模拟：新的 Scaling Law——Simile AI 的 Joon Sung Park

2024 年，当我们首次讨论模拟式 AI 之夏时，就知道那会是一个短暂的夏天；但最近，模拟式 AI 卷土重来，势头凶猛：4 月有 SimGym，如今又有 Simile AI 估值 20 亿美元的 Series B。该公司获得了 GreenOaks 和 Index Ventures 的支持，知名支持者包括李飞飞和 Andrej Karpathy；它为 CVS 等《财富》100 强客户运行数千万次模拟，并宣称相较于人类焦点小组实现了 85%–99% 的准确度。

是时候了解一下，为什么这个模拟的第二个夏天能够奏效了！

---

从打造 Smallville——这篇具有里程碑意义的 2023 年 Generative Agents 论文展示了 AI 角色能够记忆、规划、社交，并发展出涌现行为——到如今构建人类行为基础模型，Joon Sung Park 正在尝试回答一个大得多的问题：如果我们能在现实世界中做决策之前先模拟这个世界，会怎样？ 在本期节目中，Simile 联合创始人兼 CEO 将拆解从生成式 Agent 到数字孪生的发展路径，解释为什么今天的前沿模型仍无法捕捉人类实际的行为方式，以及最终模拟地球上全部 80 亿人需要具备什么条件。

我们深入探讨了 Simile 的 人类行为建模 方法：长篇访谈、观察与交易数据、随机对照试验、人口层面与个体层面的模型，以及围绕 人们为何做出决策的因果机制 进行后训练。Joon 解释了他的研究如何创造出数字孪生，其复现人类行为和态度的准确度达到 人类复现自身回答准确度的 85%；为什么以理性为优化目标的模型，可能会成为非理性人类的糟糕模拟器；以及为什么要理解“社会物理学”，可能需要改变模型权重，而不仅仅是对前沿 LLM 进行提示。

我们还探讨了 模拟背后更大的雄心：在产品和政策部署前对其进行测试，找出通往目标结果的反直觉路径，对整个社会中的涌现行为进行建模，以及潜在地应对 气候变化、民主制度不稳定和 UBI 等问题。Joon 还反思了 模拟的 Scaling Law、数据中心规模的模拟世界 的经济性、它与 Thomas Schelling 和心理史学的联系、为什么模拟与绘画惊人地相似，以及我们是否已经生活在某个模拟之中。

---

• Smallville 与 Generative Agents 如何促成 Simile 的诞生

• 为什么 Joon 的团队提出：“如果我们能够直接 重建我们所生活的世界 呢？”

• 为什么实用的个人 Agent 需要具备 对用户的深层模型

• 记忆架构、Markdown 文件，以及 提示的局限性

• “社会物理学”与 行为基础模型

• 为什么网络数据更多捕捉的是人们说什么，而不是 他们实际做什么

• 访谈、交易、观察数据，以及 随机对照试验

• 为什么预测未来不如理解 如何塑造未来 重要

• Simile 如何创建 具有代表性的模拟人口

• 模拟与预测的区别，以及它与《基地》心理史学 的联系

• 如何评估模拟，而不是简单地堆叠 LLM 幻觉

• 为 1,000 名真实用户创建数字孪生，并达到 85% 的行为准确度

• 为什么前沿模型可能难以复现 真实的人类行为

• 为什么优秀的模拟需要复现 人类的偏见和错误

• 基于 随机对照试验 对模型进行后训练

• 人口层面的模拟与 个体层面的模拟

• 人类模拟的 Scaling Law

• 最终 模拟地球全部 80 亿人 的长期雄心

• 模拟能否帮助解决 气候变化，或侦测濒临崩溃的民主制度

• Thomas Schelling 与 基于 Agent 的建模 历史

• 为什么未来模拟可能需要一整个 数据中心

• 多 Agent 模拟，以及 模拟人类相互交互 时会发生什么

• 以 合成人口 替代昂贵的人类调研小组

• 为什么市场研究只是 模拟的起点

• 为什么 Joon 认为模拟与 绘画 惊人地相似

• 使用模拟研究 UBI 等问题

• 我们是否已经 生活在模拟之中

• 为什么 AGI 与模拟 可能是高级文明的双生技术

---

---

00:00:00 开场与 Joon 从艺术走向 AI 的历程

00:01:46 Smallville、Generative Agents 与模拟的起源

00:05:03 “我们就创建一个世界吧”与个人 Agent 的未来

00:09:53 社会物理学与行为基础模型

00:14:08 预测与模拟：如何塑造未来？

00:16:59 Simile 如何对真实人物和人口进行建模

00:25:35 评估模拟、数字孪生与 85% 准确度

00:30:23 对模型进行后训练以复现人类行为

00:40:04 Scaling Law 与模拟 80 亿人

00:43:10 从 Schelling 到社会规模的 Agent 模拟

00:46:13 模拟世界的成本与经济性

00:52:05 现实世界用例、合成人口与市场

00:57:27 模拟、绘画与 UBI 的未来

01:04:23 我们是否已经生活在模拟之中？

01:06:08 创建 Simile 与招聘

---

Vibhu [00:00:00]： 今天我们请来了 Joon 做客播客。很期待开始这期节目。这是一家非常令人兴奋的公司。我想从一个问题开始：请你讲讲你的人生故事。你是如何走到今天这一步的？

Joon [00:00:13]： 当然。我非常高兴来到这里。我的人生故事是这样的：我出生在韩国，在那里生活了大约 11 年，之后我们家搬到了波士顿。我们是在我 11 岁时搬过去的。我的父母都是医生，当时正在进行博士后研究。我父亲是外科医生，所以他在波士顿儿童医院度过他的学术休假期。因此，我在那里成长，和科技圈并不算接近。我当时非常偏向音乐和艺术，也喜欢绘画。

Vibhu [00:00:49]： 绘画。

Joon [00:00:49]： 没错。我稍晚一些，在高中时开始接触绘画，但那就是我过去主要做的事。离开韩国后，我大部分时间都在美国东海岸长大。我在新罕布什尔州住了很多年，之后去宾夕法尼亚州上大学。我是在大学期间开始更多地接触科技圈的。最初我接受的是艺术家训练，我以为那会是我的职业生涯，所以它不是一种爱好，而是“我要靠这个谋生”。后来，我逐渐对这样一个想法产生了强烈兴趣：最伟大的艺术家往往会创造属于自己的媒介，而我们今天拥有的最佳媒介是计算。因此，我决定更深入地探索这个领域。随后，一件事接着一件事——我们可以继续深入谈这个过程——我逐渐对研究产生了兴趣，于是就走到了今天。

Swyx [00:01:46]： 你在研究这一部分包含了很多内容。你发表了 2023 年最好的论文之一，也就是生成式 Agent 论文，通常被称为 Smallville 论文。

Swyx [00:01:58]： 你可以回到你刚刚提到的任何内容，不过大多数人可能都是通过这篇论文认识你的。你知道有多少人读过它吗？arXiv 会给你一些统计数据，对吧？

Joon [00:02:10]： 是的，这是个好问题。至于有多少人读过它，我不太确定。

Joon [00:02:14]： 我知道我们会持续跟踪引用次数，而它们增长得很快。

Swyx [00:02:23]： 对，Google Scholar 显示它有 7,200 次引用。

Vibhu [00:02:25]： 我感觉它的影响力比这个数字更大，而且它是一篇相当具有奠基性意义的论文。它被引用了非常多次。

Swyx [00:02:34]： 当人们问“你最近读过最好的论文是什么？”时，它经常就是答案。

Vibhu [00:02:39]： 我觉得其中的记忆组件被低估了。那是一个非常好的早期记忆系统，也是最重要的论文之一。

Joon [00:02:47]： 是的，也许我可以稍微讲讲这篇论文是如何产生的。我是在 2020 年进入研究领域的，当时我开始在斯坦福攻读博士，而那一年 GPT-3 即将向外界开放。那时我们已经有 GPT-2，你能感觉到，市场上正开始出现一类全新的模型，团队对此产生了浓厚兴趣。大家普遍会问：“这种模型真的能用于什么吗？”“很奇怪，这些模型并不是为了执行任何特定任务而训练的。”但我们决定押注于它。因此，斯坦福的一大批学者聚集在一起，由我的一位联合创始人 Percy Liang 领导。

Swyx [00:03:35]： 提出了 foundation models 这个术语的人。

Joon [00:03:36]： 对，他提出了 foundation models 这个术语。我们写了一篇论文，这个术语就是从那篇名为《基础模型的机遇与风险》的论文中来的。在那个过程中，我真正开始深入思考的是：这里有一个在我们生态系统中根本全新的模型。它之所以新颖，是因为它并非为了完成任何特定任务而训练，但它的前提是可以完成任何事情、一切事情。借用生物学类比，它就像一个干细胞。

我对这个想法很感兴趣：如果我们真的思考，这项技术能够实现的杀手级应用会是什么？我的许多同事用它做简单的分类、简单的生成。这些模型能够做到这些很有意思，但从交互的角度看，并没有那么有意思。我们已经知道如何做这类事情几十年了。最终我们得出的结论是：这些模型是在非常广泛的网络数据上训练的，对吧？这些是人类行为数据，包括社交媒体、Wikipedia，以及所有这类数据。因此，如果你从正确的角度去触碰它，就能看到相当真实的人类行为涌现出来，而这是我们以前从未见过的。

