Doug Camplejohn
(00:00)
Welcome everybody. This is Doug Camplejohn, your host of Revenue Renegades. This week I am pleased to welcome Siqi Chen to the show, the CEO of Runway. I’ll let him explain it, but it’s an Andreessen Horowitz-funded company that’s re-imagining financial planning. Siqi, I have to say you have one of the funniest things on your LinkedIn profile. I was scrolling for this interview and it was your background where you talked about one of your companies, saying it was a shittier version of YouTube. What was this? Vio, like YouTube, but not as good, as a dev crash site for two hours, then made into a PM. It actually made me laugh out loud. Welcome to the show.
Siqi Chen
(00:57)
Thank you. True story, though.
Doug Camplejohn
(01:00)
That’s awesome. As a founder, I always love founding stories. Tell me a little about how Runway came about and what was the inspiration.
Siqi Chen
(01:11)
My job before this was CEO of a company called Sandbox VR. We’re a retail company with VR stores in malls where you can play holodeck experiences. The company was doing very well, then COVID hit and shut down the world. We were an outdoor experience with friends, so suddenly everything changed. Every institutional investor, including Andreessen, was creating a new portfolio, protecting their own companies. We had to make a case for why we were worth saving, and we were probably at the bottom of the list.
Doug Camplejohn
(01:46)
Mm-hmm.
Siqi Chen
(02:01)
As part of that work, I had to create a bunch of scenarios for how long COVID would last and what we’d do in those scenarios. This was March 2020, when people like Elon Musk said it would be over by June, and Andreessen wanted to see a two-year scenario. We thought that was crazy, but it became reality. We created different spreadsheet variations on Google Sheets. Eventually, the outcome was that we bridged the company so we could hibernate through COVID. I had to lay myself off, the entire leadership team, 95% of the company—from 400 employees to 10. I wasn’t a founder; I was the chief product officer, then CEO. The founder hibernated the company. Thirty minutes after that, I spoke to our CFO and asked: “Instead of Google Sheets, is there something better we could use for planning? Figma, Notion—these are great for other things.” He was an experienced CFO from Netflix and said, “No, this is as good as it gets.” I had a conversation with Andreessen and suggested someone build a Figma for finance—something accessible, powerful, a great tool for thinking. They wanted someone who understood the problem and, for some weird reason, wanted to work on finance software. I raised my hand and said, “I want to do that.” That’s how Runway got started.
Doug Camplejohn
(03:56)
That’s amazing. Your background is software engineering, not finance, correct?
Siqi Chen
(04:01)
Yes, I majored in math and started my career in software engineering—worked for NASA JPL during college, then at YouTube (the aforementioned not-as-good site), then PowerSet, which Microsoft acquired and became part of the Bing team. Then I started my first company. I haven’t done full-time engineering since then. My background is consumer products, growth. I’ve led product at Zynga, Postmates, and ran growth there too.
Doug Camplejohn
(04:42)
A lot of folks—I’m an engineer by training, though I haven’t coded in years—find finance kind of a scary, black box for product or technical people. We all deal with imposter syndrome, which we can talk about at length, but you’ve described finance as inherently intimidating.
Siqi Chen
(05:02)
Yes, and the tools themselves contribute to that culture.
Doug Camplejohn
(05:11)
You argue it shouldn’t be so scary. Often it’s the tools that make people feel inadequate. Can you expand on that?
Siqi Chen
(05:19)
It’s absolutely the tools. From first principles: What is finance, what’s a finance model? The spreadsheet seems complicated and filled with financial terms, but it’s actually important because it’s a simulation of your business. The model tries to represent every component: marketing, sales, product, retention, every person. When you see it as a business simulation, you realize its value is that you can simulate the future and understand current operations. We’re used to building these for reporting to boards and investors, but their real value is for internal understanding. For that to work, the model must map to concepts in your head—not just a wall of numbers. Our mental models aren’t endless grids of numbers; they’re product roadmaps, marketing plans, strategies. This disconnect is where tools fail.
Doug Camplejohn
(06:56)
So you get the idea, Andreessen says you’re in the small Venn diagram of crazy people who want to solve a difficult but meaningful problem. How did your team approach building something radically better than a spreadsheet?
Siqi Chen
(07:02)
It was a combination of first principles thinking, experimentation, and learning from customers—a very creative process, definitely not purely scientific.
