re not getting replies. So what they would do is show an active person 30 inactive profiles, knowing that inactive person is not there to reply — but that will be the notification for that person to come back and reactivate their subscription. They have to pay to go read their inbox and be able to reply. So it’s almost intentionally a horrible experience for the person who’s there trying to find somebody, in order to reactivate all the churned members. In the first 100 matches they’d show you, something like 50 to 70% were all just inactive people they wanted you to send a notification to.
Siqi: That sounds like a Zynga-trained PM.
The Zynga Office: LED Tunnels, Dogs, and PM On-Call [00:22:00]
Shaan: Dude, I went to the Zynga office once back in the heyday. It was the craziest office I’ve ever seen. Sam, did you ever go?
Sam: Yes. It was right — it shared a building or was next door to Airbnb. It had a huge bulldog in the front. And I think at one point they owned the building, and the building was worth more than the company. Like, hundreds of millions of dollars.
Shaan: When you’d go in, there was a giant tunnel — like an LED tunnel — you’d walk through just to enter. And when you’re there, first of all, there were dogs everywhere. It was bring-your-dog-to-work. Just herds of dogs running around. It was insane.
Shaan: And we met the chef, and the food operation was more sophisticated than any company I’d ever been a part of. The cafeteria was better than my actual company. They had a staff of 60 people on the culinary team. They had a roof where they were growing vegetables. They had a giant fridge the size of a swimming pool — you’d walk in and there were cows hanging upside down because they had their own butcher process. They didn’t serve soda — they brewed their own sodas. Everything was soup-to-nuts custom. I was like, if the food is like this, I don’t even want to know what the teams that actually do the work are like.
Shaan: And I walked into one PM meeting and it was like a stock market. His screen had so many metrics, fine-tuned in real time. They were running so many tests simultaneously. It looked to me from the outside like the most data-driven operation I had ever seen. What we did looked like cavemen compared to what they were doing in terms of data sophistication.
Siqi: The food thing reminds me — I was the person who lobbied for us to serve real bacon. After a few months we finally started serving real bacon, because we never had real bacon before. It was only turkey bacon. Mark Pincus didn’t like to kill pigs.
Siqi: One of the things people did at Zynga in the product org: we had PM on-call. And I’ve never seen this at any other organization. The PM on-call for every game would daily send an analysis of what changed day over day. So if there was a drop, you would segment it. “Oh my God, there’s a 50% drop in Mexico for FarmVille and we’re not sure why, because usually on Wednesdays at 2pm it shouldn’t be like this, and it lasted 3 hours.” They would have to explain: “It’s because the World Cup is happening.”
Shaan: What comes first — we were with MrBeast a few weeks ago and I had the same question, but he wasn’t able to articulate or answer it. What comes first when you’re that data- or analytics-oriented? Are you that way, and scale comes because of that? Or can you only behave that way because you have scale?
Siqi: It depends on the environment. If it’s a data-friendly environment like Facebook — where virality lets you grow from zero to a billion — then data is all that matters. But you can see in the experience of Zynga that that didn’t translate to the mobile industry. Mobile was less about virality, it was just difficult to get redistribution. A lot more about creativity. That’s why Supercell had such a huge advantage — they actually built very fun, new games.
Siqi: So I think it’s completely environment-dependent. If it’s easy to get early distribution, then data makes the difference at scale. But if it’s quite difficult to get early distribution, then you want to be creative and innovative — brand and creativity matter a lot more.
Sam: Do you remember any random game change — like color red to blue, or flashing lights — that just generated like $10 million overnight?
Siqi: I always remember thinking that the best way to generate $30 million overnight is to just say, “Hey, it’s going to take you $10 to unlock the game today.” People would pay.
Siqi: One of the more interesting ideas is this concept of crew. We had this idea of collecting materials — you’d aspire to materials — and one of the mechanics we invented at Zynga is the idea of “crew.” For whatever thing you want to unlock, you have to get at least 20 unique people to help you. Because what we saw in the data is that when you do materials, you ask the same two people over and over. But if we required a unique spread, that increased distribution and it ended up being a pretty large boost in engagement.
Siqi: Another thing we saw was just the power of segmentation. There was this one day where our numbers went down anomalously and I had to figure out why. It was a particular typo bug in the drop rate of a particular treasure that was creating opportunities for people to share. You just spend all day doing things like that. It’s rarely something huge — it’s all the details added together that make a difference.