Joon [00:04:45]： 这让我们非常感兴趣。于是，我和这一组同事——Michael Bernstein、Percy Liang，以及我自己，后来他们也成为了我在 Simile 的联合创始人——坐下来玩了一个我们称为“时间机器游戏”的游戏。

Joon [00:05:03]： 想象一下，我们坐上时间机器，快进 10 年后再回头看：到那时，最重要、最有趣、最鼓舞人心的单一应用会是什么？当我们想到“如果我们能够重建我们所生活的世界呢？”时，很难有比这更宏大的目标了。就像是：我们就创建一个世界吧。

Joon [00:05:24]： 这就是我们的起点。最初，我们有一篇作为 Generative Agents 论文前身的论文，叫作 Social Simulacra。

Swyx [00:05:32]： 在你继续之前，在这个时间机器练习中，还有其他最宏大的候选方向吗？第二名或第三名是什么？

Joon [00:05:44]： 有一个非常接近的第二选择，是我们当时正在考虑的自动化工具，尤其是那种真正为个人量身定制、能够替你完成事务的 Agent 愿景。

Swyx [00:05:59]： 这件事也正在发生。

Joon [00:06:00]： 它也正在发生。但对我们而言，这很有意思。我们决定选择模拟这个方向，一方面是因为我是一个科幻迷，而创造模拟的想法让我个人非常着迷。我很喜欢这个概念。看到一个这样的游戏小镇，以及其中这些 Agent 生活在里面，真的很酷。

但与此同时，我的判断是：如果你想基于这项技术打造一个真正出色的个人助手，首先需要的是一个出色的用户模型。我对一个模型说：“嘿，能帮我买份晚餐吗？”它却点了一份夏威夷披萨，而我不喜欢披萨上有菠萝。那它就完全失败了。它要避免犯这个错误，唯一的方法就是对我是谁有深刻理解。我这里举了一个非常简单、甚至有点愚蠢的例子，但你可以想象，这种对人的核心理解是多么重要。这也是为什么我们的家人和最亲密的朋友会对我们形成良好的心理模型，那是我们社会连接的基础。

所以，我们的另一个判断是：围绕模拟的这项技术——创建对人的准确表征——理应先于那些将要自动化我们所生活世界的、更复杂的 Agent 出现。这就是我们的判断。但个人助手是一个非常接近的第二选择，我现在仍然对它非常着迷。我认为这个领域有很多有意思的工作正在进行。

不过，我在这里有一个可能有些争议的观点：我认为我们还没有看到真正有用的个人助手，其能力还没有达到这条研究路线最初设想的雄心。已经有一些早期应用很有意思；而且即使你现在和 ChatGPT 或 Claude 对话，它们也知道很多关于我们的信息。所以我确实认为，它们正在生成的很多内容都更加个性化了。但我认为这个领域的雄心非常大，而且我们还没有完全拥有所有正确的要素。

Swyx [00:08:01]： 那么，对于 OpenClaw 和这些个人 Agent，你希望它们具备哪些目前还没有的能力？

Joon [00:08:09]： 我确实认为它正在慢慢接近，但总体而言，我希望它们对用户拥有更深刻得多的理解。现在，你看这些模型。OpenClaw 所利用的是一个 Ma

_When we first dicsussed theSummer of Simulative AI in 2024 we knew it would be a brief summer, but it has recently come back with a vengeance with SimGym in April and now Simile AI’s $2B Series B, backed by GreenOaks and Index Ventures with prominent backers like Fei-Fei Li and Andrej Karpathy, running tens of millions of simulations for Fortune 100 clients like CVS and 85–99% accuracy vs human focus groups. _

Time to catch up on why this Second Summer of simulation is working!

---

From creating Smallville , the landmark 2023 paper on Generative Agents that showed AI characters could remember, plan, socialize, and develop emergent behaviors , to now building foundation models of human behavior, Joon Sung Park is trying to answer a much bigger question: what if we could simulate the world before making decisions in it? In this episode, the Simile co-founder and CEO joins us to unpack the path from generative agents to digital twins, why today’s frontier models still fail to capture how humans actually behave, and what it would take to eventually simulate all 8 billion people on Earth.

We go deep on Simile’s approach to modeling human behavior : long-form interviews, observational and transaction data, randomized controlled trials, population-level and individual-level models, and post-training on the causal mechanisms behind why people make decisions. Joon explains how his research created digital twins that reproduced human behavior and attitudes 85% as accurately as people reproduced their own responses , why models optimized to be rational can be bad simulations of irrational humans, and why understanding “social physics” may require changing model weights rather than simply prompting frontier LLMs.

We also explore the much larger ambition behind simulation : testing products and policies before deploying them, finding counterintuitive paths toward desired outcomes, modeling emergent behavior across entire societies, and potentially tackling problems like climate change , democratic instability, and UBI. Joon reflects on scaling laws for simulation , the economics of data-center-scale simulated worlds , the connection to Thomas Schelling and psychohistory, why simulation is surprisingly similar to painting, and whether we might already be living in one.

---

• How Smallville and Generative Agents led to Simile

• Why Joon’s team asked: “What if we can just recreate the world that we live in?”

• Why useful personal agents require deep models of their users

• Memory architectures, Markdown files, and the limits of prompting

• “Social physics” and behavioral foundation models

• Why web data captures what people say more than what they actually do

• Interviews, transactions, observational data, and randomized controlled trials

• Why predicting the future matters less than understanding how to shape it

• How Simile creates representative simulated populations

• Simulation versus prediction and the connection to Foundation’s psychohistory

• How to evaluate simulations instead of simply stacking LLM hallucinations

• Creating digital twins of 1,000 real people and reaching 85% behavioral accuracy

• Why frontier models can struggle to reproduce real human behavior

• Why good simulations need to reproduce human biases and mistakes

• Post-training models on randomized controlled trials

• Population-level versus individual-level simulation

• Scaling laws for human simulation

• The long-term ambition to simulate all 8 billion people on Earth

• Whether simulations could help solve climate change or detect collapsing democracy

• Thomas Schelling and the history of agent-based modeling

• Why future simulations could require an entire data center

• Multi-agent simulations and what happens when simulated people interact

• Replacing expensive human panels with synthetic populations

• Why market research is only the starting point for simulation

• Why Joon sees simulation as surprisingly similar to painting

• Using simulation to study questions like UBI

• Whether we are already living in a simulation

• Why AGI and simulation may be the twin technologies of advanced civilizations

---

---

00:00:00 Introduction and Joon’s Path from Art to AI

00:01:46 Smallville, Generative Agents, and the Origins of Simulation

00:05:03 “Let’s Just Create a World” and the Future of Personal Agents

00:09:53 Social Physics and Behavioral Foundation Models

00:14:08 Prediction vs. Simulation: How Do You Shape the Future?

00:16:59 How Simile Models Real People and Populations

00:25:35 Evaluating Simulations, Digital Twins, and 85% Accuracy

00:30:23 Post-Training Models to Reproduce Human Behavior

00:40:04 Scaling Laws and Simulating 8 Billion People

00:43:10 From Schelling to Society-Scale Agent Simulations

00:46:13 The Cost and Economics of Simulating the World

00:52:05 Real-World Use Cases, Synthetic Populations, and the Market

00:57:27 The Future of Simulation, Painting, and UBI

01:04:23 Are We Already Living in a Simulation?

01:06:08 Building Simile and Hiring

---

Vibhu [00:00:00]: Today, we have Joon in the podcast. Excited to kick this one off. Very exciting company. I wanna kick off and ask you the question, talk us through the story of your life. How have you gotten here?

Joon [00:00:13]: Yeah, for sure. I’m really excited to be here. A story of my life. So I was born in Korea, and I lived there for a good 11 years or so of my life, and then my family moved to Boston. So we moved when I was 11, and my parents were doctors, so they were going through their postdoctoral studies. My dad was a surgeon, so he was doing his sabbatical years at the Boston Children’s Hospital. So I grew up there, not too close to tech. I was very much a music and artsy, painting kind of guy.

Vibhu [00:00:49]: Painting.

Joon [00:00:49]: Exactly. I got into painting a little bit later, in high school, but that’s what I used to do. And then I grew up mostly in the East Coast after Korea. So I lived a good number of years in New Hampshire, and then I went to college in Pennsylvania. And I got into more of this tech scene, in college. So I was originally trained to be an artist. I thought that would be my professional career. So it wasn’t a hobby. It was like, “Hey, let’s make a living out of this.” And then gradually, I got really interested in this idea of, hey, the greatest artist often creates their own medium, and the best medium that we had available today was in computation. So I decided to go deeper into that, and one thing led to another, and we can go deeper into this, but I decided that research was something that I gradually got interested in, and here I am.

Swyx [00:01:46]: So there’s a lot that you packed into the research components. You had one of the best papers of 2023, which was the generative agents paper, commonly known as the Smallville paper.

Swyx [00:01:58]: Feel free to call back to anything else that you mentioned, but most people would have heard of you from this. Do you have any statistics on how many people have, like, read it? arXiv gives you something, right? Some stats.

