Hi All,
Welcome to the first episode of Differentiated Understanding. Today we speak to Jasmine Sun and we dive into everything from her recent trip to China, covering China as an ABC, how the capital in China vs. Silicon Valley shapes the markets, and talent flow.
In today’s world, there’s no shortage of information. Knowledge is abundant, perspectives are everywhere. But true insight doesn’t come from access alone—it comes from differentiated understanding. It’s the ability to piece together scattered signals, cut through the noise and clutter, and form a clear, original perspective on a situation, a trend, a business, or a person. That’s what makes understanding powerful.
Every episode, I bring in a guest with a unique point of view on a critical matter, phenomenon, or business trend—someone who can help us see things differently.
For more information on the podcast series, see here.
See transcript below.
[Subtitles may contain minor typos of names as its AI-generated]
00:00 Introduction to Jasmine Sun
06:01 Journey into Writing and Substack
11:24 Reflections on China and Personal Connections
22:13 Exploring AI Anthropology
36:04 Comparing Tech Cultures: US vs China
52:36 The Shift in Silicon Valley Culture
01:01:26 Addressing AI Psychosis and Regulation
01:10:59 Misconceptions about AI and Technology
01:13:31 Tips for Aspiring Substack Writers
01:17:00 Unconventional Beliefs about AI
Grace Shao (00:00)
Hello, I'm your host, Grace Shao, and joining me today is Jasmine Sun. Jasmine is an independent writer and podcaster covering AI and Silicon Valley. She writes at jasmin.news and has also been published in the Wall Street Journal, Business Insider and the San Francisco Standard. Previously, she led the core product team at Substack and worked on AI policy for Schmidt Futures and Mozilla. It was so lovely seeing you in Hong Kong few weeks ago, Jasmine, especially right before you went on your big trip up in mainland China.
So I can't really wait to get into that trip with you and your latest reflection piece titled America Against China Against America. But before we get into that, I would love to learn more about you. I feel like there are so many people who follow your work now. You cover everything from AI to, you know, ⁓ tech culture to now dabbling a little bit of China as well. So why don't we start with that, Jasmine? What led you to write about AI and tech on Substack? Who are you and what do you do?
Jasmine Sun (00:49)
Thank you for the introduction. Yeah, it was so fun to get to see you in Hong Kong. I think it's like always special when you know someone's work over the internet and you admire it and you actually get to meet in real life all across the ocean. So that was great. Yeah, so I am currently an independent writer. Like you said, I mostly cover technology from relatively more of a cultural perspective. And these days that's a lot of AI stuff, like you mentioned. I actually, started writing full-time only relatively recently. I...
I used to work in Silicon Valley full-time. I was a product manager at Substack for four years. And then only at the end of last year did I decide to leave my job once I had been there for four years and take the leap into full-time writing. So that's been really fun. And I've enjoyed it a lot in part because I just remember the feeling when I still had my day job of like, I loved my job, but I could sense that so much was changing. Like a lot of that in AI, a lot of that in culture and technology in general.
I just couldn't keep up part time. I think I need all of my hours in a week to actually stay abreast of what's going on. Yeah, so I sort of took the leap and now I'm mostly just sub-soc and I freelance and I'm trying to make that work out too.
Grace Shao (01:57)
Awesome. Tell us about how you got involved with Substack actually, because you were one of the earliest employees and I think it's really just blown up in the last two years, especially in more mainstream area. So yeah, tell us about that.
Jasmine Sun (02:08)
Yeah, so it's actually been really fun to see. was reflecting on this with a friend last night because I think I started my Substack personal, which was a tiny, tiny blog. I was just literally blogging for myself and 10 of my friends. And I started that blog in 2018 on Substack, so quite early and probably before a lot of folks had heard of it. I think I was following Azeem Azhar exponential view and that might actually be...
He was, think, the first sub-sac I ever heard of and kind of what inspired me to start my own. was literally in undergrad at the time and I just wanted somewhere to host a blog because I had always loved blogging, loved writing. I had Medium before that and WordPress before that and it was just always something I did for fun. So I'd been using the product and then I graduated from undergrad at Stanford and I worked in tech policy as my first job. So I actually was at Schmidt Futures and their internal think tank, actually a lot focused on US and China and AI stuff that was in 2020.
sort of the end of the Biden admin or sorry, the end of the first Trump admin going into the Biden admin was when ⁓ we were, I was working on some tech policy stuff there, but I had always loved journalism and always loved writing. so there was this thing that was happening where I was like, on one hand, I would love to be a tech journalist because I think the thing that I am most interested in with technology is actually like sense making and trying to understand it and communicate that to a wider audience and thinking about
broader perspective rather than necessarily just like one product in one company. And at the same time, it was a very bad time to be a journalist in 2020. Like it was COVID, so there are tons of layoffs and the industry was frozen. Everyone knew that business models were dead, like pivot to video didn't work out. No one really knew what the next big model was going to be. And so it was just kind of a bleak time to try to be like a junior staff writer anywhere. And as I was just studying the landscape, I guess, of media,
I was like, well, I think Substack might be doing something interesting because at that time they had like Bill Bishop, they had Azeem, they had like a few people who were like starting to, I guess, like really build a name and a business for themselves that seemed quite successful. And it was a real business model. People were being really creative with it. And also it sort of combined my interests and like, I want to work at a startup and I want to like get experience actually working on product because I think that will make me a smarter tech analyst if I also have experience on the product side of things.
and I really care about the future of the media. And so I joined ⁓ at the beginning of 2021 and worked there until the end of 2024. So about four years, there was like 25-ish people when I joined. It was about a hundred and someone I loved, but like the user base had like exploded during that period of time. Like now my friends know what subpsych is when I tell them, even strangers know what subpsych is usually when I tell them, which is like really cool to feel. And there's notes and there's video and there's podcasting. And you know, like when I joined, it was literally just like...
you can send emails with text in them. And so it's very cool that like it's a whole ecosystem now. And in a way what it feels like for me is that I joined because I kind of like wanted to try to help fix the media ecosystem. And I was fortunate to like join a group of people who had a really good vision for what that looked like. And obviously subs like didn't fix the media ecosystem, but in many ways it has already succeeded. And like,
building, like showing that you can actually blog full time and make a living doing that. You can be independent, you can have a subscription model, you can do all of these things. And so ⁓ once we, once like I'd gotten to be a part of that team and like building out that infrastructure at a certain point, I was like, well, I still really want to be a writer. And maybe now instead of getting like a staff writer job at like a small local newspaper and working my way up from like
getting paid $40,000 a year on like a tiny beat, now there's a different way to be a writer. I've seen it happen at Substack and so like, I want to try doing that by myself.
Grace Shao (05:40)
That's really amazing. And I think we bonded over this a little bit as well when we were when we met in person, like I left media, like traditional media, probably in 2019, 2020 as well. And it was just like a pretty grim feeling, to be honest, like when you kind of grow up and idolizing a lot of these like traditional media publications or these platforms, but you just like every day you're just seeing people getting cut and laid off. And a lot of the interesting beats were just no longer kind of the click baity ones that would make money for them, unfortunately. And then I think to your point, like
Jasmine Sun (06:06)
Yeah.