AI Business Idea #3: AI-Powered Research Behind Paywalls [00:28:00]
Siqi: I’ve got a third one, which is a lot spicier. But on a more serious note — I was doing a bunch of medical research and I noticed that you can’t access paywalled articles. So I really wish someone would make a version of Deep Research that lets you enter your paywall credentials so you get full text access.
Siqi: And this goes further — I think there’s just so much data behind paywalls that you can’t get to. You can only search the open web. The more private data you can get access to, the more useful these agents become. So that’s probably the number one thing I’ve been thinking about.
Sam: What’s the name of the company — starts with an R, based in England? We’ve talked about them a bunch. It’s like an acronym. An academic publishing company.
Shaan: Elsevier? I think Elsevier is the biggest one.
Sam: Yeah, it’s really controversial because it has like the second highest profit margin of all publicly traded companies in the world, behind public storage. And the shtick — and why everyone hates it — is that researchers at universities don’t get paid anything for the research. In fact, sometimes they have to pay tuition to even be at these places. But the company takes your research and puts it behind like a $30,000-a-year paywall.
Siqi: That’s the whole industry, right? Elsevier is the biggest one. All of these are extremely annoying businesses — they charge an arm and a leg for access, the researchers get paid nothing, the peer reviewers get paid nothing. All they’re doing is taking the text and putting it somewhere and maybe printing it into a journal.
Shaan: So how would this work? You told Sam Altman: “Hey, can I just give ChatGPT my credentials and then it goes and logs in for me and uses that information when I ask it questions?” How would a founder do this?
Siqi: You’d need access to O3, the full model, first — because that’s what OpenAI’s Deep Research is based on. But it’s not terribly difficult to create something like a Deep Research. There’s a company called Jina — I’m friends with the founder. He used to be VP of search at Baidu, which is the Google of China. And they made a version of Deep Research just using a different model. So that’s relatively easy to build. The storage of credentials and logging in is a little more tricky.
Sam: It also just seems like somebody could create like “SciencePal” or something — specifically for researchers, a ChatGPT-like interface but specifically trained on or going to access all the journals and academic research. Just do a vertical thing.
Siqi: So I’m deep in this space. There’s a company called Elicit, there’s a company called SciSpace — they let you search abstracts where available, and you can upload individual PDFs too. But there’s nothing I’ve seen that indexes the full text of all these journals with your credentials.
Shaan: Why are you so interested in this? What’s all the research for?
Siqi: My daughter was diagnosed with a rare brain tumor last September. And we’ve been doing everything we can to find new treatments for it because it’s so rare. You really get deep insight into the cracks in the structure of the medical community that way.
Siqi: One of the big learnings is that what is available as standard of care — meaning what’s available if you go to a hospital or talk to a doctor — and what is available at the frontier, there’s a huge gap. And then further, when you have a rare disease, the amount of research and data and treatments available is also just thin, because you need enough critical mass for the research to be worth it to recoup the research costs.
Siqi: The other interesting thing we’ve learned is that the IP issues are really weird. For example, there’s this drug that’s been FDA approved, non-prescription, since the 70s, to treat pinworms. And over the past 20 years there’s a huge amount of compelling data that it might be a pretty good treatment for different kinds of cancer. And there’s been no clinical trials for it, because there’s no money for it, because you can’t patent a pinworm drug. All the money — hundreds of millions, billions of dollars — is going to new patentable molecules, even if things already available might work.
Siqi: So I’m doing a lot of primary research to find repurposed drugs that might already exist that could treat this very rare brain tumor.
Doing Your Own Medical Research as a Founder [00:33:00]
Sam: It’s an interesting logistical thing — you’re able to just research potential cures or ways to help your daughter on your own and come up with a solution? Is that what you hope the outcome is?
Siqi: I’m not the first founder type who’s been in this situation. The co-founder of Clubhouse, Rohan Seth — his daughter has a rare disease. Her name is Lydia. He runs the Lydia Foundation and has gone so far as manufacturing his own drugs for very rare diseases. You can do a lot.
Siqi: This is what I mean — if you’re sufficiently motivated, and there’s no one more motivated than a dad with a sick kid, you can go so much further than what is available as standard of care.
