Shuo Chen J.D, MBA | 交大美洲校友峰会系列讲座之五：展望科技的未来
In this alumni summit talk, Shuo Chen – a former Goldman Sachs banker, UC Berkeley faculty member, and venture partner – delivers a wide‑ranging discussion on the future of exponential technologies. She begins by demystifying AI, showing concrete demos (AI‑composed music, Google Assistant calls, automated restaurants) and then balances hype with data, citing Gartner’s 2020 report that AI created more jobs than it destroyed and explaining which roles (high‑creativity, high‑compassion) are likely to grow. She categorises AI applications into three pillars – data, compute (hardware/cloud), and algorithms – and illustrates the impact with a case study of Rescale, a high‑performance‑computing platform that saved Boeing $180 million and now serves aerospace, automotive, and life‑science customers. The conversation then shifts to blockchain, where Chen runs a quick true‑false quiz to expose common misconceptions and walks through the Gartner hype‑cycle for blockchain, noting that the technology is still in the “inflated expectations” phase but moving toward the “slope of enlightenment.” She highlights concrete ecosystem examples: the Decode nonprofit that bridges Berkeley and Stanford, interviews with industry leaders (Michael Li of Coinbase, head of data‑centers at AMD, Aave’s founder), and startups such as Arcade that turn NFTs into financial assets. Chen also addresses the metaverse, framing it as the convergence of VR/AR and blockchain‑enabled digital assets, and discusses what “living forever” might mean in a digital identity context. She stresses that the metaverse is currently a small part of personal identity but is growing, especially among post‑2000 undergraduates. Throughout, she stresses the importance of community building, new C‑suite roles like Chief Community Officer, and strategic networking via alumni and local ecosystems. She shares personal productivity advice – 80 % focus on core venture work, 20 % exploration (teaching, board work, hobbies) – and reflects on her motivation to invest in data‑driven life‑science ventures after personal family experiences. The talk concludes with a Q&A, logistical details (LinkedIn, WeChat), and an invitation to stay connected through Decode and future events.
Chapters
谢谢 大家， 谢谢 大家 听 得 见 哈， 谢谢 啊， 非常 荣 幸 啊， 今年 啊， 我是 很 快 介 绍 一下 我 自己 哈， 我是 呃 呃 上 海 交 通 大 学 啊 那个 船 舶 动 力 机 械 工 程 的 八 七 届 的 啊 校 友 啊， 然后 九 零 年 来 的 美 国 啊， 在 美 国 已经 三 十 年 了 哈 啊， 今 非常 荣 幸。 能 够 今年 能 够 啊 成 为 啊 我们 这 一 届 的 理 理 事 长， 能 为 各位 校 友 服 务 啊， 也 同 时 反 馈 给 啊 母 母 校 啊 啊 今天 可能 还 有一 些 是 我们 非 较 大的 校 友 啊 朋友 们 也 也 也 在， 所以 我 今天 很 快 介 绍 一下 我们 啊 校 友 校 友 总 会 的 一些 情况 哈。 我们 交 通 大 学 美 洲 校 友 总 会 的。 啊， 是 由 五 所 交 通 大 学 啊 组 成 的， 其 中 包括 上 海 交 通 大 学 啊、 西 安 交 通 大 学、 西 南 交 通 大 学、 北 京 交 通 大 学， 还有 啊 我们 的 这个 呃 呃 原 来 叫 新 竹 交 通 大 学， 现在 啊 刚 刚 改 名 成 名 扬 交 通 大 学 啊。 我们 在 北 美 地 区 公 设 有 十 个 分 会， 包括 多 伦 多。 新 英 格 兰 啊， 就是 我们 包 波 士 顿 地 区 哈， 纽 约、 华 盛 顿、 芝 加 哥 啊、 密 歇 根 啊， 然后 还有 啊 休 斯 顿、 加 利 福 尼 亚 南 部、 北 部， 还有 啊 西 雅 图。 我们 交 通 啊 大 学 校 友 总 会 啊 最 初 成 立 于 很 很 久 了 啊， 一 九 四 三 年 由 我们 三 位 交 通 大 学 校 友 啊 前 学 生 啊 还有 及 其 其 他 两 位 校 友 在 洛 杉 矶 建 立。 啊， 然后 呢， 不 断 的 发 展 哈。 现在 当 时 在 一 九 六 五 年的 时候 已经 达 到 四 百 名， 如 今 我们 已经 有 成 千 上 万 个 交 通 大 学 的 校 友 就是 居 住 在 美 国 和 加 拿 大， 所以 我们 的 校 友 会 也 已经 成 为 交 大 人 每 到 一个 地方 寻 找 到 的 温 暖 家。 啊， 作 为 我们 北 美 历 史 最 悠 久 的 也是 最 大的 华 人 校 友 会 总 会 的 组 织 之 一， 我们 啊 校 友 会 致 力 于 建 设 一个。 充 满 活 力 的 啊， 也是 啊 持 续 发 展 的 一个 总 会 机 构 啊， 为 我们 回 馈 校 友、 回 馈 母 校 做 服 务 啊。 我们 从 啊 呃 一 九 六 五 年 开始 啊 办 第一 次 校 友 联 谊 会 以 后 以 来 呢， 每 隔 五 年 我们 在 啊 啊 挑 一个 北 美 的 一个 大的 城 市 吧 举 办 我们 校 友 盛 会。 啊， 也 变 成 了 我们 全 球 交 大 校 友 会 所 熟 悉 的， 并 且 为 此 为 交 大 人 北 美 交 大 校 友 们 为 骄 傲 的 一个 交 流 的 平 台 啊 啊！ 所以 这是 我们 交 大 大 学 啊 的 总 会 的 一个 简 单 的 一个 介 绍 啊。 今天 呢， 我们 呃 也 利 用 这个 机 会 哈， 然后 啊 也是 我们 呃 交 通 下 呃 我们 的 校 友 会 组 总 会 主 办 的 啊 系 列 的 讲 座 之 一。 啊， 今天 我们 的 嘉 宾 哈， 我们 简 单 介 绍 一下 为什么 我们 请 来 今天， 为什么 这个 内 容， 为什么 我们 今天 请 的 陈 硕 啊？ 因为 啊， 这是 大家 也 知道， techn ology 现在 发 展 非常 非常 快。 Does n 't matter you like or not, it 's coming 啊！ 所以 我们 说 未 来 已经 到了， 嗯， 你 来 不 来 啊？ 所以 今天 呃， 我是 非常 一个 非常 偶 然 的 机 会 能 够。 啊 啊 认 识 了 陈 硕， 是 我们 在 我们 啊 我是 在 安 永 美 国 工作 哈， 是 安 永 美 国 的 合 伙 人。 我们 呢 啊 今年 刚 刚 在 三 月 啊 的 第一 个 星 期 有 一个 星 期 的 我们 叫 techn ology 的 一个 program， 我们 叫 dis rupt ive technology program， 是 我们 把 全 球 的 六 十 个 全 球 合 伙 人 飞 到 硅 谷 啊， 然后 一个 集 中 的 三 天 半 的 一个 活 动。 啊， 其 中 啊 有 请 来 各 个 大概 请 了 十 六 个 spe aker outside speaker， 然后 我们 陈 硕 是 唯 一 的 一个 woman and minority and a Chinese face speaker. That was amazing speech. 啊， 我 我们 所有 的人 都 被 惊 到了 哈， 所以 我觉得 啊 分享 是 一 种 美 德， 所以 我 第一 时间 把 他 请 过 来。 当 天 晚上 我 跟 他 聊 到 十 二 点。 啊， 说 服 他 了， 请 请 来， 他 也是 非常 乐 意 为 我们 校 友 会 来 分享。 所以 从 啊 他的 一些 啊 他 我 让 他 来 自 自己 来 介 绍， 他 不 单 是 一个 entrepreneur， 自己 拥 有 一个 啊 呃 一个 VC 的 firm 啊 是 general partner， 然后 呢 又 是啊 UC Berkeley 的 faculty， 自己 也是 UC Berkeley 毕 业 的 哈， 然后 呢 又 是 在 又 是 荣 获 这个 thirty under thirty。 啊， 然后 在 各 种 各 样 的 board 上 任 职， 呃， 所以 then all in all she 's just a super efficient and multit asking person and very very successful. So, uh, given that I 把 这个 话 筒 我 交 给 今天的 主 主 主 持 人 啊， 于 迪 你 来 啊， 谢谢。
谢谢 谢谢 小 青 学 姐， 小 青 学 姐 她 自己 本 人 也是 非常 厉 害， 她 在 安 永 做 了。 超 过 二 十 五 年 是 大 趴， 也是 我们 非常 敬 仰 的 学 姐 啊！ 谢谢 她 邀 请 陈 硕。 呃， 在 我 介 绍 陈 硕 之前， 我 还 想 要 呃 介 绍 一下， 就是 这 次 我们 组 织 活 动 的 团 队， 还 包括 我们 这 边 北 加 州 的 谢 东 学 姐， 她 现在 是 总 会 的 会 长 ； 还有 我们 的 Frank 师 兄， 他是 我们 总 会 的 这个 副 会 长。 对 啊， 谢谢 谢谢 我们 的 这个 te am members。 那 陈 硕， 我 跟 陈 硕 呢 其实 很 早 就 认 识 了。 对， 之前 有一 些 线 下 的 e vent 就 觉得 他 特 别 的 char ming， 非常 有 亲 和 力。 不 光 有 一个 s uccess ful career， 而且 他 就 真的是 跟你 交 沟 通 的时候 让 你 如 沐 春 风， 而且 特 别 特 别 年 轻。 所以 啊， 其实 我想 大家 应 该 都 已经 看 过 他的 那 些 tit les， 那 我就 不 赘 述 了。 不 赘 述， 而且 啊， 其实 今天 他 还 在 悉 尼 出 差， 所以 特 别 感 谢 他 今天 抽 时间 来 跟 我们 做 一个 分享。 OK, without further ado, I will pass the floor to Sh uo.