Siqi Chen
(07:28)
We talked to over a hundred finance people, learned the state of the art, and discovered surprising things. The pain of spreadsheets builds until everyone seeks alternatives. Updating and collaborating is painful; buying expensive products still can’t eliminate spreadsheets. We also built prototypes—tried different things—and thought deeply about what models actually are and what you’re truly trying to accomplish. We realized finance is much bigger than finance—it’s everything that happens in a company. Everything you do affects the bottom line, but software rarely reflects this. You invest millions in R&D, engineering, product, design, for clear reasons: improving contract value, close rates, retention—all with dramatic financial impact. Software doesn’t capture that connection.
Doug Camplejohn
(09:17)
When starting and experimenting, what was the first thing that resonated with customers?
Siqi Chen
(09:30)
I initially built something for myself as an operator—this hit a pain point, but we learned valuable lessons from it. The first version was a fancy pivot table. You import expenses and transactions from QuickBooks and then pivot and view them instantly as you want. Typically, your QuickBooks report is a group of accounts, but you think in terms of engineering costs, sales costs, departments. Our tool let you instantly view all that however you wanted and mix pivot views—for example, engineering costs by vendor and sales revenue by customer. You could even correct categorization mistakes, just by dragging them—like Mint, but more powerful and intuitive.
Siqi Chen
(11:01)
If your Google hosting costs were lumped into G&A by mistake, you could drag them to engineering, and it’d recategorize them going forward. That first version gave a clear idea of where money went.
Doug Camplejohn
(11:28)
Nice. So when did you feel product-market fit, and what was the timeframe? Was this before or after ChatGPT?
Siqi Chen
(11:46)
We started the same month GPT-3 was released—so about three years before ChatGPT.
Doug Camplejohn
(11:58)
Got it. That first version scratched your own itch. I love building things for myself; it doesn’t guarantee everyone has the same pain, but you found it resonated with others as well.
Siqi Chen
(12:27)
That’s right.
Doug Camplejohn
(12:28)
At what point did AI enter your product roadmap?
Siqi Chen
(12:37)
Fairly early. I was never the primary engineer, but I did one pull request—for an AI feature we shipped in May 2020. We were one of the first to get access to GPT-3. We built a feature that picked the right emoji for account names. It’s gone now, but it was an experiment.
Doug Camplejohn
(13:25)
Deep, meaningful use of AI. Love it.
Siqi Chen
(13:28)
We wanted it to feel like a super product—like Notion with emojis. That was our very first AI—2020. Then we kept iterating on applying AI to financial modeling and forecasting. Language models are great here because people are good at language, but bad at mental math. So, a spreadsheet model you could just talk to seemed ideal. Even with GPT-4, understanding spreadsheet context and intent was difficult—but reasoning models that came out recently are promising. We now use AI extensively internally, as well, automating qualification for signups, for instance.
Siqi Chen
(15:33)
When our website went viral three years ago, getting multiple signups a minute, we couldn’t qualify by hand. So we built automation: email signups were sent to Clearbit for enrichment, then a GPT evaluator for qualification, then piped into Slack. If qualified, we’d draft clever “flight” themed emails and put them into Gmail drafts. Today you can buy that off the shelf.
Doug Camplejohn
(16:06)
Ha.
Siqi Chen
(16:17)
We built a Notion bot called Jam Session. You mention it just like any user: “@Jam Session with Steve Jobs,” and it critiques your document as Steve Jobs, or Elon Musk, or even Daffy Duck. We also built automatic email classification. I even created the #1 education GPT on the GPT store—Universal Primer—for recursive learning.
Doug Camplejohn
(17:28)
It’s fascinating—you built multiple things that could each be a Y Combinator company.
Siqi Chen
(17:38)
That’s just how things work now. AI has democratized building in amazing ways.
Doug Camplejohn
(17:51)
So, AI started with emojis and internal tools; how does it impact Runway products now? You describe AI as an Iron Man suit for finance.
Siqi Chen
(18:44)
Agents are a more obvious way we bring AI in. These agents should feel like coworkers: you reference them in comments, you can chat or email them, and they show up in Slack. But what’s more interesting is rethinking familiar UI paradigms in an AI-native way. One inspiration is ChatGPT’s sidebar, where chats name themselves—subtle, understated, but fundamentally magical and not possible before. Traditionally, lists require manual naming. Now it’s automatic.
Siqi Chen
(20:27)
There are many examples—some in production, some experimental. Another inspiration was Joel Spolsky’s comments on Google versus Microsoft engineers: Google engineers use Bayesian filters where Microsoft uses if statements. AI-native products should use LLMs the way software uses if statements—probabilistic and intelligent.