Siqi: Even the last time we met with our primary care team, they were proposing two particular paths — one involving pretty aggressive surgery and radiation, another involving this drug. I proposed a third path, and they were discussing it — neurosurgeons and clinicians. They came back a week later and said, “Actually, your path makes sense, it’s worth trying.” And the reason I was able to do that is that this disease is so rare that I am more knowledgeable about it than anyone in the room, because they have to study 50 different cancers.
Sam: Wow, that’s pretty incredible. My experience has been that doctors, once you get off standard of care — I just had a knee injury and I was asking my doctor, “Hey, would PRP or stem cells help? Is there a peptide? What about that?” And he was sort of like, “My hands are tied. I can only recommend what I can recommend.” So I felt like I was on my own. Once you do this path, are you basically outside of the standard medical system?
Siqi: That’s a great question, and I’ve had an almost identical conversation. The neurosurgeon said, “When we did this operation, we did it for this reason — doing it for some other reason is not part of standard of care.” But our primary clinician, who is a principal investigator of one of the only two clinical trials that treat this disease, said, “Yeah, you’re right — we don’t have a lot of evidence because there hasn’t been a lot of research. So sometimes you have to argue from first principles. If it’s a rare disease, they’re more open to being creative.”
Siqi: In our case, our clinician was able to convince the neurosurgeon this was the right path, or at least worth trying. And she told us she was excited to have a partner who seems well-informed and is willing to think outside the box. I said straight up, “I do not care what the standard of care is. I think the standard of care is crap.” And she basically said, “Yeah, I think so too. I just can’t say that.”
Siqi: To answer your question — if you do find your own drug, you do have to show enough research and be well-informed enough that someone is willing to prescribe it off-label on a compassionate use basis. You just need to convince one doctor. And if you can’t find one, you can find others. It’s a lot easier if it’s a fairly serious, rare disease.
The Crypto Fundraising Saga: Mira Coin and Pump.fun [00:31:00]
Sam: Can you tell the story about — you tweeted about your daughter’s condition, and then this crazy crypto chain of events happened where someone created a coin, millions of dollars were raised, people got mad at you. I don’t even fully understand it. Can you explain what was going on? And also — is this just degenerate gambling, or did you stumble upon a novel way people might fund research in the future?
Siqi: I think I have a better idea which one it is after some reflection. So okay — it was Christmas. I was going to Japan to ski with some friends and the family. I was on a flight and I had a plan to start a GoFundMe for a lab at the University of Colorado — the H lab — because they’re the only lab in North America that researches this particular tumor called craniopharyngioma. And the treatment we’re on, they found. So I thought, for Christmas, it’d be great to do a GoFundMe, use my network, raise some money for this lab.
Siqi: So I tweeted a thread with this GoFundMe link and ended up raising about $4 million to the lab, which is I think the biggest donation they’d had up to that point.
Siqi: But what happened is some people started asking, “Hey, can I donate crypto?” So I posted my ENS — my Ethereum Name System address — and people started donating a little bit of ETH. And then some other people said, “Hey, do you have a Solana address?” I didn’t have a Solana address. I was aware of Solana but had never touched it.
Siqi: So the next day, once I landed, I created my first Solana wallet, tweeted the address out. And basically an hour later, I looked at my wallet and it had $400,000 in it. It was zero an hour before.
Shaan: Why were all the crypto guys doing this for you? What’s the motivation?
Siqi: I found out why later — this answers the question of whether it’s gambling or something else. But anyway, I posted a Solana address. Someone had created a coin on pump.fun — which is a platform where you can create a coin in literally like 30 seconds.
Sam: Wait — pump.fun. Have you guys seen this business? These guys made like $500 million in profit last year.
Shaan: More than that. And this is like the majority of the traffic on Solana. People are basically trading these new coins trying to ride them. It’s musical chairs. You’re buying trying to catch one of these thousand-X waves but it’s going to dump and you just have to know when to get off the train.
Sam: It looks like a GeoCities website. There’s like flashing banners. It’s all real-time. Yesterday during the Super Bowl, somebody created a coin instantly for Dave Portnoy — “Jail Stool” — and it ran up to like $100 million market cap or something.