非常 非常 感 谢， 也 非常 开 心， 在 三 月 份 的时候 有 机 会 可以 认 识 这个 小 青 姐， 尤 其 是 在 安 永 的 确 是 能 做 到 大 趴 的 这个 华 裔， 真的 非常 非常 不 容易， 跟 小 青 姐 学 到 很多。 然后， 玉 丽 也是 真的 认 识 很多 年 了。 呃， We 've actually worked on a number Number of events together, including 去 年的 时候， 这个 在 美 国 Stop Asian Hate 也是 非常 大的 一个 一个 movement。 然后 也 跟 于 丽 一起 做 了 一些 小 小 的 贡 献 和 努 力 吧。 其实 呃， 于 丽 真的 做 了 很多， 然后 也 非常 开 心 今天 有 机 会 可以 跟大家 做 一个 简 单 的 分享。 要 不 我 简 单 的 自 我 介 绍 一下 吧。 我 其实 是 在 北 京 出 生 的， 但是 呢 很 早 的时候 就 来了 西 汉 这 边。 我 大概 小 学 一 年 级 的时候 来 的 这 边， 所以 其实 说 实 话 一直 觉得 我 中 文 讲 的 不是 特 别 的 好。 所以 可能 待 一 会 儿 讲 到 这个 更 多 technology 的 内 容 的时候 会 转 换 成 英 文， 但是 其实 咱 们 可以 尽 量 中 英 文 交 叉 的 去 做 这样 的 一个 交 流。 然后 也 非常 希望 今天 有 机 会 可以 更 加 了 解 大家 的 背 景 啊， and see how I could be helpful。 然后 我是 最 开始 自己 创 业 做 了 一个 呃 f int ech 的 一个 创 业 公 司， 我们 当 时 主 要 是 呃 create curriculum and software。 主 要 是 呃 服 务 于 非常 大的 金 融 公 司 和 大 学， 然后 也是 通 过 这样 一个 机 缘 巧 合 的 机 会 开始 讲 课。 所以 在 B er keley 做 f ac ulty 和 在 B er keley 讲 课 已经 有 大概 十 多 年的 时间 了。 呃 最 一 开始 是 在 商 学 院 讲 课， 然后 现在 是 属 于 这个 工 程 学 院 旗 下 的 f ac ulty。 在 创 业 之 后 呢， 呃 觉得 发 现 自己 其实 有 很多 需要 学 习 的 地方， 所以 呃 后 来 加入 了 高 盛， 主 要 是 负 责。 和 这个 很多 中 美 之 间 大的 科 技 公 司 上 市 和 M A 的 一些 呃 项 目， 包括 像 阿 里 巴 巴 的 上 市 啊， 还有 这个 滴 滴 和 这个 滴 滴 的 一些 战 略 融 资、 美 团 点 评 的 这个 合 并 等等， 都是 当 时 我在 高 盛 主 要 负 责 的 一些 项 目。 所以 从 自己 创 业 非常 非常 小 的 公 司 到 高 盛 负 责 这个 很多 跨 境 非常 大的 项 目， 到 现在 过 去 七 年的 时间 一直 在 做 自己 的 这个 风 投 基 金， 其实 还 蛮 有 趣 的 一些 一些 pers pect ive。 From looking at very, very small companies to working with some of the largest companies, to now again helping introduce smaller companies to larger companies. Hopefully, I can share some interesting stories with all of you. With that aside,其实今天是想跟大家做一个非常简单的一个分享。虽然我们今天讲的是这个 introduction to exponential technology, I want to make sure it's relevant for everyone. really here, um, that's sitting in the session today. I know we have actually nearly 90 people here now already. Um, so to make sure that it's relevant for all of you, I do encourage all of you to send questions. I have the chat open. So if there's anything that you want to learn more about or you want to pause, we can totally address this and make this as interactive as possible. And to get us started to why we're even talking about exponential technology today, I want to just start with this overall statistic that I think is quite meaningful for why we should even be thinking about the future of technology. And that is this latest data that's released by IBM. So let me go back. That 85% of jobs that will exist in 2030 were not even invented just 5 years ago. And I start with this data because when we think about the future of technology, we tend to think, oh, it's very far away. But if we just look at people that are around us, for example, people who are in C-suite positions, we realize that oftentimes they're in C-suite positions of jobs that didn't exist when we first graduated. For example, I currently sit on the board of a nonprofit alongside the chief privacy officer at Google. And we were chatting and he was saying, well, hey, data science didn't exist as a field and data privacy certainly didn't exist as a field when he first graduated. So being able to now be the chief privacy officer at Google is hugely different because it's completely changed and shaped by technology. So this is really where, uh, why we're talking about future of technology today. And what I'm hoping to spend time on today, and we can adjust based on what would be helpful and interesting for everyone here, is I want to talk a little bit about how technology augments the work that we do in terms of empowering the way that we do work day to day. I want to spend a little bit of time talking about how technology realigns incentives between all these different stakeholders. And I want to give you some practical action steps for how we can adapt to all of these changes very quickly. And in order to really launch into this, I want to get to know all of you a little bit better. So I thought it would be fun to start this first section with a very quick game so that I can get to know how much you already know about everything we're talking about and how, how comfortable you are, how interested you are in talking about all of this stuff. So in this first section, I want to talk a little bit about how AI has been able to augment the work that we do. And I think by now everyone has certainly heard of AI. And I want to start with a very quick game, but I do see a hand that's raised. Maybe I'll take the raised hand first. I think I see Wen Hua has her hand raised. Maybe you could unmute.
Sorry. Hi.
Hi.
Hi. I just want to sort of ask a question which is really tangential technology, but it's not exactly on the exponential technology. Since you actually worked for Goldman and worked for a few of the ADRs in the boom years, and I just want to see if you have any thoughts on where we are in the ADR market. Why is this permanent damage, or this is just part of the market cycle? And if you can share any of your thoughts where we go from here. Thanks.
That's a really good question. So for context, I have to say that for the past 7, 8 years, I've been on the venture side and mostly working with early-stage companies. So I'm not sure if I'm the best person to comment on ADRs, but I have worked with them in the past, dated as of a couple of years ago. So from all of my former Goldman friends, some still at Goldman, I will share some top thoughts, top-level thoughts. So number one, I think ADR from day one has been a really interesting category of assets because it's essentially in a category where it really relies on a lot of cross-border opportunities, at least in terms of the clients that I was covering. Most of them were technology companies headquartered in China looking to list in the US. Now, that is increasingly becoming difficult, mainly because the level of geopolitical tensions that have risen since then and the different kind of government policies that have been established has made it significantly more difficult to pass both political as well as financial regulatory oversight. So it has been increasingly difficult to list as an ADR just in terms of pure policy. The second thing I will say is that there's also more robust hubs of financial service centers that are now emerging in different cities around the world, where you have certain companies that will list in specific areas. If you talk to some of the largest founders in technology, they will tell you that they will gravitate towards your local center. So for example, a lot more unicorns in China now are looking at opportunities to list in Hong Kong, or even locally, domestically in Shanghai and Shenzhen instead of abroad. So that's very high-level comments on the world of ADRs. But of course, as everyone knows, technology stocks have been seeing a lot of changes. Where I think to relate that back to our topic today, where that's relevant for our conversation today is that because of how quickly technology is changing, we see 2 really interesting trends. One, we see because technology stocks have been struggling, particularly here in the US, actually including a lot of publicly listed companies here in the US, What we see is a huge movement of talent that are now leaving the umbrella of large companies to start their own startups. So that makes our conversation today really relevant, because even though we're talking about the future of technology, we're really talking about future of technology driven by startups. So now that a lot of not just junior, but very senior people are leaving these large companies, they have this opportunity for the first time to actually build very meaningful startups. And this is really where early-stage capital is important, not as much as later stage. And then the second thing where that is relevant to our conversation today is when we look at hubs of innovation and hubs of technology, we've certainly seen different geographic regions kind of adopt or own different segments of technology. So for example, in the US, we've certainly seen a lot of technology driven by— a lot of innovation driven by underlying technology. So for example, whether it's at the core core AI developments or at the core blockchain developments, whereas a lot of the larger companies we still see emerging out of Asia and Southeast Asia are still predominantly operational driven. So of course, a lot of Asian companies are talking about thinking about AI for science, more technological innovation. But really, at the core, some of the largest companies that are at least publicly listed are still driven off of operations. So it is increasingly the case where I think there will be opportunities cross borders, both in China as well as in the US, where companies will get to learn from the other. For example, a lot of US companies are now learning best practices from operations, product growth from Chinese companies, and vice versa. A lot of Chinese companies now developing their own research and development centers to really hunker down on the deep technologies that previously we didn't have time to dive into. So I think this is more of an opportunity for cross-border learning than anything else.
Yeah.
Politics and policy aside.
Yeah, but so thank you so much, Shuo. For the audience, if you guys don't mind, if you can type your questions in the chat spots, then we can save all the questions at the end. So let Shuo finish her presentation first. Okay, you can be— you can use— you can do a short one, you know, you can— we actually count on you to control the speed. Okay. But otherwise, I think Q&A, let's leave that in the end. Okay, thank you.
