Why Siqi Chen is building an AI Thinking Platform for finance leaders at Runway
In this episode of Future Finance, hosts Paul Barnhurst and Glenn Hopper welcome Siqi Chen, founder and CEO of Runway, a finance platform built to help operators understand their business without needing a degree in finance. The discussion explores how traditional FP&A tools fall short, and why Runway is reimagining financial modeling by combining flexibility, design, and deeper thinking.
Siqi is the founder and CEO of Runway, a modeling platform designed for operators and finance professionals who need flexible, intuitive tools to manage business complexity. He previously served as CEO of Sandbox VR, VP of Product at Postmates, and Head of Product at Zynga. Siqi has also invested in over 100 startups, including Clubhouse, Amplitude, Pipe, and Italic. Early in his career, he worked at NASA, where he was awarded a Congressional Space Act Award while still in school for his work on the Mars Exploration Rover.
In this episode, you will discover:
Why FP&A platforms often fail to replace spreadsheets for financial modeling.
How Runway is addressing the core pains of collaboration, complexity, and data integration.
Why flexibility and speed of thought matter more than workflow automation.
How Siqi thinks about AI’s role in making finance tools more usable and accessible.
A remarkable story of how a meme coin helped fund life-saving research for his daughter’s brain tumor.
Siqi shared his remarkable journey from NASA engineer and tech executive to founder of a finance platform built for modern operators. His experiences across product leadership, startup growth, and personal resilience highlight why tools like Runway are redefining how finance professionals think, plan, and collaborate. This episode is a must-listen for anyone looking to move beyond outdated systems and embrace a more flexible, human-centered approach to financial modeling and decision-making.
Follow Siqi:
LinkedIn - https://www.linkedin.com/in/siqic/
Company - https://www.linkedin.com/company/runway/
Join hosts Glenn and Paul as they unravel the complexities of AI in finance:
Follow Glenn:
LinkedIn: https://www.linkedin.com/in/gbhopperiii
Follow Paul:
LinkedIn - https://www.linkedin.com/in/thefpandaguy
Follow QFlow.AI:
Website - https://bit.ly/4fYK9vY
Future Finance is sponsored by QFlow.ai, the strategic finance platform solving the toughest part of planning and analysis: B2B revenue. Align sales, marketing, and finance, speed up decision-making, and lock in accountability with QFlow.ai.
Stay tuned for a deeper understanding of how AI is shaping the future of finance and what it means for businesses and individuals alike.
In Today’s Episode:
[01:48] - Siqi’s Background
[04:41] - Runway Journey So Far
[07:11] - Approachable Finance Tools
[14:28] - UX & What Makes Runway Unique
[20:26] - Thinking Tool Idea
[24:28] - Excel Flexibility Trade-Offs
[33:46] - AI as a Financial Exoskeleton
[40:21] - The Meme Coin Surprise
[44:40] - Wrap-Up
Full Show Transcript:
[00:01:35] Host 1: Paul Barnhurst: Hi. Welcome to another episode of Future Finance. We're joined today by Siqi Chen. Siqi is the founder and CEO of runway, a startup that enables operators to understand their business without needing a degree in finance as an operator and founder. Siqi was the CEO of sandbox VR. He was the VP of Product and Growth at Postmates. He was also the head of product at Zynga, and he's also an investor in nearly 100 companies clubhouse, amplitude, pipe, italic and others. Previously, Siqi worked at NASA, where he was the only person to ever receive a Congressional Space Act award in school for his contributions to machine vision technology. On the Mars Exploration Rover. So we're super excited to have Siqi here with us today. Siqi , welcome to the show.
[00:02:36] Guest: Siqi Chen : Thanks for having me. Great to see you, Paul.
[00:02:39] Host 1: Paul Barnhurst: Yeah, great to see you. It's always good to catch up. And we're really excited to jump into our interview here. And Glenn's going to get us started today with the first question.
[00:02:47] Host 2: Glenn Hopper: When Paul said we were interviewing you, I was really excited. Just listening to your journey from NASA to gaming to Postmates and now runway. That's a broad spectrum, and I'm wondering how those different experiences sort of led you to where you are and how the experiences shaped your approach to building. Now a finance platform.
[00:03:06] Guest: Siqi Chen : Runway really feels like a culmination of all the things I've worked on up to that point. So as Paul mentioned, my background is technical. So I actually majored in math, and I've always wanted to start a company and build great products. And having founded a few companies operated at the companies I founded at some scale. What I've always felt really insecure about is understanding finance and business at all. And that's kind of not uncommon in Silicon Valley, right? Like we have VCs giving these kids literally millions of dollars, and these kids really have no business background. I didn't, and so I remember having a CFO, a fractional CFO, part of my team, my first company, and getting that first forecast model and income statement. And I was like, what does any of it mean? And part of me was thinking, well, I don't really need to understand what it means. I can just build a product, but that's no way to run a business. And that sense of insecurity really never left me. And it was really the turning point was when I joined a company called sandbox VR. Then I started building models. I was appointed CEO about six months before Covid hit, and having spent the time to create a bunch of scenarios for how long Covid was going to last and doing on these spreadsheets. That's when I asked a question to Andreessen, who was our seed investor. There's got to be something good that people use for this, right? Like a figma for finance or a notion for finance. And there was nothing of the kind. And so I thought, this is exactly what I want to work on.