Siqi Chen
(21:00)
We built “Explain Mode”: hold the option key, mouse over any number or button, and you get a tooltip that explains what it is. It’s an AI-native tooltip you can demo visually. These are magical moments—new paradigms for UI in AI-native products.
Doug Camplejohn
(24:40)
That is so cool.
Siqi Chen
(24:50)
We patented the core ideas. There’s just so much possibility: AI-native text input, natural language, inline chat. Moving beyond the idea of the LLM as external creature, we make the tool for thought, to help people think clearly in human language.
Doug Camplejohn
(26:14)
I’ve recently been exploring voice tools, especially after breaking my elbow. Monologue is great on Mac, and I’ve used natural language filters for CRM—just say, “Show me all California companies between 50 and 500 employees using Salesforce,” and never touch the keyboard. The metaphor of spreadsheets and tables is resilient. AI-native products don’t necessarily replace them, but add delightful magic moments throughout.
Siqi Chen
(26:58)
Absolutely.
Siqi Chen
(27:04)
I just had a conversation with our AI pod—I’m the lead PM for it. The analogy is like Minority Report: futuristic interfaces, but traditional tools like keyboard and mouse still persist for ergonomic reasons. Making paradigms like buttons, dropdowns, text inputs, and tooltips truly AI-native is valuable.
Doug Camplejohn
(28:45)
I remember an early TED talk with the first pinch-and-zoom tablet demo—pre-iPhone, post-Minority Report. Now, I still have a keyboard on my laptop, but the right metaphor depends on the context. It’s not one-size-fits-all. I could geek out on UI forever, but let’s shift gears and talk about personal integration. You’ve written that work-life balance is overrated. How do you integrate family and work?
Siqi Chen
(29:49)
I’m probably not qualified—I work a lot! But that’s relevant and important. The idea behind “work-life balance” suggests dualism, but work is an integral part of life. I don’t work to live, and most humans don’t. Work provides fulfillment, purpose, and value; losing work harms happiness. So the distinction between “work” and “life” feels false. I provide for my family, model value creation, leadership, and communication for them. I also work on family directly—spending time, conducting medical research for my kids. Is that work or family? It’s intertwined.
Doug Camplejohn
(31:46)
For context: How old are your kids?
Siqi Chen
(31:49)
Five and eleven years old.
Doug Camplejohn
(31:52)
Are you in-person, remote, or hybrid at work?
Siqi Chen
(31:55)
We’re hybrid: a third in San Francisco, a third in New York, a third elsewhere. Founded during COVID, remote was default—but in-person is better, so we bring everyone together quarterly. Offices: San Francisco twice a week, New York almost five times a week (younger team).
Doug Camplejohn
(32:34)
We’re remote—my co-founders and I worked together in our last company, finish each other’s sentences, so quarterly face-to-face meetings work. High bandwidth in person—my execs in Portland, I fly regularly. Definitely higher bandwidth face-to-face.
Siqi Chen
(32:45)
Absolutely.
Doug Camplejohn
(33:03)
The phrase “work-life balance” implies work is drudgery, home is fun. For people like us who love our work, it’s not about shutting it off—it’s more porous. I’ve closed rounds from Mexico, done team meetings from Italy. No strict Monday-Friday.
Siqi Chen
(33:50)
Exactly.
Doug Camplejohn
(34:01)
Switching topics: You’ve started and invested in dozens, maybe hundreds, of companies. What lessons or takeaways have you gained—what did you want to replicate or avoid at Runway?
Siqi Chen
(34:40)
From investing and seeing other operators, the biggest takeaway is the combination of reliability, discipline, and energy. It’s valuable. Jessica Livingston and Paul Graham at Y Combinator learned that startups aren’t just a test of intelligence, but of stamina. Dillon Field (Figma) once said the most surprising thing was how few people made it to the end of the journey. In early Figma days, half the team left. Longevity and stamina matter most.
Doug Camplejohn
(36:57)
It’s wild. I met Dylan at a Greylock event; even three or four years ago, it wasn’t clear Figma would make it. The Figma “ramp” took six years before revenue. In today’s market, would investors wait that long? The patience for AI dev platforms is different.
Siqi Chen
(37:40)
Right. Six years to a million dollars ARR.
Doug Camplejohn
(37:55)
Now, if you don’t do a million in three months, it’s considered failure. Fast, easy capital rules.
Siqi Chen
(38:05)
But Dylan kept the end vision in view. At LinkedIn, we used InVision—great, but now it’s gone. Figma’s survived because it did something unique.
Doug Camplejohn
(38:29)
You’re known for radical transparency—what uncomfortable truths should founders hear about finance?