Siqi: Right. So people were looking for new narratives to gamble on. And the more interesting and viral-seeming the narrative, the more valuable it becomes if you’re in early. So anyway, I posted my Solana address and about an hour later I looked at my wallet and it was $400,000. It was zero an hour before.
Siqi: I tweeted a screenshot and kept updating the thread. So it was $400,000, and I was like, “What is happening?” It turns out someone created a coin called Mira on pump.fun. There were a billion tokens, and they bought half the supply and just sent it to my wallet. So I had 500 million tokens. And once I tweeted, I checked again and it was $4 million. An hour later it was $8 million. A couple hours later it was $15 million. At one point it was $20 million in the same day. The peak market cap of the whole thing was around $60 million.
Sam: And you had 40 million of it?
Siqi: Yeah, because I sold 10% immediately just to capture something, and I sold a bit more in the liquidity pool. So I got around a million dollars out. But I still owned 30% of it. And immediately I said: every dollar in my wallet is going to charity — to the research lab researching potential treatments for my daughter’s tumor.
Siqi: So I said, “I’m not going to move anything without 24 hours notice, and I’m going to try to be very transparent about this.” And I thought: I’m going to start selling $1,000 every 10 minutes until we’re done, because I just don’t have time to run a crypto project and I want everyone to know where this is going.
Siqi: That happened, and the price started going down. But because it was such a big story, a bunch of rare disease organizations started reaching out — “Wow, this happened to you. How do we get in on this?” Average annual budget for one of these organizations is like $100,000 to $150,000. This was more money than the community had ever seen.
The “Zero” Coin Mistake [00:38:00]
Siqi: So I was asking some crypto friends: how do we make this a sustainable thing? The idea of turning Mira into a launchpad for other rare disease tokens — where Mira is a liquidity pair token — was talked about. So I thought, before I do that, I should probably figure out how a coin works and how you launch one. So I went on pump.fun and I thought: let’s find out how one of these works.
Siqi: I created a coin called “Zero.” I entered a description that said: “Don’t buy this coin. It literally says don’t buy this coin. It’ll never be worth anything. I’m never going to do anything with it. It’ll be worth zero.” I pressed the button.
Siqi: What I didn’t realize is — because I’d been on Ethereum for a long time, I had what’s called a “watch wallet.” People track whales: what is Vitalik doing with his wallet, what is such-and-such doing. They think you’re the man because your wallet is old or you have a lot in it. And because I was the main character of crypto Twitter for a few days, there were bots just monitoring what I was doing and copying whatever I was buying.
Siqi: So within 100 to 200 seconds, the market cap of this token I told people not to buy was three and a half million dollars.
Sam: Oh my God.
Siqi: I was just sitting there panicking. I said, “Okay, I don’t want anything to do with this.” I had half of the entire supply and I sold it. That was my main mistake. I should have burned it — everyone would have been fine with that. But because I sold, I crashed the price of the token, and people got very upset because it was considered a “rug.”
Siqi: I made like $80,000 — even though it was $3.5 million in market cap, there was only about $100,000 in liquidity. And people started saying I was a scammer. Within a minute I realized this was a mistake, so I bought back into the coin and burned it all. Already made it neutral. But no one cared, because of the anonymity of crypto — they thought I had like a hundred other wallets that I’d pumped and dumped.
Siqi: So I decided this is really shitty and I don’t want any part of it. What I decided to do is — with some help from people at Coinbase and others who didn’t want to be named — I did on-chain analysis. Everyone who held in the first 200 seconds and sold at a loss, before the market cap fully recovered 43 minutes later — I just airdropped them Solana out of my own pocket. Ended up being about $450,000 out of my own pocket. Just paid everyone back. Which has never happened on any pump.fun coin before.
Shaan: This is like a crypto Larry David situation. It’s like you’re walking through Times Square and one of those fake monks puts a bracelet on your wrist and then expects you to give him $20.
Siqi: And maybe like 10% of people heard about this. I tweeted about it, kept responding, but every time I tweeted for the next month someone would say “you’re a scammer, what are you still doing here” — and I’d say “I paid everyone back” and they’d just not say anything.
Sam: Would you have given back the $150,000?