[00:04:41] Host 1: Paul Barnhurst: I love it. And how long have you been working on the runway? Now it's been four years. Five years. How's the journey been so far?
[00:04:48] Guest: Siqi Chen : I think there's been a lot of lows, a lot of highs. Probably more low than highs. You are both intimately familiar with the space of finance and into this market. The surface area of the products that you have to build is just so wide, right? You have to build on the order of half a notion, half of Excel, and then half of work just to integrate with like 800 different services. That's just to get your foot in the door. And it is very, very different from, you know, let's say a ticket tracking system or project management system. The amount of software and the complexity of software is extraordinarily high. And so it took us, you know, three ish years to really have our first business really running officially on the runway. That company was Angel list. And since then, you know, we've been growing at a pretty accelerated rate. And once you have something built and people enjoy it, then you can grow a lot faster. But the first couple of years was tough.
[00:05:47] Host 1: Paul Barnhurst: Yeah, I remember I think we talked for the first time probably, about three years ago.
[00:05:50] Guest: Siqi Chen : Yeah.
[00:05:51] Host 1: Paul Barnhurst: That's right. I saw your first demo, and I know you're still doing quite a bit of work. You're like, I hope to be out of, I think at the time or something like February, you're hoping to kind of launch and be out of it. I think it ended up being August about six months later. So yeah. Yeah, I definitely know it took a long time because I've never seen the website years ago and going, okay, when is this product going to come to market? Yeah.
[00:06:11] Guest: Siqi Chen : Yeah. Well what's funny is we actually did have customers six months in and we had design partnerships and we were learning. But the best analogy I have for that experience is, uh, Dylan Field from Figma was one of the investors. And you describe the process of getting Figma to market. So Figma, it took six years for Figma to get to $1 million in AR. Right? And we're a good chunk, you know, of that now. And we're five years in. But like he described Figma in the early days, it was like trying to build a house. And you build this house and you want people to move in and invite people over and they're like, oh, cool house. You're like, yeah. Would you like to move in? And they say no because there's no kitchen. You're like, okay, so cool. You go build a kitchen, you buy people back over and you ask them you want to move in. And they're like, well, is there a bathroom? You're like, no, to go to the bathroom. Uh, and that's kind of like the process. Rinse and repeat until someone decides to move in. And that's basically what we were experiencing in the first couple of years.
[00:07:11] Host 2: Glenn Hopper: What I noticed digging in, though, that time that you spent, I mean, it's a robust platform now. And what I love about it is that it doesn't have that feeling. And I know Paul was going to ask a little more about this, but it doesn't have that feeling um, around. You can tell it's not speaking to finance nerds, you know what I mean? Like, it's more approachable to, uh, to everyone, it seems like. And it seems like a more plain language. And it's not just digging through your ERP and trying to, you know, build your custom reports and have the, the 1990s looking dashboards and and all that.
[00:07:43] Guest: Siqi Chen : I mean, I would argue outside of Excel, there's really no platform that appeals to finance nerds like Excel is amazing, right? Like. And generally, if you're a financier, you don't want to leave Excel and a spreadsheet because it's a great piece of software. And so we saw the opportunity in terms of design. Right. Like we have Figma notion of these amazing consumer grade products for knowledge workers. And the question we ask is like what is something of equivalent quality for a finance professional. And so I would say it's less, you know, we certainly do want to appeal to everyone in a company. And we see amazing adoption across different departments because of the accessibility and understandability and familiarity of it. But really, we also want to just build this offer for the finance professional. And we think a lot of that is that's underrated and being deeply undervalued in the space is good design.
[00:08:40] Host 1: Paul Barnhurst: Yeah. And you know, kind of talking about that. I know you've had some kind of cheeky advertising you've done around, you know, finance and how many people hate their platform. In fact, I think the current tagline on your website is the finance platform your team doesn't hate. Yeah, right. So a two part question to that one. I'm curious how that resonated in the market? What kind of things do you hear from finance people? Yeah. And then two, in your view what was broken in traditional planning? Yeah. No. You spoke to the wider business with what you built, but why was the runway needed?