Grace Shao (06:09)
Azeem Azhar and Bill Bishop, like they started the newsletter business a lot earlier and actually just recently wrote a guest post by Azim on China's AI ecosystem as well. And I think it's really interesting working with people like them because they're just so savvy. Like they're not just media people, but they actually position themselves beyond just a reporter. And this is something we could go into because I think you and I, both of us started off our interest and passion in journalism, but a lot of it has actually taken us.
you know, than, you know, identifying as a journalist, which is mostly to tell the story of what is happening and just relaying the facts, right? And we do definitely need people like that. But I think what you do now, you call yourself an ⁓ anthropologist. I actually call myself an analyst, tech analyst. I think it's really interesting how we're positioning ourselves, but content is still our main, I guess, source of distribution of our thoughts or knowledge or insights or observations.
Jasmine Sun (06:42)
Mm. Mm-hmm.
Grace Shao (07:03)
On that, I wanna double click. So what is a AI anthropologist and what is it that you actually really focus on with your writing? Because you do cover a wide range of topics and I think that's also why you obviously write beautifully. I think I love your writing and it's such an easy read even though they're quite long. But I think that's why you've also captured a really wide audience of Substack. It's not just like a niche kind of group of people. So I think I just wanna hear about that from you.
Jasmine Sun (07:20)
Thank you.
Yeah, I mean, I struggled with it because I think that it's always in some ways, it's easier to have a niche, right? Like you can be like the number one writer in like electric vehicles or in open source models. And you write about that thing day in and day out and build an audience that way. And I really admire those people. At the same time, I think like just noticing what I like and the way that I learn. think that's sort of where the anthropological perspective comes in. And I think for me, it's about two things. So one is that I really think by connecting the micro and the macro, so like noticing some like very specific small thing and then trying to ground that in a bigger trend. also when I look at a big trend, like when I look at stats about how many, say like Chinese ⁓ AI PhD graduates are moving to the US, like I look at the charts and like the charts matter to have the broad context, but then I become very interested in like
Why are people doing that? I want to go interview some people who have made that choice and to deeply understand that choice. People say, will AI automate everybody's jobs and how many jobs will be automated? And one way of looking at that problem is just to interface with the LLMs and the models and to see the capabilities in this kind of theoretical or technology focused way. But then the way that I kind of think about it is, well, presumably people are already trying to do that in their organizations. I would really like to actually like...
go inside a firm that is trying to take one of their roles, automate them with AI, and I want to see up close, like, what does that look like? Where does it work and where does it doesn't? Because I think there's a level of, like, specificity and groundedness that you get from close observation. And I do want to ground that in the bigger trend. Like, I think that you need both perspectives, but,
moving between like very specific things and very broad analysis is like what I spend a lot of time doing. And I think the practice of anthropology, like academic anthropology is very much about that as well. the classic image of the anthropologist is like, you know, like they go to another country and like live with like a small tribe or something like that. But like for me, I'm sort of like, well, Silicon Valley is a pretty weird place. Like startups are weird places. Like you can take a very similar kind of approach and embed yourself in that world and really try to understand like what are the
rhetoric people use, what are the rituals they practice, what really motivates people to build what they build. And I do think that it's globally consequential, what beliefs that people have and what inspires them to build these technologies, because those beliefs sometimes do get exported onto the rest of the world and the user base. And so I just really enjoy trying to understand that from a cultural perspective and from this very specific perspective in addition to taking the broad lens of analysis.
Grace Shao (10:07)
Yeah, I think on that, let's zoom in on your recent trip, right? You just got back to Silicon Valley from China. did a big, I think it was like a 10 day, 15 day trip, right? And you wrote this beautiful piece called America Against China Against America. I say beautiful because it's like you said, it's really not just a story about tech development or tech innovation or the economy, which a lot of people write about, including myself even. But actually you weaved in your personal life and your personal family story into this. And I thought that was really well done.
⁓ Just want to kind of zoom in on that. What was your biggest takeaway from that trip? ⁓ How did it inspire you to this piece? And then I think I would love to hear about your takeaways on each of the cities as well that you visited.
Jasmine Sun (10:48)
Yeah, it was like a really fun both like trip to go on and peace to write. can give a bit of context. So basically what happened was, so for me, I'm an ABC. So I was born in the United States around Seattle, but my parents are immigrants. So my parents were born in Shanghai and most of their family still lives in Shanghai. And so I grew up here, but I grew up sort of going back to China, mostly to Shanghai, like every two to three years, mostly just to visit family. And one thing that's happened since I like, you know, went to college, like
started having a job and started being able to take more trips by myself as I sort of became more interested in China now as just like, this is like where my family lives. So I should just go back and hang out with them, but also to understand Chinese culture, to understand the political system, to understand like what's going on in tech. And so it sort of like started to intersect and I started trying to go more intentionally and spend time both connecting of course with my extended family, but also trying to
understand the broader ecosystem of like for me, I'm really interested in technology. So I really wanted to know what was going on in Chinese technology. And basically me and five of my friends who were also quite interested in the Chinese tech scene decided this year to take a it was a week and a half trip sort of going between Hong Kong, Shenzhen, Hangzhou, Yuyao and Zhejiang and then Shanghai to sort of meet with different sorry about the noise.
meet with different tech founders and technologists and researchers and engineers and just understand what's going on. And ⁓ we did it because I think right now in Silicon Valley, China is sort of a backdrop and a mirror and a foil to a lot of what is happening in US tech and whether that looks like this new interest in manufacturing in the US, right? Like I think a lot of that is looking at China's advancements in
solar in batteries and EVs and saying like, China is doing really well in like high tech manufacturing and the US is not what's going on with that. think obviously DeepSeek and open source AI more broadly than DeepSeek has been a huge wake up call. But I think at the same time, I noticed that while Silicon Valley has become very interested in Chinese technologies, of course, like TikTok, et cetera as well, and sort of recognizing its strength, not that many people are actually going and seeing it for themselves. And that really results in a lot of generalizations that I'm not actually.
you know, sure, always true. And so I also kind of felt like it would just give me personally a more nuanced perspective. And that was a lot of what motivated me and my friends to go. In terms of our takeaways, there's so many, was like, really, really, there's a lot to that I thought about. I think for me personally, the thing that I thought the most about that a lot of my piece ended up centering on was like talent flows, actually. So really thinking about like an economy is made out of people, like technology is made of people who have technical skills.
And when do people choose to migrate between one country and another? And like, what are the implications there for tech development? And I think this is partly because I have some personal connection to it. Like my grandmother was born in ⁓ Indonesia. She's Chinese Indonesian. But then she actually in the, I believe 1940s ended up moving from Indonesia to Fujian province. And that was because the PRC had just been founded and they were
calling back overseas Chinese mostly to study science and engineering to help build the new China, right? And so in that case, was like, PRC was trying to bring overseas technical talent and young, ambitious talent to the country and nurture them. And so my grandmother ultimately became a math professor and that was her career. Then when ⁓ her daughter, my mom, my mom grew up in Shanghai, she went to Fudan for university. She also studied math and computer science.
But at that point in time, it was the 1980s and Deng Xiaoping was saying like the thing to do right now where China is more global, we want to be connected. And so you guys should actually go abroad and learn. The top students should learn ⁓ from other countries, learn from the West, ⁓ go do master's programs abroad. So my mom and my dad went from Fudan to Duluth, Minnesota in the U.S., did a master's program in CS and math. And then that sort of, was at that point in time, the U.S. was just so much more advanced.