Great. Yeah, that sounds good. Maybe what I'll do is I'll breeze through the kind of the session that I had planned so that we have more time to really discuss, because I'd love to hear your thoughts too. That was just my high-level thinking, and it could be totally off because you're in different industries, say healthcare, where you're seeing a different trend. So love to learn more from you as well. So I wanted to start with a quick game to get to know all of you a little bit better, and you're very welcome to use This is a really simple game where what I'm hoping to do is make 5 statements. And out of these 5 statements, there is one false statement. So promise, 5 statements, 4 are true, one is false. And would love to see you put in chat the number of the statement that you think is false. So this is really to help us take a quick look at the ability of AI in many ways to replace the work that we as students do. humans otherwise do. So statement number 1 is that AI is capable of now writing symphonies. Statement number 2 is that AI can do calls on behalf of humans. Number 3, AI can completely automate and run restaurants. Statement number 4 is that AI has the ability to make humans dance, even though humans don't know how to dance. Statement number 5 is that AI has the capacity to read your mind. So these are the 5 statements. I'll say it one more time, and then would love to see which one you think is false. So AI can write symphonies. AI can call, even to the extent humans don't know that it's AI calling. AI can automate restaurants. AI can make humans dance. AI can read your mind. So lots of guesses streaming in. That's great. So I see a lot of 3s and 5s, actually a lot of 5s. Mostly 5s, occasional 1, maybe like a 4 question mark, mostly 5s. Okay, great. So none. I promise it's not a trick question. I promise there is one false statement. So great. If you have more guesses, feel free to put in comments as well. But let me reveal which one is false. So the first statement, this is Ava Technologies, and— I'll play this music in the background. But what's really interesting is that AIVA as a company stands for Artificial Intelligence Virtual Artist, and this is one of the many symphonies that AIVA has written purely by AI algorithms. So this particular piece is called Genesis. It's a symphonic fantasy in A minor, Op. 21. The name itself looks very much like it is composed by a human, but this whole symphony, surprisingly, is actually written by AI. So of course, all of you can come to your own decisions as to whether or not you like this symphony, whether or not AI-written symphonies are good. But currently, it is— it's completely realistic and possible for AI to be writing symphonies. The second statement about AI being able to make calls, many of you are familiar with this, I'm sure. As I said earlier, we're likely already seeing this from Google's announcement. Where they've developed AI that is able to make calls.
Hey, you want to ask Google?
So this is Google Assistant's latest announcement, and this is a call that you can see in real time.
Oh, happening out here. Hi, I'm calling to book a women's haircut for a client. I'm looking for something on May 3rd. Sure, give me one second. Mm-hmm.
So on the left-hand side, what you're hearing is the Google Assistant.
At 12 PM. We do not have a 12 PM available. The closest we have to that is a 1:15. Do you have anything between 10 AM and, uh, 12 PM? Depending on what service she would like. What service is she looking for? Just a woman's haircut for now. Okay, we have a 10 o'clock. 10 AM is fine. Okay, what's your first name? The first name is Lisa.
So what we saw
just now in this whole interaction
is a completely
automated call by AI.
So many of you might have seen this live if you were there in the conference, or you might have seen this in the news afterwards, but currently it is certainly possible. Almost all of you are right, none of you thought this was false. We can completely have AI automate calls and have the human on the other side having no idea that they're speaking to AI. Which then brings us to the third statement. Some of you thought this may be false, but very few of you. So it currently is possible to leverage AI robotics to automate entire restaurants. And actually, this very first one is a Haidilao that operates in Beijing. And this entire restaurant experience is automated. So I'm curious to hear how many people have actually been to this restaurant before. If anyone has actually been to this one, feel free to text in chat. It's a really fun example. Um, and then that brings us to the 4th statement, um, which AI can make humans dance. So this is actually a research project that's coming out of UC Berkeley now under our College of Engineering, where what these PhD students are doing is basically taking source videos. Uh, this project is called Everybody Dance Now. And what they're doing is they're taking a source subject and they have a target subject of a person who doesn't know how to dance. But what they do is they map the source video to the target video to make people seem as if they're dancing, even though they've never danced in their lives. So technically, the first 3 statements are true, but the 4th statement is false because you're not actually making the human dance. I'll turn down the music so hopefully you can hear me talk a little bit better. Um, but this is a classic case where It's technically false because it's not like the human is actually dancing, but AI is certainly making it seem as if she's dancing. And what's really interesting is if we fast forward in this video a little bit, you can actually take videos of expert dancers and make people who have never known how to dance in their lives really look quite close to as if they are dancing. And actually, this one in particular that I really like is This guy doing his PhD, who I've met before, he's a UC Berkeley PhD student. He's never danced in his life before. And on the right-hand side, this is how he's configuring the system where he's like literally making swimming poses. But in this case, you're mapping Bruno Mars's music video in real time to make it seem like he's actually dancing, even though he's, in his own words, has never danced like that in his whole life. And If you look at the different source videos that they take, obviously this is a pulp music video, but they're able to take much more complicated professional dancing and apply it to the target video as well. And it's actually really fun because in this one, they're taking this video of a ballet dancer. He's configuring the system like this, and immediately he looks like he's dancing ballet in real time. So this is a really, really fun example. And of course, with our visual eye, with this video enlarged, you can kind of tell that there is a little bit of weirdness. But as AI improves over time, it can certainly make it seem like you really have a dance video. And so if the fourth statement is technically false, then this means that this last statement is true. This is also a project that's being done out of UC Berkeley. It is currently possible to read someone's mind with neuroscience.
Our main focus is on the visual system, and vision is a really important thing to work on because humans are very visual animals. And when humans have damage to their visual system, they're really incapacitated. And natural vision is a little like watching a movie. So essentially, we were trying to build a model that describes how your brain responds to movies. So we've, uh, The external world is a certain kind of movie. We want to be able to build a dictionary that will allow us to accurately predict what the brain activity would be that would be evoked by that movie.
So what they're doing here is they're taking in all the brain signals from when someone actually watches a movie. And what's interesting is they, by observing brain activity, they can backtrack what you're likely seeing or thinking about.
So the brain activity is painted here on the side of the brain, but the brain's kind of crunched up in the skull, so it's hard to visualize. So we kind of inflate it and then we flatten it out, and we can see a flat map of the visual cortex and the rest of the brain as the person watches the movie. The problem here for us is to translate between these movies that occur and this pattern of brain activity that occurs. And in the study we talked about in our paper, we just modeled this very back part of the brain, the early visual system, primary visual cortex. And that part of the brain responds to the little local features in the movie, little edges and colors and pieces of motion and texture. But this part of the brain doesn't know anything about what the objects are in the movie. First, we put people in the magnet for several hours and we show them movies.
So what he's describing now is exactly what he does. And the result is really interesting because, um, at the beginning, they, they ask people to watch movies, but here they're able to reconstruct what the movie they're watching is based on the
Right, is our reconstruction. When the movie that we showed has a fairly common object, like a person, our reconstructions are actually fairly accurate. When the movie that we showed is something rarer, as you'll see in a second, like this abstract thing, then our reconstructions are coarser.
So that's really the project here that's been running for a couple of years now. So of course, it— can it fully read your mind in terms of what you're thinking about? Not fully, but it certainly can take a graphic image of what you're seeing. And leveraging AI, they're able to reconstruct what you're likely seeing. So in many ways, this is the beginning stages of AI getting to an ability to read your mind. And I see Eugene's comment about, you know, a challenge of this is whether or not people could pretend and hide their actual mind. So it depends on which part of the brain is being read. So as the professor was explaining, they were really just reading one specific region of the mind that's particularly responsive to visual stimulation. So in terms of other thoughts, that would, of course, be a completely different story. But I share all of these examples to look at what AI is capable of currently doing, including writing symphonies, making calls, automating restaurants, making people look like they're dancing, and potentially even reading your mind from a visual stimulation point of view. Because it fundamentally comes down to the question of what does it mean to be human with the future of technology changing so quickly? Quickly. And I asked this question because there are things that AI is really good at, and then there's also things that AI is really bad at. In fact, it's kind of fun to also look at what AI is bad at. So currently, this video is here, but really what AI is bad at is it's not very good currently at really looking at writing funny stories. So this is also a research project that came out of UC Berkeley where Professors are trying to develop an AI system to be humorous and to tell jokes, but they're really bad at writing funny jokes. In fact, they oftentimes write unintentionally funny jokes. Like, this is one that the AI system wrote: Henry Squirrel was thirsty. He walked over to the riverbank where his good friend Bill Bird was sitting. Henry slipped and fell in the river. Gravity drowned. The end. Right? So AI is very, very bad at humor still. That's certainly something where humans are still better at. And AI currently is very bad at talking to each other. In fact, when you look at AI talk to each other, This is a research piece that came out of Columbia where they just hooked up bots to talk to each other.
How are you? How are you?
I'm okay.
That's good. Are you good? Yes, I'm good. I just said I was. No, you didn't. I said I was. So you are a robot? No, my name is Cleverbot. Yes, you are a robot and your name is Cleverbot. I am a robot. Yes, I'm making a million robots.