[00:09:13] Guest: Siqi Chen : I mean, it usually gets a wry grin because, you know, it's not something we made up, right? If you look at NPS scores. As fo FP&A a category, it has like the lowest the net promoter scores out of basically any software category. And so it is not a well loved category generally. And so, you know, this is not a surprise to anyone who's been in finance. But we spent a lot of time talking to customers and finance professionals who use these platforms. And we're trying to answer the question, well, why do people not like it? And our answer to that is it's actually fairly subtle and non-obvious. So if you think about why someone adopts an FDA platform, um, it's actually pretty weird, right? Because finance professionals love Excel, and Excel is an amazing product. So why would anyone actually, like, want to move off of it? Right. And there is the reason is because as a company scales you experience, you experience increasing pain on the spreadsheet. And we categorize the pain from what we heard in like basically three categories. So the first category is actualizing your model and keeping you up to date right. So you have more data sources. You have higher data volume. And you know who has a 10 million cell limit right.
[00:10:31] Guest: Siqi Chen : So you know that's like over a number of years that's not actually that many rows you can actually model. And so at some point you need to have a scalable system that automatically connects to all of your sources of sources, data and shapes it so you don't have to spend a bunch of time transforming your data and keeping your model up to date. So that's the first problem. The second problem is that there are more people in your company, right? And so the way every finance professional gets budgets in and makes sure it gets done is we create this empty spreadsheet template right for every department. And then you email it to department heads and then you slack them. Please fill this out and they ignore you because who wants to do your paperwork right. It's now the bottom of the to do list and you keep on stacking them. At some point they're like, okay, I'm sick of eating out. I'm just interested in numbers. I don't even care about these numbers. Who cares? And then, you know, back to you. You paste pastes that paste it back. So that's another painful workflow that you want to solve. And the third thing that happens is the business just gets larger, it gets more complex. And you're at some point you lose trust in the integrity of your model.
[00:11:37] Guest: Siqi Chen : Right. Like if you have like 10 million cells, what if one of them is off and you just don't. How do you find it? How do you even know? So model complexity, the collaboration with those departments and those workflows and data integrations and keeping actualize your model. Those are like the three categories or problems. And what's interesting. And so that's why you buy one of these things right. So you go on a website, you search around the platform and see how it integrates with everything that I use. Great. Oh I like to give people tasks and I can do my budgeting workflows. Great. Oh, it has great dimensional modeling. And so that's going to simplify my model complexity. And so basically everything that you can buy in the space will actually solve those problems because those are the problems that you want solved. And that's not why people don't like the platform because everyone kind of does the same thing. What is fascinating and most surprising about the space, and the reason why I think people don't like it, is what happens six months a year after you implement, which is That. As it turns out, finance is not just about reporting numbers and getting the budget said. Like you have something about a business change.
[00:12:45] Guest: Siqi Chen : You have to think about a new scenario. You have to think about a new go to market notion, some new business unit, and you need to change your forecast and your model. And that's where it gets annoying because you realize that, oh wow, it was so easy to do that on my spreadsheet. Excel is super flexible and the thing I bought doesn't actually do that thing. It does. The first three things I mentioned the pains that I had in the spreadsheet, but it doesn't do this thing that I have to do. And that thing is thinking. Thinking through something new, making something, making a new model, making some kind of new plan. And that's the issue. And that's why it's not well liked and said before, not being well liked is what happens then? What happens then is you then have to still stay on your spreadsheet. So every single customer we've talked to, without exception, on an FM platform, still maintains a forecast on a spreadsheet because of this specific reason nobody was able to get out of the spreadsheet. Which is why. Because that's the entire reason why you bought this thing in the first place, and that's probably why you hit it. So that's like our insight into the problem.
[00:13:48] Host 2: Glenn Hopper: And that's kind of that's kind of what I was alluding to earlier. So I know it starts with, I mean, the UX is very approachable and it feels we've all seen the giant spreadsheets that go around and you try to have your sort of dashboard page on the front and your executive summary that's showing all that. And it just almost gets so bogged down into like, you can when you look at a lot of these spreadsheets, you can just feel the build happening under it, whereas this data just feels accessible the way you guys do it. But I know you guys have a lot more going on under the hood, and I want to unpack that a little bit. But I know you said at some point that, uh, that startups don't win by being better. They win by being different. And I can tell that's a fundamental approach here. And I think what you said at the runway, you realize the bottleneck isn't speed, it's understanding. And I think so is the understanding in the UX. But also maybe if you could delve a little bit into some of the some of what makes it unique, what's going on under the hood as well? Yeah.
[00:14:44] Guest: Siqi Chen : It's actually really hard to say no words, which is why we had these videos on our website. And so we just like to show you the product. Like when you look at it, it just makes sense. Um, but I'll do my best with the words. So there's um, the first thing you said, it's interesting, which is that, you know, you start with what might be different, right? And so our analogy here is that if you're familiar with design as a category, uh, design software. Um, so there was this company called, there is a company called Invision. And what the software designers do is you upload files to the cloud and then you can, as a team, leave comments on it and get approvals. And it goes through this workflow so that software and that company was worth about $1 billion. And today's worth is basically zero. And the reason why is because Figma exists. Right. And Figma is a tool where designers can now do design in your browser and everyone can collaborate. Leave comments. Go through the workflow directly. Right? And so replace the original software on the desktop. The designer uses the design. And the interesting thing here, and Figma is obviously like one of the most valuable private companies of all time. It's like 20 plus billion dollars now. But what's interesting here is like if you rewind 12, 13 years ago when we first got started. If you ask a designer, it's like what they wanted. Zero would have said, hey, you know what I want? I want to do the work that I'm doing on my sketch.