Technology was more advanced. The cities were better. They were wealthier. And it was like very clear that that was the place to go. And then now it's sort of like, I wouldn't say it's flipped again, but like the dynamic has changed where it's really interesting. For the first time on this trip, I met many Chinese technologists who had studied in the U S and many of them even worked in Silicon Valley companies in the Bay area. And then they still decided to go and move back to China and to build businesses there. And that it was sort of a realization that it is no longer the obvious
choice that the US is the best place to be if you are an ambitious technologist. And I can notice that in even my own friends here, it's really stressful to navigate visa situations and changing immigration politics and all of this. There's also a lot of suspicion right now of Chinese nationals who are working in AI, which is something I've covered in the past. But I think the lens of just trying to deeply understand both
what motivates talent who could be bicultural, who could choose to live in either country? Like what makes one place a better place to both build a life as well as to build a business or to build technology? And then what does that imply for like the future of the, I guess, technical innovation? Because like a lot of Americans in tech, like I very much worry as an American that if global talent no longer wants to work in the US, like we're not gonna be able to make as many scientific and technical advancements because
So many of the companies, so much of the research is done by like immigrants coming from around the world.
Grace Shao (16:12)
I like absolutely love that you share this story and I wish we talked about it in person when we met, but it just feels like, know, I completely echo that a couple of years older, but I think, you know, have very, very similar upbringing. my parents were sent to Canada, essentially also part of that, like go abroad, like study abroad. you know, like it was like, they were kind of like the more top business how at that point to set up businesses in the West, also in the West coast of Vancouver.
Jasmine Sun (16:31)
Mm-hmm.
Grace Shao (16:40)
I'm commuting back and forth between Seattle and Vancouver a lot. Similar to you, my grandparents, again, very privileged in that generation that they went to university and also studied medicine and science. So that whole generation, they were like engineer professors and doctors, and they didn't really genuinely believe staying in China versus actually having the option at that time to go to the US actually. A lot of their families did leave.
Jasmine Sun (17:04)
Mmm.
Grace Shao (17:05)
was a better option because they thought like, you know, China is going to be on like an upward trajectory. But it's really interesting to look back and because actually everyone was essentially living in poverty, but they still had that kind of faith in it at that time. And that brings me back to like kind of my own personal experience and how I relate to what you were saying is like, so I spent like majority of my life in the West. I did my education in Canada, I did my education in US and then my father was so adamant about pushing me to go to China for an exchange program.
Jasmine Sun (17:18)
Mm-hmm.
Grace Shao (17:34)
like when you go through kind of like assimilation and then go through like growing up in the West, you don't really identify that Chinese except for the fact that like you have a family in China. So really reluctant to go at first in university. And remember I had like my whole German study abroad thing set up and I was like, I'm going to go to Germany and learn German. I was taking German in school. My dad was like, I'm not going to pay for you to frolic with German boys.
Jasmine Sun (17:57)
Hahaha
Grace Shao (17:58)
And he was just like, you better go learn Chinese first before you learn German. And I just thought he was so close-minded and backward thinking at that time. And we actually had a lot of arguments around this. now looking back, it was just such an eye-opening experience and a complete change in my life. Because what happened was I went to Tsinghua, which is 2012, and I met people around the world. They were Dutch people, Spanish, Australians, Americans, and everyone.
where they were trying to learn Chinese language and they're trying to learn about the culture. that was also peak, like China being very open and it was such a good, vibrant environment. I think it was post like, ⁓ wait, Olympics, right? There were so many diplomats, journalists, intellectuals, authors, artists in Beijing. It was a great place. And I think after that, it completely changed my mind. was like, wow, if they're learning Chinese, I should be learning Chinese and I should
Jasmine Sun (18:30)
Yeah. Mm-hmm.
Grace Shao (18:50)
learn about China. And like you said, I started like kind of intertwining my personal life and my professional pursuits then and there. And then, you know, 10 years, well, 13 years later now, I live in Hong Kong and I've been moving around between Beijing, Shanghai, Hong Kong, Singapore for the last decade. And I think to your point on the talent flow, it's something I also talk to people about. think when I first moved out here in Asia, a lot of friends back home were like, my God, like, what's your lifestyle like?
Jasmine Sun (18:52)
Mm-hmm.
Grace Shao (19:16)
There's a lot of like over stereotypical generalization of even what life is. I'm like, dude, I actually have like a really normal life. I love my I actually get to be both like I get to be both Canadian and Chinese. And people actually recognize that because there's so many of us essentially like the whatever BBC is now in in these big major metropolitans. And people really understand you can straddle the two cultures and be bicultural, bilingual.
Jasmine Sun (19:20)
Yeah.
Grace Shao (19:43)
and have that identity. And then I think on top of that, on the AI tech ⁓ talent flow, also noticed when people used to go to the West for a better life, which is what our parents did. And I think objectively, GDP per capita was just much higher. Your Your of living was much higher in the US or Canada. But nowadays, when I talk to people, like you said, people who are Stanford grads like yourself or MIT grads, can frankly easily find a job in the US.
Jasmine Sun (19:57)
Yeah.
Grace Shao (20:11)
well, I wanna go back to Shanghai or Hong Kong or Beijing because it's my culture, it's my identity, it's my language. So that's a huge ask of it that people actually now recognize that. And the other part of it is like the center of life is essentially the same or if not even higher in some places because things are somewhat still cheaper but you can still access quality food, quality education, quality housing, right?
Jasmine Sun (20:31)
Mm-hmm.
Grace Shao (20:36)
And I think that really changes things. So I really enjoyed your piece. I thought it was really nuanced and it really reflected on like what the average talent, average person's really real mentality instead of just like overgeneralizing a high level geopolitical kind of narrative. ⁓ Yeah, so sorry, I go in a rant just because I got really touched by your story and I wanted to share mine as well because we've never talked about it.
Jasmine Sun (20:51)
Yeah.
No, thought that was, no, I'm really glad that you shared that because I think that that's actually one of the things I really enjoyed about writing the piece because it was like meaningful for me to write because I was sort of connecting my, like you said, my like professional intellectual interests with thinking about my family story and then getting to meet other people and hear their stories and like how they thought about this decision for themselves, for their children. Like it's just like the place that you build a life in is like one of the most consequential.
and personal things that you can imagine, right? And so, and I think for people who, there is now this group of people who sort of like have access to both countries and cultures, which in itself is a privilege, right? Like many people cannot make that choice, but I don't know. think that there's, it's an interesting sort of entry point for thinking about the bigger issues. And that's also kind of what I mean by the anthropological stuff is like, how can you understand talent flows?
if you only look at the charts of how many people are going or like how many papers are being published. Like those are obviously inputs, but like you said, like I think things, concerns around lifestyle, around culture, around like even just like quality of life stuff, like that showed up so much in my conversations with folks. And so yeah, it's just taking a little bit of a more human lens. I'm like, okay, let's actually talk to the, when we talk about talent flows, let's actually talk to the people. Like how are they thinking about this?
Grace Shao (22:13)
Yeah, and I think on that, you know, let's kind of talk about the cities you visited. ⁓ Shenzhen, Hangzhou, you know, have really gotten a lot of attention in the West recently after the deep seek moment per se. But if you follow China's scene, you know that these cities have been vibrant in the tech sector for a very long time. For context, know, like there's six dragons of AI startups in Hangzhou, know, Alibaba.
Jasmine Sun (22:18)
Mm-hmm.
Grace Shao (22:39)
⁓ Alipay, these mega tech behemoths, they're all based out in Hangzhou. It's like always been known kind as a big tech hub, but also a massive consumer hub because of all the supporting roles around Alibaba. It's historically a big center for like merchant and trade, for art and culture. Shenzhen on the other hand, you know, actually was like, essentially manmade. It was like
the center for entering exiting China. was like the bridge between mainland and Hong Kong. It was seen as like a pilot trial zone for like capitalism. And you know, it's really played a really, really major role in China's economic rise. And I think we can talk a little bit about that. what was your feeling like, like, we're on the ground? Like, what's the vibe?