Joon [00:02:10]: Yeah, it’s a good question. How many people have read it, I’m not sure.

Joon [00:02:14]: I know we do keep track of citations, and they are going up quite fast.

Swyx [00:02:23]: Yeah, Google Scholar has 7,200 citations.

Vibhu [00:02:25]: I feel like it made a bigger hit than that, and it was a pretty instrumental paper. It got cited so many times.

Swyx [00:02:34]: It is frequently the answer when people ask, “What is the best paper you’ve read recently?” It’s this one.

Vibhu [00:02:39]: I thought the memory component was pretty underrated. It was a very good early memory system, and one of the biggest papers.

Joon [00:02:47]: Yeah, so maybe I can talk a little bit about how this particular paper came together. So when I got into research, it was back in 2020 when I started my PhD program at Stanford, and that was the year, when we were about to get GPT-3 to be available. So we already had GPT-2, and you could sense that there was this new class of models that was just becoming available in the market, and the team got very intrigued. And the general consensus was, “Well, is this model going to be useful for anything?” “It’s really strange that these models are not trained to do any particular task.” But we decided to take a bet. So a large group of scholars at Stanford, and it was led by one of my co-founders, Percy Liang, and we came together

Swyx [00:03:35]: Who coined foundation models.

Joon [00:03:36]: Who coined the term foundation models. We wrote this paper, where that term came from called Opportunities and Risks of Foundation Models. And during that process, really the thing that I started to think deeply about was, here is a model that is fundamentally new in our ecosystem. The reason why this was new was it wasn’t, again, trained to do anything in particular, but its premise was it could do anything and everything. It was like a stem cell, if you were to take a biology analogy. And I got really interested in this idea that, well, if we were to really think about what are the killer applications that this particular technology would enable, what would that be? Many of my colleagues were using this for simple classification, simple generations. Interesting that these models can do that, but from an interaction perspective, not that interesting. We’ve known how to do that for many decades. And what we came down to was these models are trained on this very broad data from the web, right? So these are human behavioral data. It’s social media, Wikipedia, all these data. So if you poke at the right angle, then you could see human behavior that would just pop out that’s quite realistic, and we’ve never seen that before.

Joon [00:04:45]: So that got us really interested. The exercise that we decided to do, with this particular group of colleagues, Michael Bernstein, Percy Liang, and myself, who ended up becoming my co-founder at Simile, we sat down and we played this game that we call the time machine game.

Joon [00:05:03]: Imagine we were to get on a time machine and fast-forward 10 years and look back. What would have been the single application that will have mattered that would be the most interesting and inspiring? And when we thought, “Well, what if we can just recreate the world that we live in?” it’s really hard to get more ambitious than that. Like, let’s just create a world.

Joon [00:05:24]: And that’s where we started. And initially, we had this paper that was a precursor to the generative agents paper called Social Simulacra.

Swyx [00:05:32]: Before you go further, were there other candidates for the most ambitious thing in the time machine exercise? What was number two or number three?

Joon [00:05:44]: There is a close second that we were considering, which ended up becoming more of these automation tools, especially the vision around really personalized agents that would do things for you.

Swyx [00:05:59]: That’s also happening.

Joon [00:06:00]: It’s also happening. But it was interesting for us, right, in that the reason why, we decided to go with the idea of simulation, one, I was a huge science fiction nerd, and this idea of creating simulation, I was personally really just fascinated. I loved the idea. It’s really cool to see, like, a game town like this and just see these agents live in it. But at the same time, my bet was if you were to create a really amazing personal assistant out of this technology, what you need first is an amazing model of your users. So I told a model, “Hey, can you go buy late dinner for me?” And it orders Hawaiian pizza, and I do not like pineapples on my pizza. Then it totally failed. The way for it to not make that mistake is only by having a deep understanding of who I am. And I gave a very simple and dumb example here, but you can imagine how this core understanding of people is instrumental. This is how, if we have our family and closest friends, they have a good mental model of who we are. That’s the basis of our social connection. So our bet also was this technology around simulation, creating accurate representation of people ought to precede the more complex agents that would automate the world that we live in. So that was the bet. But that was a very close second, and I’m still very much fascinated by it. I think there’s a lot of interesting work that’s going around. My hot take here, though, is I don’t think we’ve seen a true personal assistant that’s useful, in ways that meet the ambition of that particular line of work. I think there are early applications that are interesting, and if you talk to even ChatGPT nowadays or Claude, they know a lot about us. So a lot of the generation it’s doing, I do think it’s much more tailored, but I think the ambition is quite large in that field, and I don’t think we quite have all the right ingredients just yet.

Swyx [00:08:01]: So OpenClaw and these personal agents, what do you want to see from them that they don’t currently have?

Joon [00:08:09]: I do think it’s slowly getting there, but I do generally want them to have much deeper understanding of the person. Right now, you look at the models. OpenClaw, what it’s leveraging is a Markdown file, and I think it’s quite clever, right? So if you look at the generative agents paper, this was the same intuition that we had, where initially when we were creating the memory architecture for the generative agents, and, like, this is, like, back in 2022, so we didn’t really quite have the idea of even agentive architecture or the term agent. But the intuition that we shared with some of the work that’s coming out today was we initially thought, “Well, do we want to make the memory into, let’s say, knowledge graph? Do we want to train a bespoke model?” All of these things. And what we decided to do was, “No. Just forget about all this.” These language models are quite good at modeling text and understanding and reasoning about text. So just put everything in a Markdown file or a text file. You’re done. I thought that was quite interesting that we could do that, and there’s a lot of strength in doing that. But also, there are limitations. It’s the way you retrieve and make sense of data that’s extremely large, it takes a lot of work. So I think that technology is getting better. I also do, however, think, there are certain things you just cannot shape just by prompting the model. So to some degree, you do need to touch the parameters of the model itself. So there is this work that I do think does need to happen, and it is happening. The question is, how far can we take it? How do we source data, and how do you also create an ecosystem where people are continuously feeding data to this model so it’s learning about you?

Vibhu [00:09:50]: What’s the intuition between why you need to do it in the model?

Joon [00:09:53]: My intuition behind the actual when do you train or even post-train a model versus just prompt a model is if the model has to learn the underlying physics of the world that it’s operating in. So it has to learn new social physics. The places where it doesn’t have to train are the places where it already has the physics. We trust the physics. It already has the base statistics, but it’s just trying to react to an environment. Then I think you can just prompt your way into getting the actions out of it. I don’t think the models that are out in the open have yet learned the complete mapping of social physics of humanity. This is one of the core theses of Simile, right? And one of the core reasons why that is the case is if you look at the data that the model was trained on, these models were trained on the web data, like, whatever was available on the web. And these are really interesting data sets, but they are fundamentally the self-exposed attitudinal data with some behavior data that’s sprinkled around here and there. And it has yet to learn the really deep behavioral nature of people, not just what people say they do online, but what they do in real life. And this is one of what I would consider to be the dark knowledge of humanity that we haven’t quite captured. And it’s these data that would also need to get factored into the model creation.

Vibhu [00:11:21]: You call it behavior foundation model.

Vibhu [00:11:23]: There’s a good one-liner here, but outside of that, what type of data do you need? What are you changing on the model level? How do you go about modeling, doing a behavior foundation model?

Joon [00:11:35]: We think about data in three buckets. So one bucket is interview data. It’s quite interesting. Rich qualitative data is interesting. It’s not behavioral, but we would literally ask people, “Hey, tell me the story of your life.”

Vibhu [00:11:53]: It’s just what we’re doing here exactly.

Joon [00:11:54]: The question that you all asked at the beginning of this interview literally is the question we also ask. And we ask our participants to go a little bit deeper, than how far I went. Maybe I can give more of my life story in lieu of this. But the reason why that data is interesting is by learning about this very long-tail information about people, you get a lot of texture around this model, like, this person as a model. So even understanding their childhood memory or even their trauma, their first love, these things, quite informative in ways that’s really hard to predict. So that’s one. Then there are two tranches of what I would consider to be the behavioral data. One kind of behavioral data is observational. So these might be like transaction data, or these might be data that you can get by scraping the web, right? So you can imagine why these data sets would be interesting, right, because they give you the base statistics of people’s behavior.

Joon [00:12:55]: But then there is the last category of data, that I personally think is perhaps the most important, which is the data that describes the causal mechanism, the whys of people. Some of this is covered by the interview data, the qualitative, because people talk about why they made certain decisions. But really, where you get to see the most behavioral aspect of this is in randomized controlled trials, like RCTs. Imagine you have the same setup, but you have a few different variables that you are trying to tweak. Can you get realistic human behavior out of it in ways where, imagine you had this particular option. Imagine you’re even trying to choose whether you’re going to drink coffee or not. The day you drink coffee versus the day you didn’t drink coffee, does your behavior change? That’s a data set that describes a causal mechanism. This is quite important in modeling people. The reason why this is important is oftentimes when people come to us, or not just to us, but the reason why people are interested in simulation isn’t because they want to predict the future. If you’re trying to win against the stock market, predicting the future is interesting.