[00:16:06] Guest: Siqi Chen : That's all software. And I want to use my browser. And by the way, can you have everyone in the company look over my shoulder as I'm doing design? That'll be really great. Like, nobody was asking for that, right? And our head of design actually was like, I will never use that. Right. But now everyone uses it because they built different products and they solved something quite useful for designers for the team. And so our analogy for this market is that if you look at The States as a whole. They are the equivalent of envisions and that these are workflow products like they're about making sure that people are getting their budgets in on time and like everyone's aligning and you can leave comments, assign tasks and and the numbers are up to date. But what they're not is equivalent to Figma or even a spreadsheet where you are going in there to think models, shape your data, do your work. And that's how we refer to this issue differently. Instead of being a workflow tool or position as a tool for thinking, thinking together with your team. And so that's the high level differentiator. The way, the specific way we do it is we believe that it's an idea of what we call abstractions. Um, and the idea is simple. It's how we make it so that the thing you see on screen and the way you manipulate things on screen is much closer to how you think inside your head. And when we think about, like, tool technology tools, like the best technology.
[00:17:30] Guest: Siqi Chen : Tools are tools for thinking. They help us think faster and think more clearly than we can do alone in our head. And so the analogy I have for, you know, the way I think about technology tools like a whiteboard, right, is a form of information technology in that you can offload your thoughts onto this whiteboard that you can't contain completely in your head. You can't hold all the context at once. And when you offload it to this whiteboard, the people that you work with can also see it. And now you can work together, right. So that's like a similar character to a spreadsheet because a spreadsheet you're offloading in ten. But it can do these calculations that you can't do in your head really really quickly. Right. And so we think about running in those terms. And so for example we have this concept called plans. And the idea is when you as an operator or your business counterparts think about their business, it's not likely that what's in their head is literally a wall of numbers. Right? What is in their head is like, here's my marketing plan, here's my quarterly roadmap, here's my product roadmap. And each of these things affects the business in some way. Increases conversion rate, increases contract value. And so we really thought about how we like to connect it to so that you could be manipulating plans, not just cells. And these and many other ideas are all under the principle of how do we make the things that you see now map to the concepts in your head?
[00:18:56] Host 2: Glenn Hopper: This, uh, this feels like rocket science under the hood, but it's a good thing you had the work with NASA, but I really I mean, it's thinking through these kind of issues that are problems that we're facing every day. And coming up with this novel solution, I mean, it makes it a very unique product in the marketplace. It's super, super cool.
[00:19:15] Guest: Siqi Chen : Yeah. The secret sauce, I want to say, like, you know what technology at its best is like it is complicated under the hood, but it is through design immediately understandable, accessible. All right. So if you think about our favorite products like Google is originally right. What's the most complicated piece of technology? Right. But you interact with it by just typing some words. Right. And like having, you know, when you scroll through your contacts, you know, when the iPhone first came out and it just works. That's enormously complicated, but it feels natural, like that's what technology and good design can do. And that's very much what we're inspired by.
[00:19:55] Host 1: Paul Barnhurst: I love that kind of idea that you mentioned earlier, particularly the whole year, under the hood, it can be really complex, but if you can make that design to where people just understand it, they want to use it, right? Those are the best products. It's not that the people see the complexity, it's that they find a way to make that complexity simple. And the other thing, can you elaborate a little bit? You've talked about this before. You and I have I know, but just the whole idea of thinking, you know, building a tool that's designed for thinking versus workflow. When did that first come to your mind that that's, you know, what runway needed to be? That it was really about a building, a thinking tool. How did that kind of come into existence, that idea?
[00:20:38] Guest: Siqi Chen : Yeah. So it was um, as we went to market and we had more people use it, and it was clear that we had a certain opinion, a certain angle in which we were approaching this problem and the product just felt different. At the same time, when you enter, when you're building a product in order to win, you can't. A winning product strategy does not look like we'll take everything that an existing category product does. And in every future we will make it better, right? Because every company has limited resources and the startup you have like even fewer resources. So you have to figure out how you're going to win. And the way you're going to win is you have to be much better or different in the most important areas, and you're okay with being just as good in the other areas. And so it was more about like, where do we focus and where do we differentiate and what is true, and how does that intersect with both what the market needs and opportunity there and what we're uniquely good at in creating? And so the things that customers say about the runway, they'll say it's fun. It's super flexible. The number one reason people switch to the runway is, you know, there it is for the reasons of flexibility is what our customers tell us. Right. Like something changes in the business and they can't get their existing thing to make it to to model that. And so they want to be able to change your scenarios, model new business lines on the fly.