⁓ Who are the people you met and you know what makes these two cities so special compared to maybe a more well-known? So-called first-year city like Shanghai and Beijing
Jasmine Sun (23:38)
Yeah, I mean, again, I can only sort of speak to my trip, which is not so long. I've never lived in any of these places long term, but I was actually very excited because it was my very first time going to Shenzhen and Hong Kong. Like I had, I usually always go to Shanghai. And so I just like hadn't gotten to see those places. In Shenzhen, we went right after Hong Kong and we talked to an AI researcher who worked at like one of the like largerl like AI companies that does AI and got lunch with him. We talked to a representative who did AI stuff with Tencent and also got like the official Tencent tour. So at the Tencent headquarters in Shenzhen, we met with an urban anthropologist to sort of like get more of like a city planning and like cultural perspective who was born in the US and she's American, but she had just decided initially for academic reasons, then because she had built a life there to move to Shenzhen.
I think in the 90s. What else did we do? I think we met a few other folks, but those are the ones that come to me off the top of my head. Shenzhen was, oh, and then in terms of activities, we also went to Huaqiangbei, of course, we went to that big electronics market and everyone, a bunch of us bought some stuff, whether it was like, here's a knockoff AirPod Max or someone had a bunch of equipment that they were trying to do for a hardware project.
course, that's like a great place to do that. I went to some of like the urban villages. Yeah, so that was really cool. mean, Shenzhen is interesting because you can kind of… you can really tell that it is new, right? Like when you were there, it feels manmade. ⁓ I don't think that for me personally, like I could live there that easily because for me, I really like walkability, for example, and it's just like it is not a walkable place. Like the roads are so big, the buildings are like…
big and far apart and where I was like taking like really long DD rides like anywhere we have to go. And so there's a way in which it just like feels really big. But I actually, really, those meetings that we had like with the folks working in AI at Tencent and the other place, they were really memorable and they're really interesting. think the thing that I mainly felt was I was just like, wow, like people are really energized right now and they...
people in Chinese AI, and I continue to feel this when I went to, for example, Shanghai, I met some consumer AI founders in Shanghai. the way that, for example, Western US policies, like tariffs or expert controls, actually really motivate people to build in domestic AI ecosystem. And you can kind of sense the spirit or the motivation that comes from that. That, of course, came up.
There are also ways in which I think that I felt that there is similarities between the US and Chinese tech culture. The AI research we met was on Twitter nonstop, American English language AI Twitter all the time. We talked so much about Twitter. like, you can go across the ocean and see the same algorithm. I thought that was very funny as well. But overall, I think it was a very good entry point. It was a place that was like...
like you said, was sort of built to test run capitalism. And just like the idea for, at least to me as an American, the idea that like, you can just decide 50 years ago to take a swamp and be like, we're just going to build a city from scratch here. And we're going to have actually quite different, like a political approach, like the hookah system, we're actually going to give out a lot of hookahs. So to encourage people to move here and we're going to liberalize a little bit of the economic policies and have some slightly more labor rights laws and just like try all this new stuff.
and to actually see that succeed and to go into this giant thriving city with nice malls and big tech companies. I think just that in itself feels quite impressive and interesting to me, to know that it is possible to succeed in doing that and to see it happen. I think I do wish we got to see some more maybe robotics things. think we had a meeting with a robotics company that unfortunately for various reasons was
got cancelled and so I didn't get to see as much of the hardware as I was excited about. But yeah, that was my time in Shenzhen. And then Hangzhou, I'm going to be so honest with you. We loved it, but we did not do any, we actually didn't do any tech stuff. We went and we were like, the lake is so beautiful. And just decided to be tourists and...
walk around West Lake and go to the temples and like drink tea and just like luxuriate in Hongjo. Because as you mentioned, there's so much history there and it's a very like literary poetic place as well. Like it's not just about, like the tech stuff is relatively new. So I frankly, we didn't really do that much technology related stuff there. We just kind of enjoyed Hongjo as a city before going to Yuyao, which is where we went to this like materials factory afterwards.
Grace Shao (28:17)
Yeah, think Shenzhen and Hangzhou are just so sprawling. I spent like half a year or a bit longer in Hangzhou during COVID lockdown. Basically, I was supposed to go for like a two week work trip, then two weeks, well, two weeks plus two weeks of quarantine to four weeks, and then I ended up just being there for like eight, nine months, or even around then. And then it's just like such a big city. And like you said, like these cities actually, from one area to another, you have to take these like half an hour, 40 minute like car rides.
Jasmine Sun (28:22)
Mm.
Grace Shao (28:43)
And like, it's so crazy because they're all so new. Like, it feels like things are lived in and then it's just like land, like just empty land or like fields of nothing. And then another area. And they call them like science parks or like tech parks. And usually they have like kind of like a flagship company a product headquarter there, like, you know, either like Alibaba unit, like kind of taking over like a whole region of the offices and then some startup trying to like...
Jasmine Sun (28:57)
Yeah, yeah, yeah.
Grace Shao (29:10)
lean on that or leverage their resources or then, you know, have another area with like a few tech companies, I'm just like kind of cluttered, right? It's really interesting. It just feels so different from what you would see in the West. Like it's so much of it's not organic growth in that sense. And I think that's something we can talk about as well. Like, like it's
Jasmine Sun (29:26)
Actually, yeah, sorry. One tangent really quick is that I remember as I was catching up with Dan Wong after my trip because he just did his book that everyone's read, great book, highly recommend it. But he actually was like, Shenzhen is just like Silicon Valley because, not Silicon Valley is in San Francisco, which is like an old city with a lot of organic growth and history. He was like, Shenzhen is like the rest of the Bay Area. It's like being in Cupertino or Mountain View.
because there are just these like random, it's so much land, cause it's California and there's these big highways and it like takes forever to get somewhere. And then there's literally an industrial park and it's just a bunch of office buildings where like everyone works and maybe some residential housing where the office workers live. And then there's more land and then like another industrial park. And there's not that much in between. There's not that much like culture frankly like going on. But I thought about this a lot. I was like, I think this is correct. Like even in.
The early days in, I don't know if it was the 50s or 60s, but Stanford was founded in the 19th century, but then Stanford Research Park, which is where all the semiconductor companies were in the US, was created in the 50s or 60s or something in this research park right next to the university so that Stanford professors would go between there and ⁓ the semi-fab, or the semi-companies and then like... ⁓
I think there's a DOD office nearby and they would literally just sort of circle these office parks. So I think actually the US, a longer time ago, was doing a few similar things, but sorry.
Grace Shao (30:55)
No, that's really interesting actually. I think now that put it that way, I can totally see that because when I was there, I would just feel like, okay, I'm like visiting one campus of Alibaba. And then the next thing is like, I know I've got a meeting and it's like an hour and half drive out. I was like, oh my God, why are these companies like, why are the offices of the same company so far away? But actually in between, just feels kind of like a barren land, like you said. I think they're trying to face that up a little bit. And I mentioned it to you before you guys went to Hangzhou, just like there's a lot of new art areas and Hangzhou has now become like a really hip city as well.
Jasmine Sun (31:03)
Mm-hmm.
Grace Shao (31:23)
just become but because of all the DTC hype like you know a couple years ago and there was a lot of like private money going into that those DTC brands there's so much like influencer vibe-y set up things like a lot of really pretty like libraries and parks being built so that's kind of its own thing anyway we're going about it on a tangent I do want to give the listeners a bit of a context on hukou so you mentioned it I think it's not related to tech but it is really really relevant to understanding tech
Jasmine Sun (31:35)
Mm. Yeah.