So basically, this conversation goes in a very interesting direction very quickly when you have chatbots talk to each other. So even though AI can make calls, they're really not that smart yet when you look at the bigger scheme of things. And then also, when you look at the world of robotics, this is research that's done by DARPA, the US Defense Department, and each of these projects costs millions of dollars to put together just to perform very basic tasks. From a defense perspective. And this is, uh, really just basic things of walking on difficult terrain, opening doors, just walking forward. And you have these multi-million dollar robots that are just toppling over trying to enter doors and folding right over. So I shared this example to start a whole conversation because there's this really interesting dichotomy that happens in mainstream media where on the one hand it's either AI is extremely powerful and it does everything or AI is nothing and it's not capable of doing anything. And I start with this quick game to just get you guys all hopefully interested, because both are true. There are many things that AI is very good at in ways that we couldn't have imagined, but there's also things where AI is very bad at it. It is not capable of doing nearly enough. So with this more balanced view of future of technology, when we really think about AI, it's not like in pop culture where it seems to paint this world where AI is going to take over humans. And it's certainly also not like a lot of media headlines where it's saying AI is automating all these jobs and humans are going away. The more balanced approach to this is to really look at the underlying data and see the exact impact of AI as technology. So this is the latest data that's published by Gartner. And this is looking at the number of jobs that are gained versus lost as a result of AI. And you can see that 2020 was the first year where there were more jobs created by AI than the number of jobs that were destroyed. So if we were to understand this framework a little bit better, I think this is a really great 2x2 of just helping us understand what the kinds of jobs that are being created are, where the opportunities are, and where the kind of jobs that are being destroyed are. So when we look at this graph of plotting out how much compassion is needed on a job versus how much optimization or creativity and strategy a job requires, we can really have a much better sense of what jobs are being destroyed. In fact, jobs that can benefit from a lot of optimization, jobs where you don't really need a lot of compassion, these are the jobs that are really starting to go away and become automated because a lot of these jobs weren't that much enjoyed by humans to begin with. However, jobs where you need a lot of creativity, strategy, and compassion, where it's really important for you to build a personal relationship, that is the kind of jobs that we see increasingly growing and being created going forward. And, uh, of course, at the end of the day, AI is all about replicating humans. And it's really for us to think about what specific function is being automated. Sometimes it's our ability to see, and that's computer vision. Sometimes it's our physical form, and that's robotics. Sometimes it's our ability to understand language, and that's natural language processing. But at the end of the day, it's for us to really think about what kind of applications we're adopting this. So these are some of the most common applications we see here in Silicon Valley. Having AI work on basic pattern detection, foresight, customization, decision-making, all the way to end interactions. And I like to categorize applications really into 3 main areas, where either you're doing something with data, where you're having more data, better data, more transparency in data, or the second category is where it relates to hardware optimizations, where you're really thinking about compute and cloud computing, or supercomputers and hardware, the physical compute. Or the last compute— the last piece really around algorithms, which is how these algorithms or how these programs and how AI learns over time. So I tend to break down all of these applications into these 3 main categories. I want to just give you one quick example of a startup that we've worked with since the very early days that's really thinking about leveraging all of these technologies to really drive AI in a way that it empowers the work that we do as individuals. And after that example, I will pause just in case there are questions or things that you're curious about. But the example that I wanted to share with you is a founder that we've known for, for a long time. He really started his career from a hardware engineering perspective, and he had built or had been the core engineer to designing the wing of the Boeing 787. And he was based out of, uh, obviously the US, and he was part of a much larger team. And what they would do in terms of hardware design is, after he is done designing all the wing specifics, they would come together regularly across all departments to assemble all the different components of the plane together in one. And what he realized was that there were a lot of inefficiencies in terms of how they were testing these plane models because You would have the wing design, you would have the tail design, you would have the body, you would have all of these different components. But when you actually assembled everything in a hangar and tested it in a physical environment, you realized that there were tweaks that the wing design department should have done weeks ago so that it interacted better from an aerodynamics perspective with the tail, for example. So what he ended up doing was, instead of going through this very tedious and non-efficient hardware design process, he ended up designing software that would allow all these different departments to leverage software from purely a data analytics and virtual environment perspective to test for all of this in a virtual environment. And the software he designed ended up saving the wing department $180 million and 3 months in research and development time. And so he thought, well, look, I just proved the efficacy of this software. So he then brought the software to C-suite at Boeing and said, hey, we should really adopt this software across all of Boeing. We should really think about, look, there was so much money and time saved just within the wing design department. Shouldn't we go and adopt this across all of the different departments within the company? And after about 6 months of conversations back and forth, C-suite ended up coming back to him and said, hey, look, we think this is really cool that it saves so much money and time. But ultimately, we cannot move forward. And he was— this particular individual was, of course, mind-blowing. He said, look, I just proved how much money and time it saves. Why is it that we as a company wouldn't use a piece of software like this? And C-suite had a very simple answer for him. C-suite said, you know, it's because we're not a software company. And indeed, it's true. Boeing is a hardware company. They build hardware. They sell hardware. They're not necessarily in the software business. So even though this particular piece of software makes it that much more efficient for aerodynamics, energy efficiency, and even safety testing, it didn't really make sense for Boeing as a much larger company to be pursuing building of software. So what this individual did was that he left Boeing and started his own company called Rescale, and actually ended up hiring his boss at Boeing to work for him as the CTO. Rescale, we ended up investing in, in their earliest rounds. And over time, they've actually become the number one player in the high-performance computing space. They're now 10 times larger than the second runner-up in the ecosystem. And they sell into a ton of aviation and aerospace companies, including NASA, SpaceX, Department of Defense, where they do everything from aerodynamics design to flight testing to crash testing to safety testing. And they've expanded well beyond the aviation space as well. To work in the auto space. Their clients include Nissan, Toyota, Hyundai. And they also do work in the healthcare space. In fact, a lot of the work that they've done within the US has been particularly relevant post-pandemic because their clients include Harvard, MIT, a lot of publicly listed biotech companies. And they help all of these companies simulate the most promising drug iterations in a virtual environment first so that you can pick the most likely to succeed drug iterations to test in a physical lab. And in doing so, you're able to shorten the amount of time it takes to develop these drugs and to develop vaccines quicker. And so I share this example to say they've been able to tap into the data piece because, of course, they leverage better and more efficient data. They also tap into this compute piece, which they do high-performance computing in the cloud. They're leveraging mostly the same algorithms as what has existed in the industry. But as we've looked at these 3 categories, you really only need to pick one area to innovate in. And what's also interesting about this particular company is in terms of this whole thesis that we're talking about, in terms of leveraging AI to enable the work that we do as humans. So AI is not automating or destroying jobs, period. It is really automating the more boring pieces of our jobs. So instead of wasting all these engineers' times by designing these hardware pieces and then piecing them together in a plane, and then doing this iterative process over and over again, what you're really doing is testing in a virtual environment first so that you only have to build in the physical environment the most likely to work and the most promising iterations. And so by saving not just capital but also time, what you're actually freeing individuals to do is to spend more time being creative in their jobs, and to do other things that humans are uniquely good at. So this part of our conversation is to really go back to our overall theme that future of technology is obviously here, and it's driving very rapid changes. But at the same time, it's driving change in a way that actually empowers and enables the work that we do as humans and not just replacing us. So here I want to take a quick pause before we move on to the second section, because this section we're really talking about how AI can augment the work that we do. In terms of even symphony writing work or calling restaurants, hopefully it is all augmenting the work that we do. But at the end of the day, we as humans still need to do the more important, the more creative, the more strategic aspects of our jobs and spend our time focused more on building personal relationships in a way that is meaningful and impactful. So I have chat open in case we have other questions that come in. Oh, there's a couple that has streamed in. Let me look through it really quickly just to see which ones that we need to come in and go through. Okay, so there's a couple of questions. Let me try to answer the shorter ones. So in terms of chips, who has done the best job? There's a lot of startups that are emerging in this space. Of course, in terms of looking at large companies, NVIDIA and AMD by far is still the largest. But there's a lot of startups that are emerging in this space. And speaking of NVIDIA and AMD, I actually recently hosted through Berkeley and the nonprofit that I'm a part of, a conversation with head of data centers at AMD. I'll share that link in our chat later with everyone. It's a really fascinating conversation because you get to hear firsthand from his perspective in terms of the work that they're doing in the hardware space, not just in terms of chip design, but also in terms of larger how they think about providing hardware and data centers as a service to their clients. So I'll definitely share that. And I know Mark have shared that new technologies have been weaponized, with, of course, the hydrogen bomb being an example. And so the question is really whether or not AI will eventually lead to a world where, you know, these pop culture movies like Terminator has described. So Mark, actually, when we look at underlying data here in the US, in fact, I also hosted a conversation with head of data science at LinkedIn, She is Chinese American herself as well. And what's really interesting when you ask her about the underlying data here is that the fastest job title rising in the US right now is Director of AI Ethics and Policy. So I do think it is possible for AI to be weaponized. Absolutely. And it's a matter of what boundaries we drive. And it— we will cover that in the second section as well. Because we are hopefully getting to a world in which it's not, don't be evil, but it's, we're getting really to a world where we can't be evil. So how do we draw the boundaries and guidance and rules and infrastructure so it's not we are relying on the good faith of people to not be evil, but to really make sure we have rules in place so that people can't be evil? If you have all this intimate data, how can we make sure that it's not being weaponized and used against individuals? So a lot more nuances for us to come back to. So we'll certainly talk more about that as well. So AI learning algorithms, I know there's a comment that you would like to learn more about that. We can certainly come back in the very end discussion as well to chat more about the algorithms. But the short version is that learning algorithms haven't changed much since the 1950s when AI was first coined. So AI was first coined as a terminology in 1956. In a piece of paper that was published. And fundamentally, the kind of algorithms that AI learn by haven't evolved