Shaan: I would not. I would not have for that situation personally. I get why you did it, but I don’t think you needed to. You created a coin called Zero that you said “don’t buy this, this is going to zero, it’s a test coin.” If somebody went and randomly speculated on it using their sniper bots — I wouldn’t have given a shit personally. But then again, this community is so crazy that they’ll just make your life hell on Twitter for five years. It might be worth it just to clear your own conscience.
Siqi: Exactly. I couldn’t — it was so I could sleep at night. A lot of these people are fairly low income and the money is fairly meaningful. But another reason is: people just don’t read. I explained this. The coin description says “don’t buy it.” I explained it multiple times. What I realized is — you just don’t read on the internet. And as far as anyone knows, if they don’t read, what it sounds like is: “I created a scam coin using my own daughter’s name to scam people.” Right, right. It’s crazy.
Did the Crypto Actually Fund Research? [00:44:00]
Sam: In the end, is there anything here that’s interesting for fundraising for research? Or is this just: I accidentally got into a gambling pool and got some money, but it’s not sustainable for anybody?
Siqi: I’m still trying to figure it out. What I’m hoping to make Mira into: the way this works is there’s a game theory around — okay, you own a bunch of this coin because the narrative is attached to you. And in order to turn it into real-world impact, you have to sell. There’s no way around that. When you sell, you’re playing a zero-sum game against the community and they’re going to be upset no matter what.
Siqi: So my idea is: the only sustainable way to do this is to lock a bunch of the token into a liquidity pool. When people buy in or out of it, you get exchange fees — and that’s not really “selling into your community.” Imagine a version of Dogecoin where the market cap is a couple tens of billions, and every time someone sells or buys, you get fees that could be hundreds of thousands a day which you can then donate. I think that might be relatively sustainable.
Sam: This is insane. I don’t even know what to say.
Siqi: It was the most stressful time I’ve ever had in my life. It ruined my vacation. It was Christmas up until maybe now.
The Touchy Feely Class: Five Levels of Communication [00:46:30]
Sam: Can we do a quick detour? We were at a dinner once and you talked about some Stanford class — I think the code name is “Touchy Feely” — something about communication and relationships. You said something really great at the dinner and I want to hear it again. What’s the deal?
Shaan: By the way, Siqi, how old are you?
Siqi: I’m 41.
Sam: You look like you could be 22 or 41. I have no idea.
Siqi: Asian don’t raisin.
Siqi: So yeah, I took — actually twice — a class based on the Stanford Business School class called Interpersonal Dynamics. It’s the highest-rated and most popular class at Stanford Business School. Taught by a professor called Carol Robbins. Generally known as Touchy Feely. Named as such because every participant, at some point, will cry in the class. Carol Robbins is now a co-founder of a group called Leaders in Tech, which revises the class for tech leaders.
Siqi: One of the things taught in this class: the purpose is to teach you how to relate to people and build connection with other people, because people work with other people. And one of the most useful frameworks I got is how to think about your connection with other people and how to develop that connection.
Siqi: The two frameworks are: two tracks of interpersonal communication and the five levels of it. When you’re talking with anyone else, there are two tracks — a content track and a relationship track. The content track is filled with facts. The relationship track is filled with emotion. And the relationship track is what has to be filled for a relationship to get closer and for trust to increase.
Siqi: The way you fill each of those tracks is through the five levels of communication. There are five levels of increasing vulnerability and depth.
Siqi: Level one is “ritual” — “Hey, how’s it going?” Just a ritualized greeting. Level two is “extended ritual” — how’s the weather, how’s the gang. A longer version of hey, how’s it going. Level three is “content” — facts. How’s the project, is it late, what are we going to do with this idea. Level four is “emotional self-disclosure” — when you say something that discloses how you’re feeling emotionally at the time. “I feel sad, I feel angry.” And there’s a lot of talk about level four because people think they’re doing level four but they’re not — that’s actually quite unique to the English language, which we’ll get to. Level five is the deepest: “mutual emotional self-disclosure.” It’s when you express emotion you have about the other person. “I feel angry at you. I feel proud of you. I feel disappointed by you.”
Siqi: The content track is only filled by levels one through three. The relationship track is only filled by levels four and five. And we are taught — we are conditioned — to really not use level four and five in professional settings. But if you want to build a relationship, level four and five is the only way to do it.