Grace Shao (31:53)
migration. So in China, who calls essentially is your ID or like where your red, your family's registered and to be born in Beijing or Shanghai or Shenzhen or any of these so called top tier cities, it gives you an upper hand because you have the registration of that city. And then once you have that, you have the rights to buy property, ⁓ you know, enroll your kids in school and local system. So for a lot of talent, even so I did a
program at Tsinghua as well after I finished my undergrad in Canada. And for a lot of like what they call non-local students who come to Beijing to study, their bigger goal is to get the Beijing Hukou rather than just a high-paying job. So they will actually turn down maybe like MNC, you know, like at Deloitte or whatever, or PwC, but they will join a state company or a state media or state something because they will want to get the Hukou. It's common, almost like domestic immigration.
If you work there and pay taxes in that province or city for five years, you get that city local hukou. So it's really important. think Xinjiang, you mentioned, used that as a tool, essentially, to attract bright students and talent to a swampland and build up the city from scratch. And it's still kind of utilized to basically attract the best, smartest people domestically for them to move around cities that give the best economic options.
So okay on that I want to touch on China and US. You know, you did use that comparison I thought was really really smart on how you some of these like tech cities actually do look like what California did look like a couple years ago or just a couple decades ago but if you really take a step back and look at the culture and the sentiment around AI tech Do you see similarities or do you see a lot of contrast?
know, lot of people talk about tech optimism, tech-dumism, you know, driving the different areas. And you mentioned there's so much energy coming out of China. Do you see that in the Bay Area right now, or how do you compare that?
Jasmine Sun (33:53)
Yeah, it's really interesting. think sometimes one of the takeaways that I had is that I think sometimes when Americans imagine China and Chinese people and Chinese technologists, there's maybe an assumption that people are very, I guess, or something because it's like, China is authoritarian and has a single party state and that might mean that people are motivated by these beliefs to build technology that will have a certain politics to them.
Whereas at least my experience of the Chinese technologists I talked to seemed quite practical and actually ⁓ not very ideological about what they were doing. It seemed like most people's main goal was I want to build a really good, really successful piece of technology and build a successful business and make a lot of money. And relative to the AI founders and AI researchers who I meet in Silicon Valley, there's actually a lot less sort of like...
big mission talk. in Silicon Valley, of the things that people immediately notice is you're not just building AI, you're building the machine god. This is super intelligence. This is something that is a brand new species that us humans will never imagine. And then they will be the ones to rule over all of us. Whereas this kind of belief, just frankly did not hear about. Maybe there are people in China who I'm sure some people might have these sorts of ideas. But for the most part, ⁓ I think people are a little more focused on what is a particular product or technology that they are building and how can they, you know, make that succeed. Where Silicon Valley is like very much a place where everything is like justified in the background by like some like grand theory about the world, whether that's again, like AGI, the machine god or like, actually, the machine god is going to kill us all. We should be like super worried about that. And I mean, I don't know your experience, Grace, but yeah, I just like didn't see that much of that. I think the other thing that is ⁓ interesting is like, I will say like, there is definitely some amount of like
⁓ patriotism, suppose, like not necessarily from everyone, but like certainly some of the Chinese technologists I met, like they saw the thing that they were doing as part of a Chinese community or a Chinese national project. My sense is that it wasn't necessarily about like the government or something, but like more a sense of like wanting to strengthen the Chinese ecosystem in general. So like, for example, in AI, like Chinese AI, there is an identity of like, we are really excited to build great Chinese models because we want to compete with US models. We want to show that China can build just as good of technologies as America can. And frankly, one of the thoughts that I had was I like, I wonder if Chinese technologists would think this way if the US did not have a similar framing. Because in the US, a lot of technology is very much framed around. The reason we have to ⁓ build or regulate AI in this way is because the US has to beat China. We have to have technological dominance, technological primacy.
And everything's like China, China, China. That will justify any amount of spend is because it's really important that we beat China. And this is like, I don't know. I don't love this framing mostly because I don't really think technologies are about entire nations at once. It's not a monolith. The US and China each have many different technologies and companies and people within them. But I think that is what motivates, again, things like expert controls and very strict tariffs. And that then I think is perceived, at least by some of the people I met in China, as like...
okay, you guys are trying to bully us and shut us down and make life harder, which kind of sucks, but also that's gonna energize us more about working harder and collaborating more because we sense that the US is trying to slow our AI development down, which is true. The US is definitely trying to do that. So I think that was an interesting dynamic to observe, but otherwise I would say more practical. And then there are of course commonalities. I think, again, everyone's on the same Twitter algorithm.
⁓ Everyone is just really excited about building impressive technical stuff and sharing that with not just like sharing it with themselves, but I think like sharing it with Twitter and with the international community and like having other people like build on your models and your open source projects. think that sense of like recognition that you get like that's definitely certainly shared between I would say like technologists I meet in the US and in China.
Grace Shao (37:52)
Yeah, I think for sure, like one thing you mentioned, think the geopolitical narrative is so much more kind of the front and center of the discussions, even amongst the regular people. I'm not even talking about like policy people or like, know, politicians. It's just like you talk to investors, you talk to tech developers, you talk to even journalists, right? And people are like, yeah, but like the China versus US arms race narrative is front and center in the US. And that's how I feel as well with my like American
Jasmine Sun (38:06)
Mm-hmm.
Grace Shao (38:19)
like network or friends or whatever, right? Whereas in China, think exactly to your point, it's so much more pragmatic. And I don't know what it is. I think because on the surface, the country might be really political, but people actually learn to be really, really almost like detached from politics because the average person frankly has that much political involvement. So then you actually realize the average person is very much just focused on their own little pie or their
Jasmine Sun (38:45)
Totally
Grace Shao (38:47)
go around and like vote for my like my representative of my region, right? So then I might as well just focus on what I can do well to bring back almost just focus on economic gains and a lot of times or you know, at this point, like you said, I think especially the younger generation entrepreneurs are very focused on like their own pride or ego or like self actualization. If you must, you know, categorize it under the Maslow hierarchy, kind of that kind of thing. That drive is not
Jasmine Sun (38:52)
Mm-hmm.
Yeah.
Grace Shao (39:14)
Ideological I I genuinely believe that as well and I think I really agree with your observations So like when I talk to people here, it's just like okay. I did something really cool. I want to show the world How could I monetize this and then I can buy big house in Beijing. That's pretty much like this process So I think I think that's really interesting and I think on the tech doomism and optimism side of things I think again to your point like on the pulse safety
Jasmine Sun (39:24)
Yeah. Yeah, exactly. Yeah. Totally. Yeah.
Grace Shao (39:41)
AI policy, AI safety side of things. There's obviously efforts done by both sides, regulators and like AI safety community. But in general, I also feel like in China, people are less wary, at least on the average mass consumer level, because I don't know if it's just, I wouldn't say it's lack of understanding. I think it's just because in the last three decades, technology propel people's economic ⁓ prosperity so much, like in the sense that
Jasmine Sun (39:54)
Mm-hmm. Yeah.
Grace Shao (40:07)
You know, if you used to live in a 34th year city in China, like, or a small rural area, you had zero access to anything, you know, high end, essentially, you can't, can't buy certain food, you can't access certain products, you couldn't, you know, get information outside of it. And technology allow them to actually just obtain everything that you can in Shanghai, if you live in Shanghai. And I think in general, people have this, like, belief that technology will get us to a better place.