significantly. We have now much more and better data. We have much more and better hardware compute, but the algorithms remain the same, mainly because we as humans still learn in the same way. So algorithms refer to how AI learns, and AI learns the same way that we as humans learn, because that's how we as humans teach AI how to learn. So for example, You'll definitely have heard of machine learning. That's just an umbrella term. And if we were to look at some of the subfields like deep learning, deep learning sounds complicated, but all it is, is if we give the simplest example, is if we keep on throwing images of cats at an AI until it recognizes that those are cats, and then you keep on throwing photos of dogs at it until it recognizes that these are dogs. So we now have more complicated neural networks that are emerging. Mostly black boxes, but fundamentally to really say that AI algorithms haven't changed substantially. We have deep learning, we have reinforcement learning, which means if the AI correctly identifies images of cats, you give it a reward. If it incorrectly identifies images of cats, you give it a punishment or you take away. And so we're really incentivizing AI to learn the same way that we incentivize humans to learn. And so all of this is to say that algorithms— fascinating. We're certainly doing more work in this space. Nothing too world-changing yet. And metaverse, we will get to in this second section. So perfect timing for asking that question, Alan. We'll actually come back and talk about the metaverse in this second chunk in a sec. And then, of course, knowledge graph applications along the 3 categories certainly fall more into algorithms, although you could certainly map your knowledge of different pieces of data. But data at the core, I'm really talking about what data are you gathering, what data is naturally being generated in the process of doing business. And the best companies will say, well, in the day-to-day of doing business, I will have users and customers and clients that will naturally generate data. And when they generate data, I'll use the data to improve my products and services. And then by improving my products and services, I will end up attracting more users, more clients that then create more data. So basically, you have a positive feedback loop. And when we look at adoption of AI in healthcare being relatively slow, so late, absolutely true. It depends on what subfield of healthcare. So if we're talking about healthcare in terms of the actual care piece of being provided in hospitals, of course, a lot of it we rely on in-person interactions. But one of the companies that I've worked with is actually the largest company that is co-owned by Google and Johnson& Johnson, and it builds machines that would automate a lot of surgical processes. So there's certainly a lot of areas or subfields within healthcare that are being transformed by AI and the latest technologies. So the more human-centered pieces, for example, being a nurse, taking care of patients, really addressing to their day-to-day needs, That's something that we're going to need more of going forward into the future, really in line with our theory of AI augmenting the work that we do. But a lot more of the more tedious parts of our jobs are being automated away. For example, like radiologists, or being able to look through a lot of x-rays and identify when someone is prone or should already be diagnosed with a disease. A lot of that more manual process is being automated. So I think that's at a natural stopping point in terms of what we're looking at here. In the interest of time, I want to run through our second section really quickly about the metaverse. So in this second section, I want to quick play another very, very quick game. I'll run through this even faster than we did this the first time around, because I just want to understand how much you all know about blockchain as an industry, because this section is really about how to leverage blockchain as technology to realign incentives. across stakeholders. So again, it's 5 statements, one is false. The first statement is that JP Morgan, Walmart, and Associated Press are all leveraging the blockchain to build applications. Statement number 2, based on the latest survey, 25% of small-medium-sized businesses in the US are actually going to be accepting digital currencies by this year, by 2022. Number 3, there are national-level strategies now across the world Where countries are adopting national-level strategies to enable blockchain. Statement number 4 is based on the latest survey, half of institutional investors all are already exposed to some type of blockchain-enabled asset. Statement number 5 is you can wire or rewire a Tesla to mine for cryptocurrencies. So love to see everyone's guesses as to which one is false. So I see a lot of Things coming in. There's a rare— oh, there's a couple of 3s, some 1s, some 2s, and none from Xiao. I promise there is one that is false. There's mostly 3s and 2s. There's the occasional statement of something else. But what I'm seeing is— oh, more 4s, one 5. Great. So I asked this question really also as an excuse so we could talk very quickly about some of the most interesting changes that we've seen in this world. Oh, I see a 5 pop up on screen. Great. So I want to reveal the answers very quickly. So statement number 1 is true. So JPMorgan actually did announce their own coin years ago for institutional client settlements. If you're interested, you can dig more into that. Statement number 2 is also true. Walmart is actually quietly entering the metaverse. To Alan's question earlier about the metaverse, certainly even traditional retailers are getting in. And we certainly have the Associated Press starting their own NFT marketplace. Really creative usage actually, because they're really focusing, focusing on the photojournalism piece. So if they're already doing all this work with journalists, then they might as well make it easier for journalists and photojournalists to monetize their visual content. And so statement number 1 is true. Statement number 2 is actually also true. This is the latest Visa study, and there's a lot of other interesting details if you're interested in learning more, but most fascinating is that A quarter of small, medium-sized businesses say that by this year, they were already accepting payment by digital currencies. Same with number 3 is actually also true across all the governments around the world. India, Germany, and of course, Australia and China are among some of the first countries to come out and release nationwide strategies. And I share this because Obviously, India and China are more perceived as developing countries, whereas Germany and Australia are more perceived as developed countries. But what we're really learning here is whether it's a developing or developed country, everyone has been doing a lot in this space to better support the innovation ecosystem for blockchain internally within their own countries. And in fact, we see almost 20 central banks around the world developing their own sovereign digital currencies. And certainly here within the US, we see a lot of lawmakers now developing groundwork for a nationwide blockchain strategy. And of course, most recently, you see Biden signing— we see President Biden obviously signing an executive order in this space as well. But I want to point out that developing countries are really where a lot of the opportunities lie. El Salvador, for example, was the first country to adopt Bitcoin as legal tender. And it's really interesting to see how some of these developing countries in South America and in Africa are leveraging this opportunity to leapfrog a lot of the innovation in the US and in China, the same way that China leapfrogged the US in terms of getting faster to mobile payment. So I'm sure a lot of us have talked about from years ago how, because China didn't have the existing infrastructure of credit cards like the US did, we ended up having the ability to leapfrog this entire system and jump into mobile payments much faster than the US did. And now we're reaching a point where South American countries, and African countries are seeing the same opportunity where they say, hey, we don't even have the same kind of mobile payment infrastructure. Maybe this is where we can leapfrog ahead and look at blockchain-enabled opportunities. So now we certainly see a lot of expansion launching in South America and just Latin American countries in general. It's not something that we talk a lot about in the US. So I just wanted to flag this as an opportunity for us. So if the first 3 statements are all true, we get to statement number 4, 4. Statement number 4 is the one that's actually false. So it's not half of institutions. Now it's actually close to about 3/4. We're actually almost at 75% institutional investors with exposure to blockchain-enabled assets. Without going too much into detail, these are among many of the companies that have had exposure to blockchain. MassMutual, fascinating as an example because it's so old of an institution. If you have a 170-year-old company, able to clear regulatory procedures to make an investment, then it basically tells us this is certainly feasible. We have companies like Square seeing blockchain-enabled digital assets as an alternative to corporate cash, so getting a lot of exposure to the space. We have public companies like MicroStrategy, most famously, raising money, additional money, just so it could invest it all into the Bitcoin. And of course, even the most traditional funds, like pension funds, are getting in on Bitcoin. So even if you don't personally do any of your investments right now, it could also be that whatever money you've saved in pension, you're already getting exposure. And of course, again, to come back to this theme of emerging countries really active in this space, Rio in Brazil became one of the first countries or first cities to announce that it wants to put 1% of the city's treasury into crypto. So seeing this as an alternative to cash. And so if the fourth statement is false, that means that the fifth statement is obviously true. Before we go into the fifth statement, I just want to quickly share that this is now available for retail, not just for institutions as well. So for all of us here today, if you're an individual and you're interested in getting exposure, whether it's through JP Morgan, Morgan Stanley, Goldman Sachs, or BlackRock, there's a lot of retail consumer-facing products available now. where you don't necessarily have to buy your own currencies or figure out how these assets work. You can absolutely go to some of these larger platforms and understand what they have available. And so, as promised, statement number 5 is true. And this is actually how you would rewire a Tesla to mine for cryptocurrencies. And if you had one of the oldest versions of Teslas, you had free supercharging, which means that you can basically mine for cryptocurrencies with free electricity. But don't tell anyone I said that. And of course, now you have larger companies that are investing up to the billion level in new hardware compute, which relates back to what we were talking about even in the AI section, where a lot more data, certainly a lot more hardware and compute. All of this to better enable us as humans to do our work. And so in all these examples I've shared just now, it's really falling into this framework of what many of you might have seen before, as the Gartner hype cycle. I shared this game to start this whole section because I think we have a very bipolar approach to everything happening in the blockchain space. It's very similar to AI, where either people are saying blockchain is going to change the world, I'm all in on blockchain, or it's like, wait, actually, there's too much hype. I don't believe any of this is going to be true. It tends to be one or the other. And again, I want to take more of a balanced approach, which is to say there are certainly parts where blockchain is changing the world in terms of how we do things, but there are But there's also other areas where, do we really need to use blockchain, or can we use existing databases? And so the balanced view is to really look at any technology, whether it's blockchain or AI, in this framework, where on the y-axis, we're looking at expectations, and on the x-axis, we're looking at time. So with any new piece of technology, it triggers a lot of innovation, where people's expectations very, very quickly climb to this peak of inflated expectations. And then when people realize that it's not capable of doing everything is promised to in social media or in the news, expectations very quickly drop. But then we reach this trough of disillusionment when people are like, well, actually, the technology is not as disappointing as it is. And the real building starts as we start to climb up the slope of enlightenment. AI has certainly had this exact trend where, uh, you know, just a couple of years ago, we have crazy valuations from AI companies. Where there were too high of inflated expectations. Blockchain now, you could argue, is still at a point of inflated expectations. But there are certainly subfields within AI, as well as within blockchain, where real building happens. And those are the kinds of companies that I personally love working with and investing in. And in fact, I love investing in companies at this trough of disillusionment, because it means valuations are very fair and reasonable. And it certainly means that real building is happening. where people are not too worried about the financial incentives. So I think this is really a question for all of you to answer, which is, where do you think we are and why? What are the rational reasons behind why you think we are where we are? And