[00:22:05] Guest: Siqi Chen : And so when we think about what if we're going to double down on that, what does that look like? The thing that we want to double down on is the flexibility of the modeling, the experience of modeling the thinking at the speed of thought. And where we're going to be less focused on is making sure that you know, you have like this effectively a project management system where you can assign tasks and have a Gantt chart. And there's like this really heavy flow around, here's how we get a budget approved. Right? That's a different product. And the reason why we chose to bet on this aspect is because of exactly that core insight that we had is like everything that I described, from data integrations to budgeting workflows to a dimension that only exists in every product, basically in this space. And yet everyone dislikes it. And the reason why everyone dislikes it is because they can't get up their spreadsheet. So what we decided is we just need to focus on how we can get people off that spreadsheet for real and in order to get out the spreadsheet. The reason why you say it is because it's not flexible enough. The thing that you bought, and because the spreadsheet is a tool for thinking. So we have to build a tool for thinking that also solves those other problems. That's kind of how we arrive there.
[00:23:18] Host 1: Paul Barnhurst: Ever feel like you go to market teams and finance speak different languages? This misalignment is a breeding ground for failure in pairing the predictive power of forecasts and delaying decisions that drive efficient growth. It's not for lack of trying, but getting all the data in one place doesn't mean you've gotten everyone on the same page. Meet QFlow.ai, the strategic finance platform purpose-built to solve the toughest part of planning and analysis of B2B revenue. Q flow quickly integrates key data from your go-to-market stack and accounting platform, then handles all the data prep and normalization. Under the hood, it automatically assembles your go-to-market stats, makes segmented scenario planning a breeze, and closes the planning loop. Create air-tight alignment, improve decision latency, and ensure accountability across the team.
[00:24:26] Host 1: Paul Barnhurst: Yeah. And you know, people often ask me to go, I want the most flexible tool out there. Go. If flexibility is truly the number one thing you want. And these other pain points don't matter, then don't leave Excel. That's right. Because there's no database on the back end. It's unstructured data. It's a thinking tool. You could do whatever you wanted. Exactly, but just realize you're going to deal with a lot of other pain points. Yeah.
[00:24:46] Host 2: Glenn Hopper: I think about that. So you know arguably maybe Microsoft Dynamics is the most flexible ERP. But then the problem is the implementation takes, you know, and that's so custom and all that. So it's like well it'll ultimately be flexible. So if you can give them that flexibility off the shelf and not have them have to think about that level that I mean that's amazing.
[00:25:07] Guest: Siqi Chen : That's a yeah, that's a design problem. Right. Like you can make very complex things. It's a technology design problem. Right. You can make very complex things understandable, more accessible through better technology and design. This is a really great product. I haven't even heard of it. It's called Alta. Um, but what it is, is like a sequel. It's a sequel client. So it's like it'll talk to any database, but it's a completely visual SQL client. So you can like to group and aggregate and filter and join completely visually. And it's really, really cool. And it was someone's MIT thesis. But we believe there is an equivalent. We have to like to build some aspect of that to get data in. But there's an equivalent for the modeling and planning use case too. And this requires very, very time consuming, painful design thinking to get there.
[00:25:59] Host 1: Paul Barnhurst: Yeah, I can imagine a lot of design thinking. We talk a lot about that as we talk about the future. I know you've talked a lot about AI, how you think about that a lot with your product. I know recently, a few months back in San Francisco, you spoke at the CFO summit with some others about what does the finance tech stack of the future look like? Where are we heading with AI? What is it going to look like? How is it reshaping things? And so more broadly, I know we've talked a lot of FP and A in runway, but I'd like to get your thoughts just broadly on where we're heading and what that tech stack of the future looks like for the Office of Finance.
[00:26:34] Guest: Siqi Chen : We're, we're we're going to go and where we are is basically there's two factors here. One is the capabilities of the models themselves. And the other is uh, product builders, uh, progress and packaging those capabilities into applications. And so we've been pushing on AI since really like mid 2022. And it was, uh, not a great use of our time until very recently. And the reason is like, you know, we were playing with GPT 3.5 and GPT four, it was difficult to even get it to reliably output JSON, right? And meanwhile, we're trying to get it to model something very complex. Understand the complexity of business. The context window wasn't high enough, reliability wasn't good enough. But we were like, oh, we're just going to, you know, prompt engineer better, better prompts. Right. Um, and basically around September of last year when I came out and really, oh, three earlier this year, all the work that we tried to do basically didn't need a ton because now you could just model capabilities. The improvement is actually absurd. So we went from like we could barely get it to work to we're going to feed it like documentation around a role model language, and we're going to give you a picture of a spreadsheet and it can like basically recreate it in our own, like, uh, in our own language, like perfectly like it was mind blowing. And so the model capabilities have only just gotten to a point where it is like, legitimately quite useful for modeling, planning and analysis purposes.