Jasmine Sun (40:29)
Mm-hmm.
Grace Shao (40:33)
And then on your point on the whole AGI machine and this like God-like AI mission thing, I actually never really heard people talk about it in China. I think the Moonshot founder is the only guy that even talked about how their goal is like, you know, reaching the moon and they have this AGI mission. But in general, the AI community in China doesn't talk about AGI as much as how it is being talked about in the AI community, I think, in the West. So it's really interesting. I'm glad that we had the same observations. That means I'm not completely off.
Jasmine Sun (40:58)
Yeah.
No, I'm happy to hear the same because actually, like, I think that point you made about the ideology, is what that was my hypothesis as well was that, like, you know, like, I even just thinking of my parents are that a lot of the Chinese people I know, it's like, well, speaking up about your own political opinions usually doesn't get you in a very good situation, right? Like, you don't really want to be the one who's like advertising your own politics all the time. Like, it's unless you are actually like your goals, like, you know, join the Chinese political system, but like most normal people.
Jasmine Sun (41:31)
who are doing business or something. You don't want to have your own political thing going on. You want to be, fly under the radar, just focus on yourself. I think, like you said, I think it almost creates an environment where people are more, I guess, detached. And it's like, why would I worry about that? It's like other people seem to worry about, my job is to build technology. so it's just very different. Whereas I also wonder if there's a funding thing going on and where ⁓ in Silicon Valley, part of why founders in AI and beyond AI tend to be really big about their ambitions is because the way that fundraising works with BCs and with recruiting, it all operates on this, you have to explain your thousand X vision. You have to explain how just have a lifestyle business, right? And so in order to raise money from venture capital, you have to fit it into this really vast totalizing vision and
Grace Shao (42:24)
that, actually, I wrote about that as well. And I think in China, a lot of it is like you have to prove your concept before you get your money. And want to see the product to really hit the ground working first. So again, the first like zero to 0.5 requires you to really just focus on a product instead of like the PR. So like And you know, like even Chinese culture, there's like, it's kind of true 枪打出头鸟, right? So it means like, head sticks out first as a bird, you get like shot first. It's like you don't want to be the person who like really sticks out. You want to kind of lay low and then until you prove yourself. And then in the West, think, you you and I, grew up in the West and it's like a really different concept. Like it's not like pure PR, but you definitely like talk the talk more with your narrative. You start with selling yourself and you sell the vision first before you really kind of, well, you get that and get to prove yourself. I'm not saying the part is not good. It's just like, it's a different mindset.
Jasmine Sun (42:54)
Right.
Grace Shao (43:15)
seen in everyday communication. And when I first moved to Asia, I'm naturally extroverted, but I think it really irritated people that I was just like talking so much. And I'm just like so loud and like, you know, boisterous. I think, but in the West, it's kind of like, I think it's, it's, it's applauded, right? It's like, you're taking initiative. You're you're so ambitious. You, you're not scared of voicing your voice. I remember like,
Jasmine Sun (43:27)
Yeah.
Grace Shao (43:39)
my advisor pulled me aside, like, Grace, you know every, not every female feels so comfortable voicing her opinions. I was like, I don't know what say about that. That's really awkward, But like in chat, people really just operate differently. And I do, I did learn to, you know, kind of even present myself a bit differently when Chinese counterparts and business people, right? So I think that that comes in the culture aspect as Yeah.
Jasmine Sun (44:02)
Yeah. VC is the total opposite, right? The whole thing is like your first check, you can have a 15 minute conversation with a VC and they might, if they like you enough, they will give you a C check of like a million dollars without any product. You don't even have a prototype yet, but if they think you have a great vision and you have great drive and great ambition and you can express that verbally and in your presentation.
then that's where your first check comes in. Like you said, you got to prove yourself later and you can make money later. But VC allows you to spend a lot of time without making any money and without really having customers and they'll keep feeding you money if your vision is big enough. So it's just like there's a way in which the culture intersects with sort of like the financial world as well that changes. And yeah, the other thing I was gonna say that you said that really was interesting to me was the bit about ⁓ like doom and techno-optimism and things like that.
⁓ I do think this is thinking maybe less just about the technologists themselves, also the broader public and the citizenry and how people react to new technology. Whereas in the US, when you look at the Pew Research surveys of like, do people worry more about AI or are they more excited about it? It's very negative. The US and the UK as well, think most Western countries are much more negative on AI than they are positive about it.
When you ask a lot of Americans, like what has technology done for you in the last 10 or 20 years? People will say like, well, we have smartphones and social media and like TikTok and YouTube, but they'll say like, I don't know if that was good. They'll sort of be like, oh, I feel mixed. Like, yeah, like my phone's really good. I have this like super addictive short form video feed and like maybe that's fun.
But a lot of Americans will be like, yeah, this is like destroying my kids' life because like now they're addicted to TikTok. And like, I wish I wasn't so attached to my phone all the time. So people feel this like ambivalence, I think, about technology. And they're not really able to point to a lot of things that are like a uncontroversial good thing in their daily lives if they are not like someone who is really like obsessed with gadgets and the cutting edge. Whereas ⁓ my sense is like, even when I go to China every like two to three years or whatever, like between like every two to three year gap, like new...
technologies appear in my life in China to make my life easier, you know, like when they first started being like the phone charging booths everywhere or like my mom was showing me this year how she can get coupons for recycling more and like the recycling machine automatically like weighs how much recycling you put in and like gives you a free coupon or ⁓ like everyone talks about the high speed rail, but like the high speed rail is amazing or like going cashless. Like it is so much more efficient and like, I think you just like feel the rate of
material improvement and you see it around you when you look at the skyscrapers and the screens and the trains and whatever. And like, I think that gives a sense of an overall like progress is real. Like yes, there are hiccups. are not going to be every, not every single thing is going to be good, but in a general sense, this is like here to make our lives better. Whereas like San Francisco, like where I am right now, you look outside today, you look outside 10 years ago, 20 years ago, it basically looks exactly the same. Maybe you have more tech billboards, but like the billboards are annoying because like, it doesn't look cyberpunk out here, you know? Like I remember somebody told me something like, I heard that Shenzhen looks like what ⁓ San Francisco is supposed to look like. And then you show up to SF and it looks like, you know, like 1940. And so I think that makes people less pessimistic, less optimistic, because they're like, okay, I have this phone in my pocket that like, I'm kind of addicted to, and like, that's what all the money goes to, and you want to get even more money to build even more technologies that will just make me addicted. So it really makes me feel that like I wish in...
I wish in the West we had more everyday life type of technologies. think Waymo has actually been great for this. Waymo is amazing. Most people would take Waymo's, really love them. They feel much safer to me and they're reliable, they're super safe, they're a great experience. I want to see more stuff like that where it is frontier technology and it's really for the everyday type of person making your life better.
Grace Shao (47:53)
Yeah, and I think I'll give the West some credit in the sense that also like the base of standard of living was so much like you said, like when I go back to Vancouver, like if I don't go back for like a year or two, I'm it's the same as it did 20 years you know, with like one or two new high rises, but the standard living was already high. So I think a lot of the technology that was big and became big essentially only helped the people who worked there. Like, you know, they created the FAANG and the FAANG all made a of money. Essentially, it's like a new
Jasmine Sun (48:10)
That's true.