Joon [00:14:08]: But most people, most decision-makers, what they want to know is, how can we shape the future? It doesn’t really help you to hear that your sales are going to tank in two quarters. They’re just gonna say, “Wow, that sucks.” What they want to know is, well, what do we need to do now to avoid that future? That’s the causal mechanism. And this is also very hard data to come by, right, because the world is our ground truth, but it happens once. So in a very controlled setup where everything is equal except for one variable, this kind of data set rarely happens. So this is a reason why this data set is both hard to come by and quite important if you’re trying to model human behavior.

Swyx [00:14:50]: So behavior, I think, is the hardest data set to acquire. What is out there? What is even possible? You’re not going to know a lot of details about my life. I don’t even have data for myself on my own health or habits, and I just don’t log everything. So how can you have that data?

Joon [00:15:14]: So we run a lot of randomized controlled trials.

Swyx [00:15:17]: But you put people in the lab, they watch them sleep, or what?

Joon [00:15:20]: We do care a lot about the consent process. People know that we invite them to be a member of this community to both share data and have themselves represented in different forms. But we bring a lot of people to the lab, or virtual lab, where we design experiments that would pose them real behavioral decisions. And often in these experimental setups, what makes the difference between what is attitudinal versus behavioral is whether the stake in your decision is real. That’s ultimately what makes it behavioral. So in these setups, we are inspired by our colleagues in social sciences, psychology, and so forth. So when they run studies, the techniques they utilize is imagine there’s an online store that you’re inviting people to come by. Then whatever they purchase in this experiment, they actually get that item delivered. Like, these are the things that make the stakes real. So we run a lot of these experiments, and we also do partner with firms. Right now, we also have customers who are quite excited to at least give us a glimpse of the behaviors that their users exhibit so that we can get a little bit deeper understanding of how people behave in these different platforms.

Vibhu [00:16:39]: I think on the customer side, they have a lot of data about their users, who has bought. They have the action data.

Vibhu [00:16:47]: Can you walk us through an example of what someone comes to you for? What questions would they want solved? Do you customize a model for them? Do you have something off the shelf? What does that look like?

Joon [00:16:59]: Today, when people leverage our models, it’s often to better understand the population of their interest. So usually, the start of the relationship, we come together and hear about what population they want us to model, right? So it might be that if you’re a CPG company that’s selling to all of the US, then maybe it’s fairly straightforward. You want to model the gen pop of the US. But at the same time, if there is a vertical or if there’s a market that they’re trying to go into, imagine, they want to better understand, let’s say, people in their 20s and 30s living in California. That’s a much more specific population. So we hear about this population, and we go recruit these people, with consent, and with incentives, and we collect some of their data and create a model of these people. Then what our product allows you to do is query them. So it can take as input a filter that is a description of the population that you want to talk to, just like the one I just mentioned, and an environment. The environment can literally be survey questions, behavioral experiments, It can be A/B testing. Oftentimes, the core use cases are things like concept testing, to start with. But also, people sometimes want to do focus groups or one of the fun use cases that we also serve is even modeling things like earnings calls for public companies.

Joon [00:18:21]: So these are the use cases that we often start with.

Swyx [00:18:23]: Concept testing, is that an established term? I’ve never heard of concept testing.

Joon [00:18:27]: Yeah. So it has to do with they have, let’s say, different messaging, different products, different ideas.

Swyx [00:18:32]: It’s like a marketing exercise.

Swyx [00:18:33]: Okay, got it. Got it. Politics?

Joon [00:18:36]: We do, have a strategic partnership with Gallup, and of course, Gallup is deep into policy space and so forth. Right now, we have not worked deeply with politics, like that area just yet, however.

Swyx [00:18:49]: I’m curious if there is demand or if they really would have different needs that somehow fundamentally don’t mix with your existing, users or people.

Joon [00:19:00]: I think there’s certainly demand.

Joon [00:19:02]: But we are very much mindful of how this technology gets adopted and the societal impact that we’ll end up having with this technology. And I do see politics as an area where a company has to be particularly thoughtful about the way they operate and make impact. So this is where we also want to make sure that we form enough of guardrail and perspective on how to leverage this technology before we go on to serve markets like the politics.

Swyx [00:19:29]: I’ll give people an example. one of my favorite shows is The West Wing. I don’t know if people have watched.

Swyx [00:19:34]: One of the key storylines is, like, the president has, multiple sclerosis, but they haven’t. they need to figure out how to disclose it. So they run a poll with a fake governor and ask people to respond on the poll,

Swyx [00:19:47]: They try to make decisions based on the results of that poll on, like, how well they’ll be received, like where, how should we play this?

Swyx [00:19:54]: And I’m like, well, I think those counterfactual things, I would use a simulation for this if I could trust it.

Joon [00:20:01]: For sure.

Joon [00:20:02]: In that show, how’d it go?

Swyx [00:20:04]: In that show, it was, like a foregone conclusion. They were like, “We know it’s bad. We just don’t know how bad.” And then the poll came back. It was like, “It’s really bad.” And then they just did it anyway.

Joon [00:20:14]: Part of it is to show, right? So you’re, you’re looking at the idea

Swyx [00:20:17]: Maximizing drama.

Joon [00:20:18]: How bad could it be? Oh, it’s horrible.

Swyx [00:20:20]: And to some extent, I think that is part of the trick of the, or the challenge or with being a customer of yours, which is that if I know it’s. if I roughly know and can intuit

Swyx [00:20:35]: What the effect is going to be, do I need you? What sensitivity of it, of effect do I need in order to make a decision, right? So for example, if I, my approval rating is 50%

Swyx [00:20:48]: And I, they have this negative piece, news item comes out, and it drops to 30.

Swyx [00:20:52]: If it drops to 20, if it drops to 40, do I care? No. It, I know it drops. It’s negative. So when do I care about simulations?

Joon [00:21:01]: You do something that’s clearly bad, that’s not popular, and people don’t like you, like, yeah, it’s like

Swyx [00:21:05]: You don’t need a simulation.

Joon [00:21:07]: Yeah. Well, so there are a couple of things. one is, there are use cases where, like every day, developers, designers, policymakers, marketers, every single day, they create assets. They create new products. And turns out, it’s many of the decisions in hindsight is obvious. Yes, of course this is bad, but we still run those studies because understanding the magnitude and understanding how acute something is quite difficult, even if, we feel like, of course, like this makes sense. this is the reason why we make so many mistakes. Like, every time somebody goes online and say something that has huge backlash, you look at that and like, “What an idiot.” However, it’s tough. That’s one. There’s also another aspect here, which is, again, this is the reason why simulation is different from prediction. In simulation, in the ideal case scenario. So what simulation is trying to show is it’s trying to show each step of the way or each step that we need to take to get to a certain outcome, right? So in the most advanced simulations, sometimes the next step that we’re suggesting might be quite counterintuitive. The analogy that I sometimes give, and I ground it in a more realistic example, but, I, as I mentioned, I’m a huge fan of science fiction, and I don’t know how, many of the audience members have read, like, things like the Foundation series by Asimov.

Swyx [00:22:37]: Oh, yeah. We’ve mentioned psychohistory a number of times.

Joon [00:22:39]: Okay, fantastic. So I might be, talking to the right crew. If you read Foundation series, literally the first act is there’s a group of scientists who have found out that, “Oh, our galactic empire is going to collapse, and we’re going to have 30,000 years of unrest.” And they run psychohistory, the simulator that tries to teach them, “Okay, how can we keep this unrest to a 1,000 years?” And they plan this out, and the first step of that plan is to get the scientists who say, “Okay, this is coming,” exiled into this random place in this, galax- galaxy.

Swyx [00:23:18]: Terminus.

Joon [00:23:19]: Exactly. And that’s so counterintuitive. Like, what a strange move that you literally sent the group of scientists who was raising voice around this potential collapse of galactic empire into nowhere. How is that the right first move? Well, it turns out in this particular simulation, that was the move.

Joon [00:23:40]: It’s these things, right? And the reason why these reasoning is possible is because you’re showing the step function or each step that results in a particular outcome. So really what simulation allows you to do in its highest form is you give it not a problem or question, like what would people answer to the survey? That’s not what we do. What we tell it is, “Here is a goal that we have. In the context of foundation, we want to keep the unrest to a 1,000 years. What is the path that we need to take now to get to that particular future?” And that’s what simulation allows you to do. Now, translating that into real market, imagine you’re a automobile company and you’re about to release a, EV, and you’re trying to understand, well, how do we market EV, to make sure that our stock price goes up? But what if the answer comes down that, well, you can market your EV in XYZ way, but that might change people’s perception around the cars that’s not EV and make your overall sales to go down. Not very intuitive, especially all you’re trying to optimize is EV salesss, and that’s the only thing that you’re tracking, then that might result in a completely wrong solution, or at least different solution than what you would have expected, whether it’s right or wrong.

Joon [00:24:57]: That’s the power of simulation.

Swyx [00:24:58]: For listeners, we covered a similar topic with Mikhail Parakhin from Shopify, where they are working on SimGym. I don’t know if he ever talked to you about it. it’s very similar.

Joon [00:25:07]: I

Swyx [00:25:07]: The goal is increased conversion, but then the journey is very unusual.

Joon [00:25:12]: Journey is unusual.

Swyx [00:25:12]: Yeah. The-- He’s trying to look for interventions on a shopping trajectory, which is similar to what you’re saying. Like, it’s not about the attitudinal, is your word for it.