[00:28:10] Guest: Siqi Chen : very recently in the last like 6 to 9 months. And up to this point, the really useful applications have been not really around modeling and finance like the numbers, but around RFP, around accounting where you're coding, um, and, and you don't have to send emails to those people to pay you. Like, that's fairly useful. But what's going to change now and over the next year is how these models are now capable of doing, like really productive work in RFP. And so then when you look at the three sorts of categories of work, uh, in FP and itself, one is like making sure that you have good reporting and good data. And the foundation of good reporting is good data. So that's like one piece. The other is the modeling piece. And so yeah, if you're really ready to excel or if you use something at runway, you could think about new models and you can do so flexibly. But there is still a learning curve, um, and expertise. But I will make it so much more accessible. Right? You can write today and run away. You can just like it. And I'll show Glenda tomorrow. Obviously, it's like you can just mouse over any part of the product and hold on option and what we call the artificial runaway intelligence. Ari will just go and explain and start talking to you about and explain everything that you're hovering over. Right. How a formula works, why this number is like a million sets of 500,000 a month per month, etc.
[00:29:36] Guest: Siqi Chen : and then from there you can just have it do changes, right? You can say, okay, now I want to have a new business line, and it's going to be a product led growth type of model. And here's what I want, and I'll just create that model for you. And that's pretty game changing in terms of accessibility and also even more importantly, time to value. Right. One of the things that's plagued this space is just then how long it takes to implement. But you can imagine, um, these models are now capable of looking at your data, shaping it and pivoting it into your reports. And then from there, even taking a look at your existing model and forecasting it right into the implementation, and you're taking like a six months, three months implementation into six days, six hours. And that's going to happen this year. So that's going to be pretty dramatically game changing. And so the way we think about this is there's two schools of thought for how people interact and use AI. So the first school of thought is sort of this ChatGPT and agent tech type of workflows where you're treating this AI as this creature, this external thing. And so, you know, when you're talking to people, you're talking to a person, right? And with an agent, it's like this employee that's going to do stuff on its own or with some minimal amount of direction. And that's when the use case of AI.
[00:30:54] Guest: Siqi Chen : That's all very interesting. And I think Pigman talked about like the model agent, analyst agent, the scenario planning agent. And I think that's like a really valuable use case for AI. But we are a tool for thinking. And so the way we think about AI is a little bit different. In addition to these agent expressions. And the way we think about it is you think about something like cursor right. Or Copilot what the tools that engineers use to write code, right. That isn't just a chat interface, that is a cursor. An IDE is a tool for thinking. And when you're writing code, it helps you write faster. Autocomplete writing infers intent from the work that you are already doing. And so when you think about that expression, to me that feels much different than like a thing that you talk with. We think of it in terms of like you're already doing work and we're inferring intent or accelerating that work. So let me give you a really like stupid small example. One of the first real AI native features I've ever seen developed in the wild is not. It's something inside chatting with you. Obviously ChatGPT is like a revolutionary product, but there's this one small feature in Chattahoochee that I thought was fascinating that very few people talked about, which is that when it first comes out and you have a conversation, there is a sidebar, right? With all your chat history there, you don't name the chat.
[00:32:19] Guest: Siqi Chen : The chat names itself, right? Like it knew the conversation and it just gave it a name that has never happened. That's like a new user interface pattern. Never in the history of software like mind blowing, right? But that's like truly aided design. So like let me give you an example, which is like use Excel or use runway today, frankly. And one of the things that you do is you write a formula. But what else do you do? You format it. Do you want a dollar sign? Do you want a percentage sign? You want a number. You know, how many decimals do you want? You shouldn't have to do that because we know the name of the row and the formula, like a human can instantly tell what it's supposed to be. It's supposed to be a percentage. It's supposed to be Australian dollars, right? And this is an example of AI where like there is no chat, there is no even interface. But like we just know, those are the type of AI expressions that I really enjoy. That's what sort of really defines a native product to me. It's like you're using AI the way you're using an if statement. And I think most implementations today, just like we're going to put a chatbot on the side. And that's not a very interesting piece of AI. I think it's totally underutilized. It should be embedded into every piece of text and every pixel.
[00:33:39] Host 2: Glenn Hopper: And that's such a fundamental shift. And a lot of I mean, people just aren't catching on to it. And that's what I think about, you know, you talked about the different ways you use AI. And I always think of it as for now, it's like an exoskeleton, you know, it just makes takes what you do and makes you able to do it quicker. If you don't have to go through and format cells and remember which way to click the arrow and Excel to make the decimals increase or decrease and all that which I always forget.
[00:34:04] Host 1: Paul Barnhurst: But he gets that one wrong, right?