there's a different set of actions that we could take depending on where we are. If we're just at a point where we're climbing to peak of inflated expectations, maybe this is not the best time to get in on something. But if you feel like we're now starting to climb up this slope of enlightenment, It could be a fascinating time to think about it from an investing perspective or from a career perspective, because this is where the real building happens. And this is also where I personally love to look at movement of talent. So I've been teaching at Berkeley for the last 10 years. I typically get about 60% grad students and 40% undergraduate students. The grad students I get range from PhDs to students who are cross-enrolled between UCSF and Berkeley. to MBAs, to Masters of Design, and to even very, very technical PhD students in biotech or in physics. But in terms of the undergraduate students I get, especially now, all of my undergraduate students are born post-2000. So this is a very different generation. And when you look at them, I do think where they move or where they work is a high indicator of where we are in this hype cycle we are. So in the past, even though there's a lot of hype, you actually don't see the smartest and best talent going to work for these companies. But now I think for the very first time, we actually see the best talent, the smartest people making job changes in order to work at some of these promising emerging companies. So I think that's certainly something really interesting to look at. And of course, in the interest of time, I'll run through the rest of this section more quickly because now we have a good sense of what's happening in this space. So of course, blockchain is the underlying technology that enables Bitcoin and other cryptocurrencies, and they're not the same thing. And at the end of the day, blockchain is all about a distributed list of records. So it's a database, but distributed, which means that in the past, you have a centralized third party that sits in the middle, like the government, like the bank, or like a company that sits in the middle and manages all of the transactions. But in a blockchain-enabled ecosystem, everyone in real time has a copy of all the transactions. And what it really enables us to do is we can replace the work that a government would do. By having a completely decentralized currency like Bitcoin. We can replace what banks do by having completely decentralized financial systems, such as DeFi, or decentralized finance. We can completely replace the work that traditional companies do because we can have, again, a decentralized ecosystem around decentralized applications, or dApps. And this is where things become interesting because then it enables for this whole world of DAOs, or decentralized autonomous organizations. We're in the very, very early days of DAOs. I'm currently doing research at Berkeley now to really look at what are the governance implications, what are the legal implications, and what are the financial valuation implications of having DAOs as a company structure instead of our traditional companies. So happy to chat more with everyone if you're interested in this space. And as promised, I just want to recap this section to really just say if we're replacing the work that governments banks, and companies are doing, we're really changing the way we think about digital money. We're changing the way we think about digital and programmable smart money. And we're changing the way that we think about anything of value, not limited to money. Because now we can have a world where we can have photojournalism. We can have photos and visual imagery have value, unique value that's attached in a digital world. So this is really where the metaverse comes in, as promised I would come to. Because in a world that is enabled by a virtual environment, you can have virtual real estate. You can actually own a shop in a physical— not in a physical environment, but in a virtual environment. You can have virtual products. You can own certain fashion pieces. You can own sneakers. So in the metaverse, this becomes interesting because if we believe that we can live in a world where virtual items have unique value, or data has unique value, it means it's not just about money, but it's really anything of value that we want to ascribe financial and monetary value too. And as a result, Gartner predicts that by 2030, which is really just 8 years from now, or less than 8 years now, 10% of the global GDP will be stored on the blockchain. And we certainly see this very much exponential trend in terms of the startups that are raising funding in this field as well. Everything from underlying infrastructure and protocol all the way to end applications. But of course, we're in the very, very early days. And this is where I want to say, In comparison or contrast to AI, AI was termed back in 1956, but blockchain really only came out as underlying technology in the late 2000s. So in comparison, blockchain is very new as a piece of technology, and you could really think about it as if we're in the early days of the internet where a lot of infrastructure is still missing. So now if you want to build a website, you can drag and drop a few buttons and have a website up and running, but that's not yet possible in the world of blockchain. It's just like in the early days of the internet, if you wanted to build a website, you actually needed to know how to write HTML code. And now we've simplified that to make it UI and individual-friendly. So we're certainly down that trend. And the metaverse is exploding as an industry, as we know, but where it really— the only number you need to know from last year that is important to highlight as kind of a top number is that the world of NFTs, or non-fungible tokens, certainly a part of the whole virtual environment that we live in. It actually ended up surpassing $40 billion in 2021. So this is really not a matter of, is the future here? This is a matter of, the future is here, and what do we do with it? And now we see a lot of startups emerging in this space. I'll just share this one example of a startup that we've been looking at called Arcade, which is actually thinking about the world of metaverse and digital items, as a financial asset, because what it allows you to do is instead of owning an NFT, you can actually put your NFTs to work by borrowing, lending, and earning on your crypto asset. So in the past, say if you were to buy a painting, a physical painting, there's really not much you can do with the painting except store it away for safety or hang it on the wall for enjoyment. But now what you can do is you can take a digital art piece and actually gain financial value from that. Of course, in the past, you could say, well, if you really have a very famous piece of painting that's worth a lot of money, maybe you could put that as collateral and take out a loan on that. But now in a digital environment, it makes it so much easier to have a digital asset that is worth something, and you can borrow, lend, and earn on top of that. And if you're interested in learning more about how this specific startup works, they have a developer hub where you can read more about how they put all of this together. But I share this example Because it's a fascinating example of how the metaverse is not just creating a virtual world, but it's creating a virtual world in which a lot of our items of value in a digital environment is really a new financial class of assets. So there's all these different things that we can do with it now that's completely different than before. So that's a very quick rundown of how blockchain actually helps realign incentives, because now you can have decentralized of everything. So instead of us having to choose one centralized third party, we can now have individuals that give input in terms of how these communities are put together. And in line with how we opened our entire session, we have jobs that are now— people are now in C-suite jobs that didn't exist at all when they graduated. And one of the quickest emerging C-suite job titles is called Chief Community Officer, where their job is to run a community and not necessarily do any other part of C-suite work. But to just run and oversee and engage the community that makes this company possible. And blockchain is certainly one of the first technologies that enables us to really focus on aligning incentives across all stakeholders and building a meaningful community. And so here, I'll take some quick questions. I know we're already over time, so I want to make sure to take these questions and leave for any discussion. So when would non-digital currencies phase out? That's a really fascinating question. It depends on whether or not we think non-digital currencies will ever phase away. Because, for example, in China, very few people use credit cards. But credit cards are still certainly the way to go in the US. Mobile payment has become extremely easy to do in China. But we haven't quite yet gotten to a place where we're phasing out all non-digital currency. So it's a question mark as to which geographic region and how quickly the adoption is. And I think Charles has a question on, will the government or authority allow the financial control out of their hands through decentralization? So that's a really good question. There's a couple of different ways we can approach this. Number one, we certainly see a lot of countries taking advantage of this opportunity to attract more talent. So for example, you have countries like Singapore, like Switzerland, like Cayman Islands saying, look, we will have easier regulations. So if you want to build something in this space and you don't want to be in the US, you can come to us. And so there's certainly a drain of talent. I'm sure for any of your friends who are in the blockchain crypto space, you might have seen a lot of them move out to countries even like Thailand, because why pay all the taxes and deal with the complex regulation? Why not move to somewhere different? Especially now post-COVID, we have remote work that makes it possible to do work anywhere in the world. I'm sure we've all had friends who have moved out to Texas or to Hawaii or even to Portugal and these other places around the world. So we've certainly seen this as a strategic opportunity for different countries to attract more talent to move there. People don't want to stay in the US as much as people used to. And the second piece of this is, for countries that do not allow for decentralization, that's fine. But the eventual goal of where we're moving towards is where we have regulations in place where you cannot do evil, but the rest can still be a very community-driven approach. And that's why we have a lot of chief community officers. So as long as we're not taking advantage of retail and individual investors, it's still possible for a community-driven approach to take place and have it be a very decentralized process, but just as long as it follows general guidelines. And I think Will has a question around what the minimum value of one single NFT or asset will be that make use of blockchain reasonable. Because currently, the cost of transaction is so high? It depends on which network we're talking about. There are certain networks that are being developed. For example, Ethereum has been working on a couple of projects where hopefully we can do settlement for much cheaper. But if we were to draw the analogy between the world of decentralized finance and traditional finance and just look at the comparison and contrast, it's actually not as cheap as we think it is for traditional financial services institution to process a transaction. Because if you think about all the costs that go into cybersecurity to make sure that they have robust systems, when you think about all the hardware compute and data centers that are required, when you think about all the individuals that they employ to make sure they have cybersecurity in place, and when you start to really account for all of these costs— You can actually check out a really good quantifiable analysis of this. Cathie Woods from ARK has published a report on just comparing and contrast, what do traditional financial institutions and banks do? What do DeFi companies do? It's actually significantly cheaper to run a transaction in a DeFi approach, at least within a financial services setting. Now, for other types of settlements, it might be different. But it certainly is easier to do and cheaper to do micropayments in a decentralized fashion than a traditional fashion. The easiest example I'll share is, for example, right now, let's just say you want to bring $10,000 US, from China to the US, what's the cheapest way to do this transaction? You know, there's all sorts of things that we can imagine, like doing a transfer via the bank or a wire. But really, the fastest way to do that is if you were to literally fly with $10,000 cash from China to the US, because that's within limit, right? If you use any bank, it's 5 to 7 business days for transactions, and for settlement, and it usually involves a couple percentage points of transaction fees. But with a decentralized approach, you can directly transact with another person. So we currently have Wise. We have a number of these blockchain-enabled approaches that allow you to settle transactions on the same business day, and sometimes at very low, if even sometimes no transaction fees. So that's certainly something to look at now. And I know Scott is sharing a comment on whether AI is really played only by the rich, the big guys, the big companies. But blockchain is more grassroots and people who come from more humble backgrounds or small, medium-sized businesses. So Scott, that's a really interesting perspective that you bring up. So there's certainly this