Swyx [00:25:24]: It’s about behavior.

Joon [00:25:25]: It’s about behavior.

Swyx [00:25:25]: And that’s exactly the difference, right? It’s, like, not about the near-term direction about-- but it’s more about, like, how do you affect multiple turns of interactions.

Vibhu [00:25:35]: You had a good quote at the start about this as well. It’s not about people wanting to know the outcome. It’s about how they can change it, change the way to get there, something like that. But I wanna take it back to how do we know this is grounded? Like

Vibhu [00:25:47]: How do you run evals? How do you test that simulations come through? if I was to do the same thing that you described with, say, your favorite LLM, Opus, GPT-5.6, have some agent to map out these things

Vibhu [00:26:02]: How different are the answers we would get if I give it the same goal, the same objective, make a decent system? You’re saying that you need to change the model weight. You have your own solution to this. But how far off are we, and how do you check if it’s grounded? you have some interesting stuff on your site that points to how you run real evals, but if you could take us through that side. I think that’s one of the big concerns that people have. They’re like, “LLMs hallucinate.”

Vibhu [00:26:27]: “You’re just hallucinating layer after layer,” right?

Joon [00:26:30]: The way we do this, and this is the paper that we worked on after the generative agents paper that really became the, at least for Simile and also the field of simulation and synthetic panels, really became the foundation. Yeah, this is the paper. the paper is called Generative Agent Simulations of 1000 People. Here’s what we’ve done. For this paper, we brought 1,000 people that’s representatively sampled from the US to a virtual lab. And what we have done was we spent two hours collecting fairly wide-ranging data. In this particular study, we focused a lot on this interview data, that was, whose script was taken from this project called American Voices Project. And then we would also pair that with a lot of behavior data and so forth, whatever we can collect within two hours. And then we would send these people away for a couple of weeks. And during that time, I would use this data to create their digital twins. And I would bring the humans, participants back after 2 weeks and have them complete a battery of surveys, experiments, behavior studies. So we have the list here, which included things like behavioral economics games. We would run literally, like, Big Five personality test, General Social Survey. We would also go ahead and run the randomized controlled trials that were published on PNAS. And we would have their digital twins predict how the source individuals would have acted in these studies and surveys. And this is where we could replicate people’s behaviors and attitudes 85 percent as accurately as people would replicate their own. So that was the first really paper that gave this validated results that we can model individuals in an accurate way. And what we ended up finding now, of course, in AI space, so this paper came out at the end of 2024. AI space, a year and a half, 2 years, that’s a lifetime.

Swyx [00:28:24]: Yeah. Just, for listeners who are not seeing the YouTube, I just wanna say, like, the headline figure is 85 percent accuracy, like, which is a big improvement over all the other

Swyx [00:28:34]: Methods that you showed.

Joon [00:28:36]: But the part that was particularly striking to us, especially as we improved this technology even further, was the generative AI models like ChatGPT, Claude that’s coming out, it does give you the right foundation. However, what they do not consider is the true attitudinal and behavioral aspect of people, especially in the population that you care about. So what these models are really good at today is they’re trying to become the super rational, objective machines, right? So you go get their data from places like Mercor, Scale. You talk to professional programmers, scientists to create model that’s amazing at reasoning. That’s what they do. Simile doesn’t care about any of this. The models that we’re talking about here, what we’re trying to create are models that are as dumb as I am, right? So if I make some mistakes, the model has to make the same mistake.

Swyx [00:29:34]: Oh, that’s very hard.

Joon [00:29:35]: That’s very hard.

Swyx [00:29:36]: You’re solving Murphy’s paradox.

Joon [00:29:37]: That’s exactly. And this is a completely different data and training objective. This is also where we see quite a bit of discrepancy in the performance in human behavior prediction between the frontier models, Simile’s model, and the models being created in this space, where in some cases, the model performance of frontier models go all the way down to 20, 30 percent, especially if you go into that more niche population on topics that our customers would care about. On more gen pop, it might be around 50 to 60 percent. So it’s not very robust. Like, you wouldn’t want to make your decision off of these and these findings. If you can bring that up to 85 percent, that is ultimately what people end up getting very excited about.

Swyx [00:30:20]: Yeah. Do we wanna keep going on the paper, routes?

Joon [00:30:23]: Yeah, for sure. So the last one, was an interesting one. So this, paper was the follow-up paper that we had, to the 1000 agents paper, where the idea was now can we augment the models even further and post-train a model based on a lot of randomized controlled trials? So this was an interesting one. The data is always the most interesting part of modeling in many ways. The data that we got here was there’s this, there’s this platform called Open Science Framework. So some, the audience might be familiar with this. And there has been, especially in the social sciences over the past 5 years or so, there has been this concern around replicability of studies. And so it was a bit of a crisis, the scientists acknowledged, where we rerun the study and we don’t see the same finding.

Vibhu [00:31:12]: Oof.

Joon [00:31:12]: It’s tough. And the reason why it’s there-- that was often the case was there’s this survival bias where the papers that get published often need to maintain what we call the value of less than 0.05 in the experiments that we ran. That suggests that only-- there’s only 5% chance that the results that we saw is false positive. But the tricky part was all the papers that were not published, and there’s still a 5% chance that whatever we publish is totally just randomly generated. Like, there’s a 5% chance that, hey, this effect is not real, but it just happened to be real because of the sampling bias. So because of that, what scientists started to do was they started to register their studies. So before running an experiment, they would go to this platform and say, “Here is the data. Here is the population that we’re collecting, and here’s the hypotheses.” And they would just say, “Here is our hypothesis.” Like, “This is what we believe.” And you cannot retroactively change those hypotheses. This is what gives us more scientific statistical confidence that whatever effect that you ended up seeing is true. So that ended up creating this really interesting platform where there’s one platform that has now contains tens of thousands of real-world experiments and hypotheses. And a lot of these are really high-quality, like, professionally designed behavior studies and random- randomized controlled trials. So we got the data and the studies from this platform and used that to make a point. And this particular, model is not, something that we’re serving commercially because this was a part of the open science. But this particular data set, helped us make a point that by collecting a lot of these randomized controlled trials, that are really well-designed, we can make significant improvement in model’s capability to predict human behaviors. So that’s what this paper was about.

Vibhu [00:33:10]: Is this stuff done on a individual level? Like, do I need to tune the model per individual, per company? Is there foundation model changes and then some slight post-training? Anything you can share there?

Joon [00:33:21]: So this particular model was trained. the data we had at the level of individuals, but this particular model was trained. We experimented with both. And this is what we end up doing at Simile too. We always train 2, distinct model. One is what we call the population-level model. The other is what we call the individual-level model. And both take very similar input, which is the description of a subpopulation or individual and a stimuli. In this particular work, we’ve done the same. Here, the results that we are reporting are much more geared towards individuals because we do think that is a harder task in many ways, but that’s what we have done.

Vibhu [00:34:02]: You seen anything on the questions that humans can solve that models can’t solve? So like

Vibhu [00:34:09]: Currently, it’s, I live 5 minutes walk away from a car wash. It’s a 10-minute drive. Should I walk or drive?

Joon [00:34:16]: Huh.

Vibhu [00:34:16]: The model will say, “Oh, walk to the car wash.” And, you don’t have your car.

Vibhu [00:34:20]: Is anything like this a problem in simulation? You would assume, like, very simple for human to think about, but if the model is saying you should walk to the car wash, anything here?

Joon [00:34:32]: It’s less, what can we solve, but I think it’s more about what biases or mistakes do people make that models miss. Like, imagine that you are, like the. When I was still at Stanford, I lived in Palo Alto. So it’s about, I would say, 40-minute walk from the campus. You ask the model, “Okay, let’s go home. What can I, what can I do?” It would likely call an Uber or, give me, the bus time. But for the longest time, I really liked walking back. And the reason why I wanted to do that was not for efficiency. It really helped me think. And I like to walk for, half an hour or 40 minutes or so a day, where I just get to, just think about ideas, research, just get lost in my thoughts. That’s very human activity. Unless the model has seen that and understands the importance of that activity, it would miss these kinds of features. So that I think, is fundamentally what we’re trying to model. Like, what is fundamentally human might not be the most efficient thing to do, might not be the right thing to do, but things that make us who we are.

Swyx [00:35:43]: I’m curious if, there are some data sets that you really want that would materially help you. One version of this may be interesting, which is more valuable to you to acquire as a data set, all of LinkedIn, all of Twitter, all of Facebook?

Joon [00:35:57]: It’s a little bit hard to rank, in part because, there’s, there’s this product saying where no feedback is wrong because it teaches you something about your users. Doesn’t matter what feedback.

Joon [00:36:11]: I think it’s a little bit like that.

Swyx [00:36:12]: So just whatever is bigger.

Vibhu [00:36:13]: What about a different domain? Say it was. What about all of Amazon data?

Joon [00:36:17]: Oh, yeah.

Vibhu [00:36:18]: Shopping data, right?

Joon [00:36:18]: Shopping data. So Amazon data is interesting in that it’s very much behavioral, although, like, what people do on social media, you could squint and say that is also behavioral. But the transaction data is always interesting. It is also most commonly available, however.