[00:34:07] Host 2: Glenn Hopper: And I think Google Sheets does the opposite of correctly, but on which side you click. But um, you know, at the same time though, Andy Jassy just came out yesterday and said, look, we're going to, you know, jobs are going away because of AI and Microsoft. What is it like 8000 layoffs and then 6 or 6008 thousand significant number of jobs going away with that and with the interesting thing and and, you know, part of the existential crisis we're all having now is what am I going to do? But you think about every move through technology and what, 100 and whatever years ago, uh, we were 80% agriculture, you know, working in agriculture in the US and how different that is now. And so there's jobs we can't even imagine that are coming up. But when you talk about the level of insight you get, where, you know, I'm just picturing in a spreadsheet, you click on any cell and it's giving you information beyond just whatever the formula is there. And I keep going back. And Paul's going to roll his eyes because I use this quote all the time, but it's, uh, it's, uh, Clifford Stoll and I just, I think about this as the pyramid of, as automation.
[00:35:09] Host 2: Glenn Hopper: And I move up, we're getting squeezed the value that we add at the very top. But the quote is data is not information. Information is not knowledge. Knowledge is not understanding, and understanding is not wisdom. And I look at that as the pyramid. And as we automate each level up, you know, if you can think about the variance analysis reports that we do every month when we, you know, in fact, when we first get it. And that used to take a while, but now you can have I do the variance analysis and it will understand and can give the same kind of insights in two seconds, something that would take me two hours to go through. And I just it's it is the reality of technology right now. And I don't know ultimately what it means for entry level finance pros or even mid-level and senior year where you're going to add that value if it's being automated more and more up. I'm not duma around this.
[00:36:00] Host 1: Paul Barnhurst: I call Glen doomsday. Glen. Just so you know.
[00:36:04] Host 2: Glenn Hopper: But yeah, it's super interesting. And I think, I mean, in the technology is coming so quickly because I'm sure even on your team, the rate at which you can develop now, because when you can use cursor and you know, all the tools out there for development, I mean, it's wild right now.
[00:36:19] Guest: Siqi Chen : It is so wild. And this is such a fun question to talk about too. So last month I know, I think you're all familiar with Operators Guild, right, which is this, um, community of non engineering, um, employees in tech, and a lot of finance people were part of the community. So I went to their conference and I gave a keynote on AI and building. And what I did is I brought one of our salespeople from the runway, and he did part of the keynote not to sell the runway or talk about the runway at all. He demoed his workflow because he committed 6000 lines of code last week. As a sales guy, he literally shipped dark mode and he's like constantly shipping new things because cursor exists and just tell cursor what to do and it'll just build things and it can one shot all these small features, small bug fixes now just by asking it. And so the takeaway there is that there's going to be massive change in the short term though. It's democratizing building. It's democratizing knowledge where people like, you know, I call it the rise of the idea guy, right? Previously, like, nobody likes the idea guy because idea Guy always needed a technical co-founder who was the actual valuable person. Now the idea is like, maybe like the idea I can just build whatever they want on their own.
[00:37:44] Guest: Siqi Chen : So that's kind of interesting. But your question about the long term, me, I'm like, on the one hand you could say, you know, I'm so sympathetic to argument and it's very real and say, yeah, most people were farmers or, you know, in the case of the Luddites, they're weavers. And everyone found different jobs. I think there is a large extent that's likely to be true. I also think that comparing AI to basically any other technology revolution has ever happened is actually probably under hyping AI. And the distinction is that every other thing that you can name that's happened before, whether it's the loom or Movable type or internet or mobile, was all the product of intelligence. This is intelligence itself and is capable of like new inventions. Pretty soon that we're not maybe a year or two away from that level of reasoning capability. In that world, all bets are off. And so what you're saying about finance is true for every role that humans have, um, professionally, uh, including product managers, uh, including CEOs. Um, and so that's sort of the pessimistic view. The optimistic view is that proof of human is like, if you think about the things that are going to be like scarce and valuable in that world, it's probably going to be realistic because you can make more land. Um, and then the other category is basically everything that is true for human.
[00:39:05] Guest: Siqi Chen : So luxury goods are probably going to be more expensive as they happen, right? Like, dogs were expensive 200 years ago, but now plastic bags are cheap. But you still have a very expensive bag because, you know, it's made by some, like artisanal craftsmen in the south of France. There's nobody who can play better chess or better go than a computer today. But you don't watch computers play chess, you watch Magnus Carlsen play chess. And I think there's something equivalent to, um, professional work. To where? Like, unless we truly, as a society say, like, well, we're perfectly okay with AI autonomously making all decisions about business and products and everything, then all bets are kind of off. If we're not okay with that, and we say we want humans to be able to press the button and at least approve it, then the question is, how does a human have the judgment to know what to approve, and the judgment can only be acquired by experience and by actually building the model in a spreadsheet from scratch. And maybe I can help you understand how to do it faster. But this is why I think a tool for thinking is actually exponentially more valuable in an AGI world, because your ability to create and understand and hold that clarity in your head becomes more valuable.
[00:40:16] Host 1: Paul Barnhurst: I love that answer. I know we're very short on time. We're coming over. We got just a few more minutes. If you can hang with us and we'll wrap this up. I recently learned that you created a meme coin.