Siqi: So a lot of the training is about breaking past the discomfort of engaging level four and five. You basically sit in a circle with 12 people for four days straight until you can observe the impact of doing level four or five — and not doing it — and how you’re able to be closer or further from someone in emotional distance.
Sam: Has this made your running of a company better?
Siqi: I would say this is the most impactful thing I’ve ever done in my entire life. Out of any class. I always saw the company as some kind of machine. And I found it difficult to relate to people. But it completely transformed all my relationships — including my relationship with my wife.
Siqi: Even just last week, we had an onsite and I was able to do a mini version of this with our customer success team. We just sat and talked for about four hours — a very condensed version of this. And the amount of closeness people got was transformative. You wouldn’t normally sit around for a couple hours in a work setting and talk about your feelings. It’s very uncomfortable to do. But it’s intentionally so. In the case of the real workshop, it’s half a day. We were speed-running it and it was uncomfortable for about an hour. But then people were really into it. At some point everyone was crying about some disclosure they’d heard or experienced. And as a result the team got so much closer and trust increased.
How to Actually Have Level Four and Five Conversations [00:54:00]
Sam: How do you do this in practice? Because when you talk about “I feel angry at you about X” or “I’m disappointed about Y,” I could see myself not having the skills or finesse to let the end result be a positive one, versus: we start talking, you’re upset, the other person gets defensive. Can you give me an example of a conversation that typically would have gone badly — or been avoided altogether — and instead here’s how the actual conversation went?
Siqi: The first thing is easy: most people just don’t have the conversation. You wouldn’t say “I’m angry.” You’d just be angry and you wouldn’t say anything. But people can tell — when you feel a certain way about someone, it leaks. Like, even if you’re not intending to be passive-aggressive, you’re just kind of ignoring the person. Or it comes off like: “Oh my God, it’s late again,” right? And so that’s the default — you have this negative feedback cycle where you felt a certain way, expressed it unknowingly, and now the other person thinks you’re angry at them, so they dislike you more, and then they do things that you dislike more, and it just gets worse. That’s how all relationships get worse.
Siqi: So if you know it leaks anyway, it becomes easier to just say it. Your choice is: do you express it with words, or do you express it with not-words but just passive-aggressive behavior?
Siqi: And then you combine that with: everyone is entitled to the same facts as everyone else. But you’re not entitled to read other people’s minds and tell them what they’re thinking. However, you are 100% entitled to share what you’re feeling — because those are facts to you. That’s your reality.
Siqi: So the mental model is — and this is typical because people aren’t used to expressing this — if you say something like: “When I see you do this, the story I tell myself is that you don’t respect me. I don’t know if that’s true. But this is what I’m thinking in my head. And because of that, I feel angry. I just want you to know that — I know you probably don’t because you can’t read my mind. And I’m guessing you probably aren’t intending to make me feel that way. I just thought it would be helpful to share that.”
Sam: I just learned that technique in therapy last week, actually.
Siqi: Nonviolent communication framework, right? It’s very connected to that.
Sam: Do you guys do this where you get into something — whatever you want to call it, the touchy-feely stuff — and you’re like “this is the way,” but then half the time you execute it poorly, and then the business suffers, and you go right back to the total opposite where it’s like everyone has 15 minutes to fight for their job? Like, I get influenced by either side and there’s no middle ground.
Shaan: The version that happens for me: I hear this and I’m like, “Siqi just taught me something, this is great. Two content tracks. Five levels. I’m going level five, baby.” Then I go have the next conversation with my wife tonight, and I’ll give her — “I feel… no no, I feel upset, and I know you didn’t mean that…” And I try to do the whole thing. And she’s like, “What?” She doesn’t know all of this. She didn’t go to the seminar. She doesn’t have the tools. And so she doesn’t play back the roleplay that I’d imagined. And then I’m like, I don’t really know the next move. Revert back to old self.
Siqi: That’s not a bad response, honestly. Because you have to do whatever works. And the reality is, to get good at this — I did this four-day program twice. Every day was like 12 hours. You’re sitting in the circle practicing with people.
Siqi: The first time, she actually — I kind of smuggled her into the hotel room so she didn’t go to the class. But the second night when I went home she was like, “Who are you?” Because I was like, “Oh my God, I feel so bad, I’ve been such an asshole.”