Joon [00:36:33]: If we were to look at purely social media, like if you really, if I were, if I had to really pick, Facebook likely is interesting because I do think it is most a default version of people. Because you go to LinkedIn, it’s very much professional environment. So people put up their, they have their guards up, right? And that still is interesting because that is true human attitude and behavior, but it is not your base state. you go to Twitter- Twitter, people have their own crazy personas, or depending on who you are. Like, my Twitter profile and, persona is very much, initially was I was very much an academic. “Hey, I’m here to share my studies.” Now, I share, things that’s related to Simile. But Facebook is one of those more private space where people just connect with their friends. In that way, I do think it shows you a little bit more about who that person is. So if I had to pick, I’d likely pick, Facebook.

Swyx [00:37:30]: Yeah. And you’re interested in, like, the whole person and their background and philosophy. I, is it too clinical or too machine learning-oriented to just say this is just ways to inject variance and biases? The broad question, is, like, is this any better than a randomized, like, combinatorial explosion version? So we have a link to the Tencent

Swyx [00:37:54]: Billion persona paper, where they did not do any of the groundwork that you are doing.

Swyx [00:37:59]: They just did like a cross matrix of here’s all the professions in the world, here’s all the people, possible backgrounds in the world, do a dot product across all of them, and that’s it. That’s your prompt for a billion people.

Swyx [00:38:12]: This will do something. I don’t know if it’ll do what you do, but it gets you some way, some percent of the way there.

Joon [00:38:18]: So this was an interesting paper. Like, what I admired about this paper when it came out was the scale. And you do gradually want to be able to simulate really large societies and interactions. So the scale is definitely admirable. it is relying heavily on the known statistics that went into training the model. So to the extent that you believe that statistics is correct, this is not a bad way to go about this. But the thesis here, and this is something that we also have seen in the market, like if this works, then we have solved simulation.

Joon [00:38:54]: It,

Swyx [00:38:55]: Because I survey, like, okay, 5% of the US population is in construction.

Swyx [00:39:01]: The other 5% is in medicine, whatever, right? And then you just keep going down the list, and then you do the other side. 5% has, like, the big 5 personality

Swyx [00:39:08]: Of, like, neurotic or whatever. That’s it.

Joon [00:39:11]: That’s it. So if you believe that the underlying data set and the platform that we’re leveraging has all the right statistics, then this will have solved it. you’re at that point merely retrieving the knowledge that is already embedded in the model, in the model parameters. That’s not, unfortunately, what we see, where there is such detailed and also niche knowledge about people that if you just take one example, it might feel very mundane, but it’s quite rich when you put together, that you do need to do a lot of bespoke data collection to better understand people. And this is also, I think what makes this particular, job fun, which you want to deeply understand people, and the process of deeply understanding them requires a lot of attention to the details. And you do need to pay attention to and pay respect to the daily lives that people lead.

Vibhu [00:40:04]: I wanna talk about scaling simulation.

Vibhu [00:40:07]: So what can’t we simulate, what can we simulate, and how does scaling affect this? So how big are the models? What if we go from, 8B, like, couple 100 billion

Vibhu [00:40:18]: Like billion000 parameters, billion000? Do we get scaling? Any interesting emergence? Like, at a certain scale, at a certain amount of training, you uncover anything unusual and any learnings from that?

Joon [00:40:31]: What we are seeing is at Simile, so we do post-train our own model. The thing that we’re seeing is the early glimpse of scaling law in simulations. The more data about humans and more compute you ingest, you start to get predictive and predictable gains of the model performance in simulating it, simulating people.

Vibhu [00:40:51]: Ooh. We need a scaling law curve.

Joon [00:40:52]: It’s scaling law. Whenever you find it’s a beautiful thing. And we’re starting to see the glimpse of it, which is quite exciting. But if you talk about the ambition of simulation as a whole, it’s not merely about building a model. It’s about building a model, then creating the agents that become the individuals in a much larger ecosystem. So they’re creating this multi-agent simulation. Down the line, you want these multi-agent simulation to also live in a very rich environment, right? What we are really trying to get to at that point is, hey, can we create. All right, let’s do a time machine game again, and 5 years, 10 years into the future, can we create a simulation of 8 billion people living on Earth? I think that’s quite interesting. And that really is the vision. And once you get to that state, the questions that you can help answer for the society also start to change from my perspective. The answers are fundamentally about emergence of the emergent behavior of society and large groups of people.

Joon [00:41:53]: So the questions that I get excited by, and maybe this is a stodgy- a bit. I have my, academic side of me.

Joon [00:42:01]: And for me, it’s questions like, can we help solve climate change? If you look at climate change as a problem space, this is what we, like social scientists would often call it the wicked problems, problem where you have many actors with competing incentives for trying to make a very complex decision and coordinating that coordination decision. Very difficult to really solve in real life, which is also the reason why we couldn’t solve it. Can simulation help us solve that? Another one is, can we understand the signals for collapsing democracy, or can we understand or can we uncover the origin story of the monetary system? These are societal questions that we never really had a good way of answering. If we can create simulations of our society, you have to believe that these are the problems that we can solve. So that’s really the ambition of this field. And, I also think, yes, I think there’s a Nobel Prize to be won there, which wouldn’t be surprising. And I think there’s some amazing societal impact that we can have to help people make better decisions.

Swyx [00:43:04]: Nobel Prize in economics?

Joon [00:43:06]: In economics.

Swyx [00:43:06]: Oh, I see. I see. Rooting for you to write that paper.

Joon [00:43:10]: One of these days. But, one of the scholars that I was deeply inspired by, When I was coming into the space of simulation, is this scholar, named Thomas Schelling.

Swyx [00:43:23]: Schelling point?

Joon [00:43:24]: So the canonical example of the work that he’s done was he was one of the creators of agent-based modeling. So this was, like, in the 1970s and 80s. It’s very early days, but this was truly one of the first exemplars of simulations. And one of the canonical model from that time, and of course many of these simulations are trying to tackle the societal problems that’s most relevant for their era, it was called the model of segregation. So racial segregation was a big topic, that, we cared about. And what they’ve done was they created this grid world where they had red dots and blue dots. And these dots were, back in the day, like, they were the agents, and they had a simple rule that governed their behavior. If certain percentage of your neighbors are of different color and if that goes above certain threshold, then you move to a new location at random.

Joon [00:44:21]: One of the striking finding of this paper or this agent-based model was for the longest time, people thought the segregation within society was caused by explicit and overt racism.

Joon [00:44:34]: But if you look at this model, people’s preference towards living with people of the same color, that preference can be very minute.

Joon [00:44:42]: But the very small difference causes the society to segregate completely over time. This was very counterintuitive for a lot of people. And this particular work ended up informing housing policies. Mixed income housing, got really inspired by this work. And Thomas Schelling ends up winning the Nobel Prize for having laid the groundwork for very early versions of simulations. The opportunity that I do see here in the more scientific terms, is agent-based models for the longest, had impact in the 1980s, 90s, to some extent, early 2000s, but it has now gotten forgotten by the community a little bit. Because as you can imagine, red dots and blue dots is not really a rich description of people.

Joon [00:45:31]: But with the emergence of things like generative AI and, in particular, generative agents, we do have an opportunity to create these agent-based models that are high fidelity enough to help us make really complex decisions. And that’s the opportunity that I see. If that truly works, then yes, that is the work that will result in a Nobel Prize.

Swyx [00:45:53]: Yeah. For what it’s worth, and I grew up in Singapore. 80% of Singapore is in public housing, and public housing has, enforced racial quotas for exactly that reason, which is very interesting. okay, so we talk about scaling, we talk about all these, the agent possible applications.

Swyx [00:46:13]: I’m scared about the cost. if you even-- let’s just keep it to the US, about 8 billion people.

Swyx [00:46:21]: But, how much does it cost to model so many hundreds of millions of people?

Joon [00:46:26]: Oftentimes today, we don’t start at that scale, this stage of the, of industry and simulation as technology. But we can get our users extremely rich and meaningful insights even by modeling thousands, tens of thousands of people. And today what we do is every week we are collecting data on the scale of tens of thousands people’s data, and we have panel partnerships that gets us to tens of millions of people globally. So that’s what we do today.

Swyx [00:46:55]: And just as a side note once you’ve collected one person for one study

Swyx [00:46:59]: Can you reuse that same person for all the subsequent studies?

Joon [00:47:03]: That’s exactly right.

Swyx [00:47:03]: Okay.

Joon [00:47:04]: The beauty of this model and these agents is the fact that they are domain-agnostic.

Joon [00:47:08]: That what you’re really trying to understand is what is the fundamental nature of these people? What’s their social physics? And there are a lot of, a lot of, people that does change over time. Like, even, like, even things like, how many times have you gone have you been to, like, CVS the past week? that will change. But there’s so many traits about people that are also known to never change. Like, your risk tolerance doesn’t really change over time. It’s very consistent. So it’s these things that we’re trying to learn. But the scale we are operating is right now hundreds or, tens of thousands to hundreds of thousands. And in many of the core use cases that we are deployed in, and this is more than enough population, to cover those. Really, at that point, what you care about is less the number of people, but more do you have the right subpopulation of interest covered? And this is also the reason why people want a larger sample. It’s not because they want, stronger statistical guarantees. It’s more that can they filter down to any population of their interest. However, you can also imagine in 10 years, if we truly believe that the compute is going to scale, that we’ll have much more availability for compute, and our ambition for simulation is also going to scale accordingly, there’s definitely a reason for us to create an entire data center worth of simulations.