Siqi Chen
(38:57)
Honestly, finance isn’t that important early on.
Doug Camplejohn
(39:02)
Really?
Siqi Chen
(39:04)
Yes. V1 of Runway showed founders where their money went—almost always, to people, and that doesn’t change month to month. Finance becomes critical only when things take off, hiring increases, marketing spend rises. Early on, you just need frugality and to track spending.
Doug Camplejohn
(39:23)
Understood.
Siqi Chen
(39:33)
When scale comes, trade-offs and decisions become valuable—and finance gets important. But generally, finance makes people uncomfortable, so even early discussions are tough.
Doug Camplejohn
(40:06)
For example?
Siqi Chen
(40:08)
One key lesson: founders consistently undervalue intuition. I’m a third-time founder, and it’s still a challenge. The stereotype is founders operating on vibes, frustrating those seeking logic and data. Most advice in business isn’t popular because it’s true, but because it’s viral—makes people feel good, so it spreads. Every founder should know their own superpowers and leverage them.
Siqi Chen
(43:08)
Evidence of a superpower is if things that take others a long time, happen for you almost instantly and unconsciously. These insights are unique. When you hire smart people who disagree, as a humble CEO you weigh their researched opinions above your “random feelings”—this can be a failure mode. If you undervalue your superpower, you avoid explaining it and can’t lead the team effectively.
Doug Camplejohn
(44:09)
Applying that to myself: I have taste in product, intuition, and I listen to others. As my LinkedIn colleague said, “I may be wrong, but I’m not confused.” Bring better ideas, and I’ll pivot. As CEO, the challenge is wielding authority without overwhelming debate. Sometimes you just have to say, “Here’s where we’re going.” The Netflix “Quickster” story is an example—Reed Hastings had been right before, so no one objected, but it was a disaster. It’s hard to know when to use the CEO card.
Siqi Chen
(46:01)
That’s fine. Explaining your thinking helps, but it won’t always work. The CEO card exists for a reason. Quickster was a mistake, but it didn’t block Netflix’s ultimate trajectory. High conviction led to error, but that’s allowed. The popular narrative is the arrogant CEO, but everyone makes mistakes, and that’s part of growth.
Doug Camplejohn
(47:33)
I met Steve Jobs a few times—never understood his anger when younger, but I do now. It wasn’t about arrogance; it’s about seeing something so clearly in your head, being frustrated others didn’t get it fast enough.
Siqi Chen
(48:08)
Exactly.
Doug Camplejohn
(48:18)
Sorry for turning this into a personal therapy session.
Siqi Chen
(48:20)
No, this is great—where fun happens!
Doug Camplejohn
(48:25)
We’re almost out of time, so let’s finish with a little personal info. You shared about your family and work-life integration. What do you like to do for fun?
Siqi Chen
(48:41)
I run, and picked up boxing a couple years ago—spar three or four times a week. It’s stress-relieving, confidence-building, and surprisingly intellectual.
Doug Camplejohn
(49:10)
I’ve tried it few times—it’s exhausting! Watching matches is incredible.
Siqi Chen
(49:24)
Yes, three rounds will floor you—and they go twelve. Even three minutes is tough.
Doug Camplejohn
(49:37)
Outside that, what’s something people would be surprised to know about you?
Siqi Chen
(49:44)
I’m an INTP on Myers-Briggs—the most introverted type. People don’t expect that.
Doug Camplejohn
(49:52)
Got it. As product people, what’s one product you use that’s made an impact?
Siqi Chen
(50:05)
I like 8 Sleep—a big help. And Insight Timer for meditation, a very underrated app.
Doug Camplejohn
(50:07)
Same.
Doug Camplejohn
(50:18)
Why that over, say, Calm or others?
Siqi Chen
(50:23)
A particular teacher on it just has the calmest ASMR voice. It’s like YouTube for meditation.
Doug Camplejohn
(50:33)
I’ll check it out. Jeff Weiner (former LinkedIn CEO) was board member at Headspace; he always remarked how weird it was that the CEO did the narration at actual meetings.
Final question: How can listeners stay in touch with you and help you with Runway?
Siqi Chen
(51:02)
We’re Runway—runway.com. I’m on LinkedIn under my name, and on Twitter (“not calling it X”) as Blader (blade with an R).
Doug Camplejohn
(51:14)
Awesome. Siqi, this has been fantastic. Thanks for your time—I really enjoyed the chat.
Siqi Chen
(51:20)
I enjoyed this too. Thanks for having me, Doug.