Sam: This sounds like — have you guys heard of the Hoffman Institute? I’ve contemplated going. You go for four or five days, can’t bring your cell phone, completely disconnected. Everyone who goes won’t tell me what happens, but they all say it’s lifechanging and they can now develop relationships and connections with other people. I’m so tempted to do it, but the amount of time to be disconnected is scary.
Siqi: It does sound very similar. It’s not the time disconnected that’s the scary part.
Sam: Well, yeah it is, dude. Could you go seven days? Away from your family? Like it’s crazy.
Shaan: I don’t want to cry with a bunch of strangers. I don’t want anyone to see me dancing and singing.
Siqi: What’s great about Leaders in Tech is they’re all well-known-ish founders. It’s not cheap to go.
Sam: Oh, even worse. I want to be my most vulnerable self around cool people? No — actually, it’s really helpful, because you realize that all the people you look up to, we’re all kind of the same. Very similar insecurities.
Shaan: Whenever I hear about this stuff, I think of the Tony Soprano quote where he’s like, “Whatever happened to the strong silent type? Like Gary Cooper.” I go all the way down the track. Like, can I just say “hey chief” and just know that person at that level? Just yes or no?
Siqi: And then he goes and kills about a dozen people.
Siqi’s Favorite Interview Question [01:04:00]
Sam: Siqi, you have this interview question that I love. You said your favorite interview question after 20 years of interviewing people is: “What is your greatest strength that you are most worried about not coming across in this interview setting?” Why that question?
Siqi: I enjoy breaking the fourth wall. Interviews tend to be so standardized or formalized. And my greatest anxiety when I’m interviewing someone is: I might be really good at something and you’re asking me to reverse a linked list, and it’s just not coming across.
Siqi: What I find is that question breaks down the formality and lets people get excited to talk about something they’re really, really good at. You get great stories. And you get to know the person a little bit better. The particular thing is — if you just ask “what are your strengths?” it sounds really formal. But “what is the insecurity you’re bringing in that you’re worried about not coming across” — it changes the tone quite a bit.
Sam: That’s a good one. I’m stealing it.
Runway: Art and Finance Can Coexist [01:05:30]
Sam: You founded Runway, which is a serious business — like a real B2B enterprise software thing. But you have the vibe of an artist. You have this variety of really intriguing projects. You’re like a thinker, almost a philosopher. Is this a new challenge for you — to be doing something so serious and regimented? Or can you still be goofy and an artist in the B2B world?
Siqi: Funny enough, my first company was called Serious Business — and we built fun games. I feel like if you call it serious, it isn’t.
Siqi: But yeah, I actually think it’s a pretty huge advantage, particularly for things we consider serious. People have the stereotype of business and finance being super serious and rigid. But finance at its best is really about creating value, looking forward, thinking about new ideas for how you can push the business forward and grow faster. And in order to be creative, you have to be in flow. Things that are fun keep you in flow.
Siqi: What we actually hear from customers quite often — and we love hearing it — is: “This software feels fun.” And that’s not a luxury. Fun creates flow, creates creativity, creates value. When you use something that isn’t fun — something that feels slow or confusing — you’re not creative and you’re making worse decisions. So I don’t think they’re in conflict. I think they’re quite complimentary. And that is very deeply part of our philosophy.
Runway’s Viral Billboard Stunt [01:07:30]
Sam: I saw this thing — yes, that’s what made me think of that question. What’s the story?
Siqi: Our marketing team is run by Kel Frese, who was a YC founder of a company called Taika, and we also have this woman named Julie Freitas who was at Shopify. I was in the office and I just saw them come out of a meeting kind of giggling. They got out this piece of cardboard and started writing something. I was like, “What is this?”
Siqi: “Oh, you know how you didn’t want to do a billboard? We decided to just create a billboard ourselves and hold it on a freeway.” And I was like, that is such a cracked idea, of course that’s going to go viral. And I said, “I’ll do it.” They were like, “You will?” I was like, “Yeah, I’ll do it.” So we walked outside, found a freeway entrance, and I was just holding the sign.
Sam: What does it say?
Siqi: “Hate your finance platform? Get Runway.”
Co-Founder Departure and Going Solo [01:08:30]
Sam: Are you the only founder now?
Siqi: I had a co-founder, Arya Asemanfar, and he left the company about a year and a half, almost two years ago.