[00:40:26] Guest: Siqi Chen : I did not create a meme coin.
[00:40:28] Host 1: Paul Barnhurst: I heard you did from somebody recently.
[00:40:30] Guest: Siqi Chen : So, uh, this is the wildest story. But my daughter was diagnosed with a brain tumor last year, September. And, uh, we've been supporting this lab because it's very rare. Brain tumors are 1 in 1,000,000. And, um, we are done with this lab. It's the only lab that researches brain tumors, basically in North America and Christmas. I started this fund for this lab and I tweeted about it, and we ended up raising a 300 K for the lab through GoFundMe. But what happened is someone replied that it's a very viral tweetstorm about the fund and asked me if I had a crypto address because they wanted to donate crypto because they're crypto person. So I said, yeah, here's my Ethereum address, which is the only like address I had at the time. And they're like, no, no, no, I don't want three addresses. I want a salon address, which is a different blockchain, which is the, you know, the blockchain. And the Trump coin was on, uh, later it was like a month later. But anyway, so I'm like, okay, I don't know. Uh, fine. I'll create one. I woke up the next morning and I posted this address for my salon wall onto the thread.
[00:41:37] Guest: Siqi Chen : And it was an empty wallet. Right. Because, you know, I have nothing in it. And like, I woke up and I opened the wallet and it said 400,000 USD. I'm like, what? And I have 500 million tokens of coin called Mira, which is the name of my daughter with a brain tumor. And someone created this meme coin, and there's a billion tokens and sent me half of the entire supply of this token and the market cap of the coin, because it was immediately traded, was like already $1 million or 800,000. So I'm like, what? What is happening? So I took a screenshot of that wallet and I said, what is this magic internet money? And about half an hour later, I looked into my wallet. And now it's at $4 million because that screenshot went viral. So I took $304 million. I'm like, what is happening? An hour later it was 8 million. And she shot at that. And at one point I had 18 million USD of this meme coin in my wallet. And I'm like, I have no idea what to do. Um, because, like, you know, if I just sell it all, I think, like, I'm gonna just crash the value of the token.
[00:42:51] Guest: Siqi Chen : Faced with this dilemma, and like, a lot of people are just, like, investing. And I really thought about it and said, here's what I'm going to do. I don't want to surprise anyone. I don't have time to run a crypto project. 24 hours from now, I'm going to start selling this token. I'm going to sell about $100,000 every hour on the hour continuously. If you are, you know, if you want to exit, now's your time, right? So that's what we did. And we ended up liquidating about $1 million and immediately donated in crypto to the lab.it's their largest donation ever. So the University of Colorado Hankinson lab, and,it was a majority of your budget from the last five years. Um, especially with, you know, the Doge sort of shutting down research funding, like they would have had to shut down if not for that donation. I didn't create it. I don't know who created it. Some random kid in Dubai who probably made $5 million from this. Like the mean value of the coin. But from my perspective. Yeah, it was like an amazing thing that made a huge difference to kids with this particular tumor.
[00:43:55] Host 1: Paul Barnhurst: It's an amazing story and great you're able to donate there and hope your daughter, you know, all goes well there. I know that's always a challenging thing when you deal with health challenges, especially with your children.
[00:44:06] Guest: Siqi Chen : So yeah, the great news is and I haven't really talked about this, is that, um, thanks to the work of this lab, we introduced her to a new treatment for this drug. And she's now one of five people on the planet who has it. And her tumor doubled in size in January. And in the three months since we introduced this drug, it's shrunk by 80%. It's like an unbelievable response. And so that's a really exciting. You're talking about that a bit more.
[00:44:31] Host 2: Glenn Hopper: But yeah yeah that's a good story to end on. And it seems a little lighthearted to go to our silly questions. So yeah.
[00:44:40] Host 1: Paul Barnhurst: Thank you so much for sharing and for sharing the meme. I hadn't heard that. I just heard someone talk about it. And like, you have to ask him about this. And I'm like, all right, oh.
[00:44:50] Guest: Siqi Chen : Symmetries in crypto.
[00:44:52] Host 1: Paul Barnhurst: Yeah, you write a book. Well, we'll buy that children's book Adventures in Crypto. You can have me write it for you.
[00:45:01] Host 2: Glenn Hopper: That's true. Yeah. And illustrate it you could and market it.
[00:45:06] Host 1: Paul Barnhurst: And thank you again so much for joining us.
[00:45:09] Guest: Siqi Chen : So great to be back.
[00:45:11] Host 1: Paul Barnhurst: Yeah. Thanks.
[00:45:12] Host 2: Glenn Hopper: Thanks, sticky.
[00:45:13] Host 1: Paul Barnhurst: Thanks for listening to the Future Finance Show. And thanks to our sponsor, QFlow.ai. If you enjoyed this episode, please leave a rating and review on your podcast platform of choice, and may your robot overlords be with you.