Joon [00:48:35]: Or in my hunch here is I do think in the next some number of years, we will start creating simulations that will cost as much as training a foundation model. But perhaps it’s going to be so valuable to the society that it would be a no-brainer. Right now, even today, like, we are training bunch of new foundation model just so we can say we trained one and we spent tens of millions. But if we can create a simulation at the level of society that would solve climate change, I would run that today. I would raise the money right now just to run that.

Swyx [00:49:10]: Amazing. the follow-up question is, does it also compound if you let the simulations talk to each other?

Swyx [00:49:18]: Or do they already do that today? They don’t, right, as far as I understand?

Joon [00:49:22]: It depends on what simulation you’re trying to run.

Joon [00:49:24]: In the multi-agent simulation setup, the agents do talk to each other.

Swyx [00:49:28]: Right, which is exactly Smallville, right?

Joon [00:49:29]: That’s right.

Swyx [00:49:30]: But a lot of times, for example, in commerce, you’re just by yourself, so there’s no point talking. which is way cheaper.

Vibhu [00:49:37]: But they use all these levels, right? Like, you decide what you will buy based on what other people around you buy and talk about, right?

Swyx [00:49:43]: It depends.

Vibhu [00:49:44]: It depends.

Swyx [00:49:45]: Again, I’m, I’m coming at this from a cost point of view. I’m like, “Oh my God.” Like

Vibhu [00:49:48]: I think

Swyx [00:49:49]: If there is, like, some combinatorial thing of, like, thousands of people talking to thousands of people, then that one million X’s might cost.

Vibhu [00:49:56]: I have a very different view as the cost point aside. Like, running these studies in reality is a lot more expensive, right? Running any study like this is you gotta have people do it, you gotta sign people up. It’s very expensive and sometimes, like, not feasible to run the study.

Vibhu [00:50:14]: But the outcome or the decisions you make are very expensive on them, right? So spend X million on something that, the overall process costs 100 million might as well, right? There’s, there’s a lot of value to be had there. It’s a small cost, but I’m excited on the cost side.

Joon [00:50:33]: To some extent, and when you deploy technology, you often want to deploy in a way where you can replace existing budget or you can make things more efficient, and that is the best way to deploy. However, the way you capture the long-term value of the technology is making the argument that, no, it’s the upside, that by making this better decision using simulation, you have saved yourself or made yourself hundreds of millions or even billions of dollars, and that’s a case to be made.

Vibhu [00:51:06]: Random tangent question. So if you’re doing a lot of inference, a lot of model multi-agent stuff, are you at the point where it makes sense to, train a model that’ very sparse? You’re expecting to do multi-million dollar runs. Are you thinking about this in model architecture standpoint or inference efficiency, or, you’re still at the research phase of it works, we’re not super there yet?

Joon [00:51:34]: Efficiency, we do think quite a bit about. this is technology that is deployed now in some of the largest enterprise companies in the world, and we do process significant number of queries, that are trying to, simulate the populations in the world. So efficiency is a consistent thing. we don’t want to over-optimize too early, so I wouldn’t say, like, this is the higher bid Right now, but this is definitely something that we think pretty carefully about.

Swyx [00:52:05]: Yeah. Are there other case studies? So we, you talked about CVS, talked about Gallup, Deloitte, Wealthfront.

Joon [00:52:12]: Wealthfront is an interesting one, because one of the things they were trying to do, they were one of the first customers that wanted to do product testing that goes beyond just asking people what they think about, let’s say, behavior experiments and so forth. So there, really what we had to do was reason about multimodal input, so images, but also you can also imagine, like, these agents traversing through Figma mockups or websites. So some of the things that our agents can also do is it can be given a domain, like, or, like, a website URL and go use it for a while. It’s these things. And Wealthfront was one of the first, customers, that was very excited about this possibility.

Vibhu [00:52:53]: What have people been asking? Like, is there any demand that we have not covered? Like, UI testing, right?

Vibhu [00:52:59]: I wanna try a new. I wanna ship a new feature, test the UI, simulate how people will do it. Any interesting things that you’re seeing demand for?

Joon [00:53:08]: Today, a lot of the demand does come from like, the places where people have historically used human panels, we can now replace with agents, and these synthetic populations. And this is not replacing human panel. in many ways, the simulation that Simile is building is grounded. So the way that I think about this is we are trying to represent humanity at scale. And in that way, the use cases are what we would expect, but it’s the scale of deployment that surprises me.

Joon [00:53:44]: Turns out there are so many decisions that people make every day in these organizations, groups, and we want to be able to say, “We listen to people. We have consulted our users.” But in reality, that is rarely the case because getting to people and asking them many questions, it’s difficult. It’s both costly, time-consuming, but most importantly, people are just not available. If I had to answer 1000 survey questions for this one particular, vendor, even if I wanted to do that, like, I would never do it. And that’s very much the case. What simulation can do is ensure that the voices of people are always represented in rooms where the decisions for them is made, right? So all the stakeholders of this particular product launch, ideally they’re consulted. That’s what this technology really is trying to enable.

Swyx [00:54:39]: In my mind, that means it skews towards more consumer focus, right? Like, anything with a wide enough customer base where you do benefit from the diversity that you represent. What are some rough statistics, just for people who are not familiar with this market in general, what’s the market size that. I’m sure you have some, like, rough numbers. market size is, like, a vague question

Swyx [00:55:01]: But, like, how much do people spend?

Joon [00:55:03]: So market research is a $100 billion industry.

Joon [00:55:06]: But the thing about simulation is not a tool for market research. Simulation is a tool for human decision-making. So the question around what is a TAM here is quite tricky, right? Because it’s easy to say, “Well, market research TAM is roughly 100 million or 100 billion.” so is it a TAM? And not really, right? Because in many ways, you’re trying to inform all human decision-making. You’re trying to inform every decision that are made about humans for humans. What is a TAM for that? It’s really unclear. And I’ll be honest. Like, I have a scientific background, I have a research background, so I didn’t come into the field calculating, oh, what is the TAM for human decision-making? But I just had to assume, well, if we can inform every decision that is made about human for human, that has to be big.

Swyx [00:55:58]: Some- something valuable.

Joon [00:55:59]: Exactly.

Swyx [00:55:59]: To some extent, you are a unicorn founder now, and you have to care as a CEO. But, like, I do think, like, yeah, when you go into these boardrooms with people that you’re quoting millions of dollars of contracts for, like, you have to say, “Well, here’s what you spend on humans-”

Swyx [00:56:15]: “. And here’s what we save you, and it’s 85% similar.”

Joon [00:56:19]: And certainly, the value case, is something that we care deeply about. Like, what is the value that we provide to the users and the decision-makers? But this is also where, like, as a founder, I think valuation only tells one very superficial aspect of the story, and I try not to think too much about valuation, in general, because that’s not what also motivates a team or certainly doesn’t. I’m, I-- Again, the interesting thing about researchers is we are happy living in academia, getting paid next to. we get paid okay. we don’t get paid that much, as a researcher here in academia, but it’s the impact and it’s the, it’s the value that we can provide to the individuals and the society that really drives us. And in that way, ultimately what drives us is the impact. Does the simulation we provide have a real impact in people’s decision-making in ways that progresses our society forward? If the answer is yes, then yes. that has to be great business, and we see that in numbers, and we do care deeply about that upside story, but that’s the heart of it.

Vibhu [00:57:27]: Do you have any timeline predictions? So we talked about scaling laws of simulations.

Vibhu [00:57:33]: You brought up, okay, maybe one day we can simulate how to solve climate change.

Vibhu [00:57:38]: Where are we now?

Vibhu [00:57:40]: If that’s not the end state, what is an end state, and what does progress look like?

Joon [00:57:45]: So what I sometimes tell people is simulation as industry, it feels a lot like where GPT-3.5, GPT-4 was, for the AGI saga, which is we have now technology that is powerful enough to do real damage on the verticals that we are tackling. At the same time, there’s a lot of progress that is yet to come. And that’s, I think, where this is. So the way I see it, I do think there will continue to be breakthroughs both in data, in algorithms, and there will be much more aggressive scaling that will also happen over the next few years. But I think that’s roughly where we are.

Swyx [00:58:27]: I think that was about the rough set of topics. Anything else that we should have asked you or you wish people asked you more about Simile?

Joon [00:58:38]: I think the, what’s, for me, what’s quite fascinating about simulation, it is very impactful technology, but it is also very interesting technology, both in terms of, like, what it means for human society, our philosophy. And the way I sometimes interpret simulation is. So going back to my background, I as I mentioned earlier, I started my career as a painter. it was a professional pursuit, and I did oil painting, for figures. So I got my training originally in the realism studios, and that’s what I spent a lot of my, years, doing. Simulation is a lot like painting, right? The best paintings teach you something deep about the subject that you’re trying to represent. And it is always not a perfect representation. It-- No painting is perfect. There’s always some small differences and discrepancy, but what it does is it tries to highlight the thing that matters the most about the subject.