very grassroots movement within the blockchain ecosystem. But if you look at who the largest investors in the blockchain ecosystem are, they're the same as any other industry. The biggest investors, the whale investors, they're still the largest financial institutions. So if you look at any token, for example, you'll see the largest holders are venture funds, are traditional financial services companies, venture funds that are backed by traditional financial services companies. It's still the same large investors. But where it's different is in terms of how the community is built. So if you were, say, at a bank, do you have to listen to all your customers? Technically, no. But if you were building a blockchain community, Is it now culturally very important to build a community and to engage on platforms like Reddit and on a lot of chat platforms, on Telegram from day one? Absolutely. So I think the biggest difference is not who has the most financial exposure, because the big guys, the wealthy companies, they will always have the most money to invest in any sector. That's really in terms of bottoms-up incentive. So I certainly see a lot of grassroots use of AI as well. where people are now starting to think about, well, what can we do with AI? But AI does need more scale. You do need more data. You do need more hardware compute. So it is harder to play the AI game if you don't have a lot of data. So I would say the barrier to entry for an AI startup is significantly higher than a blockchain startup, because in a blockchain startup, you really just need a very innovative way of approaching things or a unique community that's built, and that works. And To Brian's point on the supply of Bitcoin being limited, once the mining of Bitcoin is finished, what will happen to miners and why? So you're absolutely right. Once we reach the cap, we really won't be mining for new Bitcoin. But what the miners will then instead do is they'll become processors of transactions. So you'll still have transactions that occur. So instead of processing transactions to be rewarded with a new Bitcoin, which is called mining, You'll be processing transactions to be rewarded a transaction fee. And that's been designed into the ecosystem from day one. And of course, Brian, you're absolutely right. Now, Bitcoin and Ethereum still use proof of work, which use a lot of energy. But Ethereum has actually been transitioning to proof of stake. Absolutely. Hasn't happened yet. What do I think of this? So again, other than Bitcoin and Ethereum— of course, they're the 2 largest ones— And we see a transition starting to happen. We're slow. Starting to happen. But I think where it's most interesting is if we look at some of the more emerging blockchain startups and what they're doing. Because oftentimes, it does use a lot of energy. But is the energy worth all the other risks that come with? For example, in a traditional bank, you may not be using as much energy because you're using a traditional data system. But you're also Easier to compromise as a system, because you really only need one person in Wells Fargo to make a mistake and not do his job properly, or to have done his job but accidentally made some sort of mistake for the whole ecosystem to be compromised. So it's hard to say if there's a fundamental better or worse. In fact, many of you might have heard of the movie played by Leonardo DiCaprio called Catch Me If You Can. And Catch Me If You Can as a movie is based on a real story of an individual called Frank Abagnale. And Frank Abagnale has been working at the FBI for the last couple of decades. And it is a really, really interesting movie. In fact, Frank's son actually works for a Chinese gaming company. His son also speaks Chinese. And he's a really, really fascinating person. And I bring up Frank Abagnale because if you talk to Frank, what he will tell you is that For any traditional cybersecurity company, their ecosystems are very easy to compromise. You just need one person within an organization of 10,000 people to not do their job properly, and then you can have a hacker that gets into the system. So the cost of that hack is actually extremely high. And when you look at some of the largest hacks that have happened in the last decade, you have Home Depot, you have US home the Department of Homeland Security being hacked. You have some of the largest banks being hacked. You have some of the largest security companies, even RSA Security, being hacked. So no one is really safe or prone from hacking. So everyone is prone to hacking. And so as a result, perhaps currently the proof of work, even though it does use a lot of energy, when you compare it to how much energy any traditional company uses to secure its cybersecurity systems, it's actually quite reasonable. And Eugene, great question. What do we think of the wealthy individuals in Silicon Valley that love to invest in life sciences, all about living forever? So one of my fellow faculty members at Singularity University is what is known as a longevity expert or the aging expert. And he starts his talk by saying this one sentence. He said, aging is a disease. So it is like any other disease where it's curable, where if we adopt the right practices, it is possible to avoid or to overcome or to treat. So that is a very Silicon Valley perspective. Now, whether or not you fully agree with him, we can look at the details of what he talks about. But there are certainly people who love to invest in life sciences and live forever. Now, what do I think? A couple of things. Number one, I think if you are super wealthy and you already have everything in life, then I guess potentially living forever may be the only thing that you have left that you really want to do. So I understand why there's a lot of very wealthy individuals that want to invest in that. Others though won't invest in living forever, and rather they live— they invest in fundamental life sciences, like how do you treat cancer. And I've actually invested in a portfolio company called Notable Labs. That is completely focused on hematological research, which is leveraging data to figure out how can we personalize cancer treatment. So right now we cannot cure cancer, but there may be 100 different ways to personalize that cancer treatment for the individual so that you're more responsive to the treatment. Currently, if you go into the hospital for cancer treatment, they have the same kind of treatment that's prescribed across different patients. But how one individual responds to treatment is not the same as another individual. And I think a lot of the times, I'm certainly not wealthy. I'm based in Silicon Valley. I don't invest in hardware or hard tech life sciences, but I certainly do love investing in data-driven life sciences approaches. And the reason for that is that it's something near and dear to our hearts, I think. My grandmother was diagnosed with blood cancer, and she unfortunately recently passed away. And I grew up with my grandparents, and I'm very, very, very close to them. And I was the primary caretaker for both of them for the last 2 years. And so when you have a personal experience like that, it's very hard to not invest in something that makes a difference from a life science perspective. So I do think that life sciences tends to be one of the things that resonate with all of us. But of course, I think everyone needs to pick and choose their own battles and think about What are you good at investing in that you understand and are able to do that? And Long's point is, miners will help record transactions and be rewarded even though BTC reaches 21 million. Yes, they will record transactions. And they are rewarded by transaction fees. So once we reach the cap, whenever you want to have a transaction that is settled, you have to attach a reward to it. And the reward is essentially the same as what we think of as transaction fees in traditional banking. setting. So instead of being rewarded with a new Bitcoin, you're rewarded with the fee that is attached to settling that particular transaction. And Solana has definitely emerged, to Bing's point, among the different networks as a leader. Absolutely. It's been doing a lot. And what's really interesting is when you look at the world of these new emerging platforms, no one even knew what Solana was until much, much later on. So I do think that it's helpful to be able to track these startups early on and understand the fundamentals of what they do. And actually, going back to Eugene's point, how about people living forever in the metaverse? So it depends on what we think of as a metaverse. So I'll make a controversial statement that not everyone will agree with, but I do think everything in Silicon Valley is cyclical, which means that at the end of the day, nothing is completely new, 100% from scratch. In fact, metaverse is really just combining what we used to call VR/AR with some of the blockchain technology-enabled digital assets. So what is considered living forever? Is it our physical bodies living forever, or is it the ability for us to leave behind a legacy, or is it us being able to leave behind the ways of thinking or the— Because at the end of the day, we don't currently yet now live in the metaverse. The metaverse is just a small part of our identity. But many of us still have lives where we physically interact with others. But whether or not we will increasingly exist in the metaverse, certainly possible. Most of my students now, undergrads, are increasingly more tied to their digital identities in the same way that they are with their physical identities. So it certainly is increasing as a trend. I know we are now well over time, but hopefully we've had the opportunity to answer questions throughout our session together. And really, the last thing I was going to leave you with is we've had the opportunity to talk about AI and blockchain, but it's really the intersection of all of these different exponential technologies that make it interesting. When we start to think about quantum computing, healthcare, life sciences, cloud computing, and all of this stuff. And at the end of the day, it's really about thinking about people and knowledge. And I think obviously this alumni network is amazing. for you to learn from each other, for me to learn from you. And I think it's really thinking about how do I expose myself to a regular network where I'm exposing myself to the relevant people and knowledge that's important to know out there. And the one thing that I personally do is I've personally been involved on the board of a nonprofit called Decode. It's the largest technology and entrepreneurship and innovation ecosystem that is co-hosted between Berkeley and Stanford. Obviously, I'm on the faculty at Berkeley, so it's no surprise that I'm plugged into this network. of you that are here today, hopefully you'll have an opportunity to plug into your own local ecosystems as well, wherever that might be in the US or even outside of the US. And through Decode, I've had the opportunity to host a lot of really interesting conversations and just learn from people. In fact, one of the latest conversations I hosted is with Michael Li. He's a longtime friend who was previously head of data science at LinkedIn and has since been VP of data at Coinbase. And he talked about the intersection of AI and blockchain. So for anyone who's interested, you would certainly check this out on YouTube. And I know this talk itself will be on YouTube as well. So if there were any pieces that you wanted to follow up on, please feel free to watch the recording and reach out to me afterwards. I'm about to host another conversation in the DeFi space with Aave. It's one of the largest— I think it's definitely top 3 or top 4 in the DeFi space. They have, I think now, well over $20 billion locked up in the protocol. And the founder is coming in to do a quick chat with us through Berkeley. So I think this is going to be a fascinating conversation. But really, I just wanted to leave that with you to say, if you want to be more plugged into Silicon Valley, Decode is a helpful community to plug into if you want to learn more about what's happening between Berkeley and Stanford. But outside of the 2 schools, there's a lot of other amazing networks to tap into, whether it's professional networks within your own company, your alumni network right here. Xiao and Yudi are the most amazing connectors and they know everyone, so I'm sure if there's anyone that you want to connect with and get to know, they are the right people to chat with. And with that, thank you so much for having me here. You can find me on LinkedIn at Shuo Chen. My WeChat is there as well, so if you wanted to add me on WeChat, I'm certainly happy to do that as well. I think if you go on to my LinkedIn and you look under contact, that should have my WeChat in there. I know a lot of this I had done— oh, actually, Alan's question was perfect. Like, if WeChat and LinkedIn both work, so whichever one that you're more comfortable with. And a huge thank you to everyone who's on the East Coast, because I know that's certainly running quite late. So if you have any other questions that I can be helpful with, please feel free to reach out anytime. And maybe the last thing that I will mention here is What has been helpful for me is to hear and learn from all of you in the chat to see what you're interested in. So if there's anything else that you wanted to follow up on that I didn't have a chance to answer, whether it's me or Xiao or Yuri, please feel free to to do that. And of course, because a lot of the stuff we've talked about is more technical and I'm more used to teaching in English, I've stuck to English.但是中文其实也可以，所以大家如果有什么问题的话，也可以随时用中文联系我。
Thank you so much for a fantastic talk. So, Xiao, if you want to say something, you go first.