Sam: That’s pretty rare — you’re very, very early in the period of being kind of the only founder running it. That’d be kind of exciting, right? You get to do whatever the hell you want. Or is it — are you going to be lonely?
Siqi: I think the reason why we parted ways is things were just slower to make decisions and neither of us were having as much fun as we wanted to. I didn’t want him to leave — he’s still on the board, I love him. He’s the most wise and high-integrity person I know. He was one of Parag’s peers at Twitter. He worked for Brett Taylor at Sierra. But things were going slower than he wanted, and he identified that the relationship was slowing us down. He decided to fire himself. I didn’t think he was right about that, but he usually is right about everything. And it totally transformed the company.
ElevenLabs: The Fastest-Growing SaaS You Haven’t Heard Of [01:10:30]
Sam: You talked about Character AI and Chai — these AI companies that are just printing money right now. Are there any other companies that have blown your mind in terms of how well they’re doing?
Siqi: ElevenLabs. I’m a fan of that company. They do basically all of the audio — translation, generation, sometimes deepfakes. When I met them they were less than 10 people. Ex-Google. Working on foundation model technology. And they are now — I don’t know what they last announced — but they’re in the hundreds of millions ARR in like a year and a half.
Siqi: What blew my mind is I’ve never seen a group of very good technologists who were also so good at commercializing the technology so rapidly.
Sam: I feel like this just came out a year and a half ago, right? They dubbed our podcast — remember that clip where it was like, “Sam, you’re speaking Hindi now”? And we were like, this is kind of insane. We’re not pursuing that more?
Shaan: There’s that graph for the fastest-growing SaaS companies and I’m fairly convinced ElevenLabs is actually faster than all of them.
Sam: When we were tinkering with them it went from “this is cool, this is cute” to “I need to use this right now.” They just crossed it. People are actually using it.
Sam: Who are they selling to?
Siqi: So Apple has audiobooks — and you’d know that a lot of AI-generated audiobooks are powered by ElevenLabs. They also do AI agents for customer service and support. So they’re the back end of other software tools as well. All kinds of product lines. They sell to book publishers, they power agents — they do a whole range.
Sam: Why are they better than what OpenAI offers? Like, the risk with all these AI companies is that the base models — OpenAI, Anthropic — can just offer those capabilities as part of the main suite.
Siqi: It’s their own foundational models. And the quality and expressiveness is just better. The form factor is better, so you can use it for all these different use cases. The Voice API of OpenAI is actually not as good or as easily usable.
Naval Named the Company, Quarter Million for the Domain [01:13:00]
Sam: How much did you have to pay for the runway.com domain?
Siqi: A quarter million dollars.
Sam: Fun fact: Naval named the company and suggested you use the .com?
Siqi: When Clubhouse only had one room, in April 2020 — it was basically the Naval app. I think you were on it a lot too. When I was thinking about what I wanted to do for my next company, I said I want to do a finance company, and I said, “I’ve got a great name for it — AI, in capital letters.” Naval said, “No. Don’t call it AI — it’s going to be dated in two years. Everything’s called AI. You should just call it Runway. And you should get the .com.” And he wrote the first check into Runway. That’s why we got the .com.
Shaan: So funny that you said it was basically the Naval app at the time. It totally was. And I was using AirChat pretty seriously for like two months, even though I knew AirChat wasn’t going to work. I was like, it doesn’t matter — this is Naval’s app, he’s going to be on it 24/7. This is like I get to hang out with Naval without emailing him like a million other people asking for coffee.
Siqi: I was on it for exactly two months. I was a power user for exactly two months and I hung out in there. That’s exactly what happened. The most fun way to use that app.
Wrap-Up [01:15:00]
Sam: Dude, you’re super fun to talk to. I like hearing all these — I feel like you’re a treasure chest and I just grab something out of the toy chest. “Tell me the story behind this.” That’s awesome.
Siqi: Thanks for having me.
Shaan: Thanks for coming on, man. Give people a shoutout for your company and where to follow you.
Siqi: Runway.com. You can follow me at @blader — Blade with an R. Or @RunwayCo. And if you’re a CFO, you should use our software.
Shaan: We’re thinking of you and your daughter. All right, see you.
Siqi: Thank you.