Grace Shao (48:20)
middle upper class or a high income bracket that was created in the society, but it didn't really help the average person as much. But to your point, when you look at logistics, digital payment, a lot of these things that are more mainstreamed in China, they actually propel the living standard because it jumped from pure cash to cashless. And they skipped the whole Visa, MasterCard, credit card ecosystem as well. Anyway, we can talk about forever, but I think
⁓ before we go into, you know, talking about just kind of the risk and fear around AI and technology, I do want to, double click on the tech talent, ⁓ kind of topic one last time. So in one of the A16Z recent, ⁓ podcast, Katherine Boyle talked, like one of their general partners talked about, there's a shift in Silicon Valley culture. And I think, you know, you're right in plugged in there. And they're saying that essentially for like 20 years or so, or 30 years with internet boom.
all the top graduates, like again, like yourself, from Stanford, MIT, Harvard, whatever, these students, these young talent were looking to join tech companies because of the sexiness of it or how glamorous your life could be or all these nice perks and high income, whatever. All of a sudden, because of ⁓ Trump 2.0, we have Elon Musk, how to get involved, well, no longer, but you have like Chamath, you have like sex, you have a bunch of these like...
big shot, big name tech people. Whatever we can talk about, we don't have to talk about their political like actual beliefs, but just people who made it quite big in technology, joined the government and served the government. And it's all of a sudden made the government agency sexy again. And you know, it's made it like, it's not even purely for money, but people want to kind of go and serve their government again. Do you think that's driven by like nationalism? Or do you think it's
Oh, are you seeing this trend, like young talent wanting to leave technology to go back into government? for a very long time, people also understood and think that if you are a young talent and you're a CCP member, you are like really ideological, like we just discussed. But actually, there's a lot of prestige to it within the domestic Chinese I would say, community.
that if you are a top student graduating from Tsinghua Beida, which is like the equivalent of the Ivy Leagues in the US, then joining a state-owned company or like a state department is extremely respected.
that like really deeply ingrained cultural value that has led a lot of top talent go into government. And it goes back to Dan Wong's book where, you know, most of the Chinese leaders are actually from top tier universities and engineers, STEM trained and very logical and rational, maybe lacking some of the more liberal arts training, but you know, there's that whole comparison. I guess this topic is quite broad, but I just want to kind of hear your thought on, do you see this trend phenomenon or?
Do you think this is just one of the things that Katherine Boyle said herself?
Jasmine Sun (51:13)
Yeah, I've been thinking about this topic a lot this year, so I'm actually glad you brought it up, because I do think one of the biggest changes in Silicon Valley culture that I've observed in the last, I guess, what did I get here, it's 2017, like last eight years, let's call it, has been a different orientation between Silicon Valley and the government, because the origins of technology and tech culture were very libertarian, right? So like in the 90s, it was all of this idea that the reason you have things like the
Declaration of the independence of cyberspace the reason that we go to the internet and we create our own communities there is so that we can transcend borders so that we can escape the law and so it's almost this like more It's this more libertarian attitude of like we can do payments that nobody can track with crypto we can ⁓ create communities that are totally different from national borders and in fact, there's a lot of skepticism of ever really like working with or collaborating with the government because like that's slow and that's old and like we're here to build something radical and new
And I think that that's over the last, you know, 10 years, there has been a recognition from Silicon Valley and big tech companies and VCs that has been affected sort of the general ecosystem that actually one is like, government is still very powerful and influential. It will probably like not be replaced by technology.
we will get hit by these fines and regulations that are going to be really challenging. And so I think Silicon Valley has recognized that institutions and political institutions are, probably not going away and be quite powerful and see that they then need to learn to collaborate and work with them. And so I think first of all, that's been sort of a broader shift even before the last couple years. But then I do think the last couple years have been an inflection point, sort of like Katherine Boyle has said. I did a reported piece earlier this year where I basically interviewed a ton of Stanford students, like undergrads, GSB business students, faculty, etc. Because I had heard from people I knew at Stanford, like my younger sister was at Stanford at the time, that it was really sexy to go into defense tech now. So working at Palantir, working at Andrel, working at SpaceX, these companies were now celebrated as like some of the most ideal places for the top engineering students to work. And when I was on campus at Stanford, the total opposite was the case actually. It was a mood of lot of skepticism towards working in defense tech, because it was associated with the war and violence and like that's bad and scary or like surveillance and things like that. And so it seemed to me, like the mood had really shifted. And so I started interviewing students to try to understand why this was the case. And there are a few reasons that I think there's been a resurgence in interest in working at defense tech companies and other companies that sort of are more government related. One is that the geopolitics simply have changed. Like, I think the conflict in Ukraine, I think the sense of a great power competition with China, I think also like Israel-Gaza war has actually just like made geopolitics more front and center in people's minds. Stanford students tend to be quite mission oriented. Silicon Valley people love a mission. People ideally don't want to just make money, but make money doing something that they feel is meaningful and that they feel really good about. I think that created a sense of mission and working either with the government or with ⁓ defense. For some people, that really resonated. There's also the fact that people like Katherine Boyle, Andreessen Horowitz are creating these theses.
A lot of Silicon Valley is just about orienting where the money is and that's where the opportunity is. Like one thing that I really thought about when I was in China was who makes the market. So like in China, it might be that the government says like, hey, like we think that AI is really important or we think that electric vehicles are really important. We're going to start putting subsidies. We're going to start putting money there. And that means a lot of ⁓ entrepreneurs will be like, let me go to that space so I can start a company there. I actually think that VCs basically play that role in Silicon Valley. So Katherine Boyle will then say - hey, we are now doing an American Dynamism Fund. We're gonna put in like $600 million or whatever it is towards startups that are working on things that will like sort of support the government and like the national security and things like that. And now that everyone knows, like all of these like 21 year olds at Stanford know that there is money available from one of the biggest VC firms in the Valley specifically for this thing, then everyone's definitely gonna be interested in that because they're like, okay, great. Like this is like what I should do. This is like exciting. This is where I'm gonna get funding. And so like, I do think that again, VCs just play such a big role in shaping the markets and deciding where people go. it's just like these people are quite young. whatever seems exciting, wherever the money is, wherever their ⁓ entrepreneur VC heroes are, that's what people are gonna follow as well. And then I think there are other things. I think that in the last six months, we've seen that Donald Trump is very interested in companies that are willing to work with the government. I think this is very different approach from previous US administrations, but Trump is...
pretty clearly like, gotta work with me very closely, you gotta say nice things about me or else I'm not gonna like let you thrive. And so then therefore tech companies even that might not want that are gonna be much more cooperative than they otherwise would be. I also think like Palantir and Anderil and SpaceX have just done really well. And like those three companies specifically kind of became the model for a really successful like defense tech company that does government contracting and is still a really valuable private business.
at the same time as they're doing these big government contracts. And so I think people feel really inspired by that. But I will finally say that like, it is definitely true that there is a resurgence in interest in working with government, but ⁓ that is not the majority. I would say the majority of folks in Silicon Valley and coming out of top schools would still probably work out like a open and anthropic, like more software type of thing, because it is still true that just like working in government contracting is like, slower and more challenging and just like, there's still less money in that space relative to say like AI broadly, even if it is something that is more interesting to people now.
Grace Shao (56:54)
Yeah, definitely. think, yeah, just on your point on Donald Trump really showing that he wants to work with tech leaders, it's interesting that we just saw that massive dinner quite a bit of shit for saying 12 thank yous in one sentence. But anyway, okay, moving on from that, I'm mindful of time, but I do want to touch on a few things in terms of the AI culture space.