Yeah, that was an amazing talk from my perspective. And when we— last time I heard Shuo talk about this, and we actually sitting in the audience, all the partners like looked at each other and said, oh my God, this go like nonstop. Was talking like an hour and a half. It's like nonstop, the information just exploding, and everybody just learning so much and thinking about what else I should have followed through and everything. I don't know what other people think, and I think we definitely— I think a lot of our audience wanted to invite you back for other topics. But Anyway, I'm sure that you're gonna— you should have a lot of followers already by now. Okay.
Thank you so much. It's too kind of you, and I'm sorry for throwing all this information at you. I wanted to make sure that at least hopefully we paused and took enough comments and questions throughout to be helpful. But if I've just been talking about things that you're not as interested in, feel free to ask about things that you're interested in separately.
We're so well nurtured by your origin stories, Shuo. Yeah, so do you have 5 more minutes? I have 2 questions that I pre-collected before the event, one in English and one in Mandarin.
Sounds great.
Okay, great. So, uh, what has— okay, first one, what has helped you get to where you are leading and influential now, and what advice would you have for others who want to set off in a similar direction?
Yeah, honestly, this is probably one of the geekiest things that I would say, but I've been following this traditionally computer science logic of exploring and then exploiting. So I think it's very difficult for— I mean, and I tell the same thing to all my students all the time as well— it's very difficult to figure out what you want to do for the rest of your lives based off of a few interactions. So I think something that I've done in my life is explore a lot. If I don't know something, I want to at least dedicate a small portion of my time to exploring how that field might work. And then from all the exploration I've done, really choosing to focus on a few specific areas. I'm still not very good at the focus piece yet. But what I try to do is, if I think about all of my time, most of my time, 80% plus, is being really focused on the core thing that I do, which obviously now is running the venture But 20% of my time I allow myself to explore because that's how I think I can become a better investor and to be better at everything else it is that I do. And within the 20% of time, that includes teaching. I teach a class every semester at Berkeley. That includes the board work that I do, which is working with fascinating companies and looking at what they're working on. And it includes even things like personal hobbies and things that I explore. So I kind of apply this balance where I always give myself some time to try something new, but it may or may not be the best thing that works for you. In fact, my significant other is very, very good at staying focused. All he does is 100% focused on what he does. And so everyone has their own styles, and you'll have to find the best one that works for you.
Yeah, great, great. So since we've known each other for quite a while, every time I check on you, You were always on the road or giving a lecture or presentation, or you know doing your VC or co-organizing some social impact events. Wow, that's so amazing!
Thank you.
And the second question, so it's in Mandarin. 我们对你的自身成长经历如同你的演讲内容一样感兴趣，请问你是如何？ 做 到 同 时 在 做 大家 一 件 都 忙 不 过 来 的 事情 而且 还 每 一 件 都 做 得 非常 好 那 在 这 过 程 当 中 你是 怎么样 突 破 自己 的 每 一次 瓶 颈
就是 真的 不 敢 当 因为 就是 我是 属 于 非常 非常的 l ucky 就是 有 这样 的 机 会 因为 其实 我 现在 做 的 很多 事情 是 属 于 ex ception 而 不是 norm As in 其实 正 常 的 f ac ulty member 是 需要 全 职 去 讲 课 的 呃， 而 不 一定 有 这样 像 我 可以 全 职 做 自己 的 基 金， 然后 再 兼 职 去 讲 课 这样 的 机 会。 呃， 我 我 比 较 lucky 是 因为 我 教 的 课 程 内 容 是 创 业 相 关 的， 创 业 是 属 于 非常 需要、 非常 非常 需要 industry practice 和 这个 real time experience 的 一个 topic， 所以 Berkeley 才 有 这样 的 structure 可以 允 许 我在 faculty 上， 但是 一个 学 期 只 带 一 门 课。 那 其 他 Berkeley 的 教 授 肯 定 是 需要 一个 学 期 可能 带。 四 五 六 门 课， 然后 呢， 同 时 还 需要 做 full time research， 所以 这 一点 来 说， 我觉得 还是 比 较 比 较 lucky 的。 然后 嗯， 其实 我觉得 如果 说 自 身 经 历 的话， 跟 上 一个 问题 回 答 还 蛮 像 的， 因为 我 其实 是 很 早 做 了 很多 的 这个 experimentation， 有 很多 我 自己 比 较 喜欢 的， 呃， 不管 是 跟 工作 相 关 还是 不 跟 工作 相 关 的， 比如 说 呃。 我 特 别 喜欢 种 菜， 种 菜 已经 种 了 十 几 年的 时间 了。 这个 也是 属 于 我 自己 花 很多 时间 做 的 一 件事情。 然后 余 力 有 时间， 咱 们 loc al 离 去 的时候 给 你 带 一些 这个 自 家 的 这个 各 种 各 样 的 ve get ables 和 f ru its。 呃， 所以 其实 我觉得 我是 属 于 精 力 比 较 充 沛， 然后 喜欢 就是 尝 试 各 种 各 样 不 一 样 的 东 西 的 一个 人。 所以 呃， 我觉得 这 一点 来 说 是 比 较 l ucky 的 吧。 但是 我 先生 就是 属 于 非常 非常 会 f ocus。 他 就是 不能 同 时 做 多 件事情， 然后 一定要 百 分 之 百 的 时间 f ocus 在 一 件事情 上。 所以 从 自 身 经 历 来 说， 我是 属 于 这个 person ality， 不管 是 这个 性 格 来 说， 还是 精 力 上 来 说， 一直 属 于 比 较 充 沛， 然后 喜欢 各 种 各 样 尝 试 的 一个 人 吧。
谢谢 说， 我 一直 说 您 是 我们 华 人 之 光， 而且
是 god of giving
presentations， 而且 特 别 孝 顺。 那个 我知道 你 一直 在 照 顾。 grandparents. Yeah, so, well, I think it's the end of our webinar today. So thank you again for answering all those questions and for the great presentation. It was a pleasure to have you with us. So this will conclude the webinar. So thank you all for attending. We hope you have learned and enjoyed this presentation. 呃， 那个 小 青 学 姐
还有 什么 要 讲？ 非常 感 谢 啊！ 非常 感 谢 啊！ 这 就是 呃 我们 呃 总 会 的 系 列 讲 座 之 五。 然后 呃 非常 希望 你们 能 够 关 注 啊 这个 呃 交 流 平 台， 然后 呢 关 注 我们 总 会 的 关 注 峰 会 的 交 流 平 台 啊。 您 希望 什么 样 的 啊 话 题 也 随 时 请 啊 在 微 信 群 里 面 交 流， 我们 会 尽 量 的 满 足， 尽 量 的 找 到。 更 优 秀 的 啊， spe aker 来 跟大家 分享 好 吗？ 谢谢 大家 啊！ 大家 晚 安。 我也 帮 我也 帮 硕 打
个 小 广 告， 硕 有 一个 You Tube channel 叫 做 D ec ode， 大家 在 那个 You Tube 上 搜， 那 上面 有 非常 多 他 去 inter view
其 他 也 非常
厉 害 的 spe aker 的 video。
Yeah. Oh yeah, I 'm sure there 's so many followers right now. Okay， 这个 粉 丝 群 已经 建 立 起 来了 啊！ 谢谢 大家， 谢谢
大家， 大家 周 末 愉 快。 周 末 愉 快！ 最 后 我 再 感 谢 一下 我们 交 大 校 友 会 的 这个 团 队， 小 青 学 姐， 还有 呃 那个 谢 东 学 姐， 还有 Frank 师 兄 啊， 他是 我们 西 安 交 通 大 学 DC 的 会 长， 同 时 还有 他的 这个 霍 普 金 斯 创 新 发 展 中 心 也 supp ort 了 今天的 e vent。 非常 感 谢 大家。 那个 东 部 的 晚 安 呀， 西 部 的 en joy dinner and sure enjoy your lunch。
Thank
you so much. See you soon.
See you triple back. I'll see you triple back. Bye bye.
See you guys soon.
Bye
bye. Take care. Bye bye.
那国内的早上好呀。 Bye bye. Bye bye.
More from Shuo
Ep 666: Shuo Chen — Fractional Founders, Frontier Firms and Why Enterprise Data Is AI's Gold
75. Shuo Chen - Boards and the future: going plural, encouraging innovation, adding value to entr...
Exploring the Future of Work with Shuo Chen: IOVC Partner & UC Berkeley Faculty | Podcast Episode
IOVC General Partner Shuo Chen on Investing and Richard Branson's Island | Seed to Harvest S1 Ep1
New Instructor Spotlight: 5 Questions with Shuo Chen on the Future of Education and Technology
Validating Ideas with Shuo Chen - UC Berkeley's Silicon Valley Startup Success