Swyx [00:59:47]: The essential

Joon [00:59:49]: The essential essence.

Swyx [00:59:49]: Yes. He, you, he’s brought up some of your work.

Vibhu [00:59:53]: Just nice to put it up.

Joon [00:59:54]: Yeah. So these are some of the works. So this is from, my, personal website that I maintain when, I was still a researcher.

Swyx [01:00:00]: I think a lot of people will say, like a Picasso, like anything postmodern is, like, very much focused on the essence.

Swyx [01:00:09]: Right. yeah, but I don’t know if any one of these evokes something that you like to tell the story of.

Joon [01:00:15]: No, it’s one of those things where, each of these paintings, drawings, whatever it may be, it is trying to surface something about the subject that you feel deeply about onto the surface. when I was a painter, and artist, the topic that I cared really deeply about was, the more mundane aspect of human lives. This shows up in some of the, some of the work that I’ve done, where, like, I did this entire study of a rural town where I went around and took photos of people for not really doing anything special, but just living their everyday lives. I thought that was the most interesting thing. I’m somebody who has this perspective where, the world is oriented around this fractal shape, and you have two choices to understand the fractal shape. You either go outward and try to explore as much as you can to understand the broader shape of the fractal, or you go inward because, the outward resembles the inward, shapes. And understanding the mundane aspect of it was very much that. Simulation has a lot of this, right? You’re trying to understand even the most mundane aspect of people. When put together- teaches you something really deep about that individual and the society. So I think that’s what’s interesting about simulation, the way, the same way that AGI helped us better understand or really think critically about humanity and human intelligence, simulation is really an exercise of understanding more about human society and our collective lives. So that I find to be, yeah, particularly interesting.

Swyx [01:01:56]: Yeah. Now you’re reminding me that some of the best biographers, documentarians, and even photographers, they’re taking a photo of you.

Swyx [01:02:05]: But before I take a photo of you, I must spend-- I must, like, follow you for a week just to understand you?

Swyx [01:02:11]: Which some artists, some do. Part of your work, there’s a very famous book called Working. I don’t know if you’ve, been referred to it before.

Swyx [01:02:18]: It’s very famous, like, to the point of having a Wikipedia page

Swyx [01:02:23]: About this like, really depth understanding and interview of people as they, about their lives, which seems mundane, but is told in a very, compelling way. Yeah, 1970s as well.

Joon [01:02:34]: Okay. It was an amazing decade.

Vibhu [01:02:39]: Before closing question

Swyx [01:02:41]: Okay, here we go

Vibhu [01:02:41]: You said that you started Simile with your 10-year question, right? If we do that now, 10 years down, what can we simulate? What would you simulate if, like, if you’ve made significant progress, are there any questions outside of the ones that we brought up? Any- anything that you think is most impactful? Anything that you would go vision 10 years out?

Joon [01:03:03]: In many ways, as I mentioned, I am somebody who is very much impact-driven. So the what would inspire me is I would want to ask, 10 years later, what would be the most important societal question that we as a society have to ask? I would love to tackle that. Like, do we need UBI? That could be an interesting one.

Swyx [01:03:24]: Ooh, has anyone done that?

Joon [01:03:25]: Well, we were thinking about it.

Vibhu [01:03:27]: Can we get access? Can we just

Swyx [01:03:28]: So OpenAI, this is, like, just trivia now. Like, OpenAI, or I think Sam Altman funded a study on this

Swyx [01:03:35]: In Africa, and the answer was no.

Joon [01:03:37]: The answer was no. But, what, was it something about the implementation?

Swyx [01:03:41]: Yeah, I know. It was a skill issue.

Joon [01:03:43]: Or was it something about, But this is the thing. See, when Sam

Vibhu [01:03:46]: Funny news article

Joon [01:03:46]: Altman funded this particular,

Swyx [01:03:50]: He spent 14 million dollars? Oh my God.

Vibhu [01:03:52]: It’s a little more.

Joon [01:03:52]: Quite a bit. But this is the thing. This is the reason why you want to run a simulation. You spend 5 years, 40 million dollars on this one study and have one finding, but if you can run simulation many times instantly, then that’s the value.

Swyx [01:04:07]: I feel like that one could-- you could have done in a simulation. Like, if you can do the housing study, you can do the UBI one. Like, I, come on.

Vibhu [01:04:13]: I think sometimes people will spend the money because they wanna verify what you think, right? Like, sometimes you just wanna. Is it right? Like, you gotta test it.

Swyx [01:04:23]: Okay, closing question. What are the chances we are in a simulation right now?

Joon [01:04:28]: So it’s a fun question, and I assert at some point I just answer, yeah, we’re definitely in a simulation. But what I do, feel, however, is, whether we are in a simulation or not, that, I don’t think that makes our experience any less real. And I think that’s fundamentally, like, what I believe in. Maybe we live in a simulation, maybe not, but for

Swyx [01:04:48]: It’s real to us. Yeah.

Joon [01:04:49]: Yeah. For me, I don’t really care.

Swyx [01:04:50]: Yeah. Unless you die and you wake up in, like, the level higher or below.

Joon [01:04:55]: That would be interesting.

Vibhu [01:04:55]: I feel like you wouldn’t care. Once you die, then you find out you’re in a higher level.

Joon [01:05:01]: I worry about it when I die.

Swyx [01:05:04]: I think the other thing that. Okay, so I like the mathematical answer to this, which is, like, the, sheer number of possibilities that you are in a simulation far outweigh the sheer number of possibilities that you’re not.

Swyx [01:05:16]: Except for the simplest answer, which is, it is computationally very expensive to have you be a simulation. okay, great. You’ve been very generous with your time. Congrats on all your success. I met you just after your Smallville paper and had no idea that you could build, like, such an enormous company. And then now you’re like, “Well, it’s a $100 billion market, but that’s just where we’re starting.” So this is, very exciting.

Vibhu [01:05:42]: I think $100 billion market was not the term. That was only part of it.

Swyx [01:05:45]: Yeah, exactly. It’s, if you’re thinking too small.

Joon [01:05:48]: Well, I do believe that, maybe my final note here might be, again, I love science fiction. You look at any advanced civilization in science fictions, there’s 2 twin pillar, technology. One’s AGI in some form, and the other is simulation. So I think the market’s pretty big here.

Vibhu [01:06:08]: Tell us about the company. You guys just raised a lot. You’re half a research lab, half a company. you’re hiring. Where are you based?

Joon [01:06:15]: Yeah. So we’re based in Mission Rock, so not too far away from, where we are right now. So we’re in SF, but we are also bicoastal. So we have our, team. I would say our headquarter is in SF, and we have a lot of our technical talent in SF, and we do have a smaller office that just opened up in New York. We are, as a company, an interesting one in that today, there are AI neo labs and then there are AI product companies. Simile truly is both. So this is a company that was founded by 4 founders, myself, Michael Bernstein, Percy Liang, Lainie Yallen. Michael, Percy, and I are all researchers. So of course, Michael was one of the authors of the ImageNet, kickstarted the AI revolution back in 2013, has been instrumental in human-centered AI. Percy coined the term foundation model, and is a, one of the greats of the AI researchers today. And Lanie is my business counterpart, where she led some of the fastest-growing AI native companies from their seed to A and B. But we have this DNA at the company where the vision of the technology that we’re creating is continuously developing, that we are getting people who were my lab mates. We are about 60 people right now.

Joon [01:07:28]: 15%, almost 20% of the company population are just my lab mates from Microsoft Research lab.

Joon [01:07:36]: And we It’s quite fun because many of them then had gone on to OpenAI, Google Gemini, and these places. And so it’s been a few years since we really got together and had a chance to work together. But now they’re coming back and really building out this vision that I find to be quite exciting, and that excitement is shared. So there’s that motion at Simile where we are a group of researchers trying to do something that no one is working on that we find to be the most impactful potentially. But at the same time, this is, again, technology that can make impact today. So we have an amazing group of engineers, product people, and designers, who are sitting here with us trying to imagine what does it look like to help people understand what simulation can do and make real-world decisions with this. Having both and then deploying it to some of the largest customers in the world today, it feels quite unique.

Swyx [01:08:30]: Yeah, it’s very compelling. One part of it was this is the call to action. Like, who are you hiring? You’ve done part of it, which is you have-- you’ve got a very talented group. Who are you hiring? Like, what roles?

Joon [01:08:41]: So honestly, at this point, we’re hiring across

Swyx [01:08:43]: Everything

Joon [01:08:43]: All, section. we are always excited to bring on, amazing research talent.

Joon [01:08:49]: So if you’re interested in working with, our lab mates, we are always welcoming of amazing, researchers. But also we, hire, amazing engineers, that some of whom I, like, I respect the most. Many of them come from places where we have personal connections with, so many of the members are from Figma, Notion, Rive, and so forth, but also more broadly from the companies that we as a team have really admired. So engineers both in the product side, infra side, we’re all looking for those hires.

Swyx [01:09:24]: Well, lots of people. I think you made a really good case. So thanks, and, we’ll see you in the simulation.

Joon [01:09:30]: Amazing.

Joon [01:09:31]: See you all there.