There's a lot of fear around AI psychosis right now. just like you mentioned earlier, like the fear in the West about AI or technology distorting our sense of reality or causing more mental health and issues for our younger generation. How do you see that right now? who are the people who should be, like how do we regulate this? And who are the people who should actually be regulating this? And how do we regulate this? Because right now, when I talk to some AI software developers,
It's kind of crazy that they're so driven by the high level mission that they kind of, I don't want to say ignore, but they just don't really think about the maybe day to day consequences. And you know that I'm about to literally pop a baby out in two weeks. As a young mother, I'm actually quite fearful of how that might look like for my future children, because you and I grew up with social media. And I think it really did have a certain effect on our mental health as young women at that time. But we were lucky that actually we were already past teenage years or at least like old enough to at least the stuff were not fruitful for us or positive us. But I think for the generation that were actually born between 2005, 2010, they really kind of had to be the guinea pigs of growing up with social media and really living with the most negative consequences that could there could be.
Jasmine Sun (58:32)
Yeah, I've been thinking about this a lot. actually literally right before this, I was having a conversation with a writer and a friend of mine who actually experienced AI psychosis last year. And so he's basically thinking of writing up an essay on his experience, both going through AI psychosis and also sort of like what advocacy and policies that he feels are necessary based on that reflection. And it was interesting because one of the themes that we talked about was like, this friend of mine, like he has worked in tech, has worked in AI, he is someone who's like very tech literate and very educated about these things. And so you wouldn't think that someone like that can be susceptible to literally he had to be hospitalized like as a result of this experience, right? But like the thing that is actually happening is because these things are so personalized, like, and it's like, they know everything about you. So they know just how to flatter you the exact right amount about the exact right things. And if you don't have a lot of other social support, or you're already stressed for whatever reason. Like it can really, not like anyone or like a lot more people than you think, I think can like fall prey to these kinds of dynamics. And so I worry about it a lot as well. I think like, I wrote an essay on sycophancy a few months ago that was not quite about psychosis, but relates to the dynamic of the AIs that, because usually psychosis starts with the sycophancy. It starts with the AI that's like flattering you even when you're wrong, even when you shouldn't be flattered. And I think one of the really challenging dynamics there is like two things. So one is that, people want to be told we're right, right? Like most of us enjoy being told that we are correct, that we are smart, like that's a great question, blah, blah, blah. And like even though most people may not necessarily admit that out loud, if you are a PM, because like I used to be a PM, so I'm like, I know that they're looking at how many messages are people sending every day? How much time are people spending chatting with this? How many days of the week do they chat with us? And I am willing to bet money that... a more sycophantic, a slightly more flattering AI is going to do better on all those metrics because just like you are going to talk more to a friend who's like pretty warm and supportive, you're going to talk more to an AI that's like warm and supportive. And so there are all of these like product dynamics where just the fact that people tend to like nice AIs will ⁓ sway AIs towards being more sycophantic if you're just taking like a PM approach of like, want to maximize like messages sent, time spent, blah, blah, blah.
And then second, of course, like these are for-profit companies. And so even if the companies recognize that there are some risk, they are by default not going to be very incentivized to curb it because they know that as soon as they make the AI like more strict or less emotional or less warm, people are going to use it less. They're going to like it less, right? Like open AI, I think about how 4.0 was known to be the most ecophantic model and
usually led to psychosis in these cases, they took it away when they introduced GPD-5 and the users literally revolted. were like, GPD-5 is so cold and so unfriendly and like doesn't have a personality. And like people got so angry at them, they actually gave the users for or back, which is a really fascinating thing to see because you see how the dynamic of like give users what they want. Like we're a company, we gotta build products people like actually made ⁓ open AI have to like.
make the AI more sycophantic than they actually probably wanted it to be. So I think that's a very interesting dynamic as well. And like what that makes me think is I think, do you think you probably want policy action, right? Like I think they're that if you just leave it up to the free market of companies and their users, these are there, you're going to have a lot more cases of psychosis and dependence and addiction and these kinds of risks. And I'm not like a policymaker. So like, I don't know exactly what it is, but like, I do think that for example, like, there should probably be guidelines around like, if someone brings up suicide, like you can end the conversation and point someone to a suicide hotline. Or like one thing that AI companies have said is that after a certain number of turns, as the conversations get longer, then the AI gets less controllable. So you might be really good with guardrails in the first 20 messages. But once the conversation goes on too long, like maybe the guardrails like start like not working anymore. And my reaction to that is like, if you know that, you should probably stop, like cut off.
the limits past the conversation length that you feel able to enforce your guardrails. Like it should not be the AI will help you talk about suicide after 30 messages. Like if you know that you should cut it off at 30 messages, right? And like maybe there should be some policy action to try to enforce that. Or like, I think also transparency is a big part of this story as well because it's a very new technology, right? And every time there's a very new technology, I think actually just understanding the scope of the problem is probably what we need to start with. And like,
The companies themselves have all of this data about like, what are people talking about with AIs? Like how many people are experiencing these things? How many people are having mental health conversations versus just like doing like business stuff, right? And so like, I think that there are ways to both like analyze and share that data in privacy preserving ways. So like just looking at the aggregates, of course, not necessarily sharing specific conversations. And I think if there's a way that whether through policy or some other sort of action to motivate OpenAI and Anthrobic in the major labs to actually like, themselves collect the data and analyze how many of our users may be experiencing something like psychosis. How do you actually measure how sycophantic a model is? Because I know they have evals to measure this. And can they actually publish those every time they release a model? Can they actually share with external researchers some of this data so that is aggregated and anonymized? So you can understand, is it just a couple people who are experiencing this mental health issue, or is it actually very widespread? And so those are some of the things that I'm starting to think about.
Yeah, I sort of do think that there's almost because American AI safety has been so focused on these more speculative risks of like the AI is going to have its own motivations and it's going to like start assembling resources or lying to like strengthen itself or something like that. There's been less of a focus then on these like very immediate psychological risks, but that are literally like killing people. Like people have committed suicide as a result of AI psychosis and so.
I'm happy to see that there's media attention on it now, but my sense is it will require some policy action because the incentives otherwise, people just probably won't do it.
Grace Shao (1:04:27)
Yeah, I definitely think more transparency on that would be really, really, really helpful for people to even understand what they're getting themselves into when using these technology. you're saying, I think a lot of the high level kind of focuses on, oh, whether they're going to take our jobs and turn around against us instead of planning the apes and we plan on the robots. And then even the talks about the environment and energy. And obviously, it's extremely important to understand the detrimental effects of just how much data centers
how much power data centers use and how it's affecting the environment. Our things are so imminent and so in our face already where individuals are losing their lives because of this or being catfished for like, you saw that news. It was really sad actually, this elderly man, he got catfished essentially by a robot and thought he was gonna go meet up with her and thought it was this young, beautiful girl that wanted to be with him and then he passed out and died unfortunately on the way getting there. So there's.
Jasmine Sun (1:05:20)
gosh.
Grace Shao (1:05:21)
There's no guidelines on this right now, which is And I don't know what it will mean for the kids. But unfortunately, to your point about like capitalism running like, or like money sometimes driving a lot of these behaviors, it's that like, the companies themselves were not going to volunteer this information, unless being forced to. And I wish there were more regulations around that, on like, you know, immediate issues. Okay, so I don't want to end the conversation super, super doomy and like sad whose head is such a lovely conversation.
Jasmine Sun (1:05:41)
Yeah. Hahaha
Grace Shao (1:05:47)
⁓ have a few quick questions for you. think one is, you know, what is the biggest misconception you think the average layman have about AI and technology and the tech development right now? this can be really broad. could be culture, it could be anything, but not. It's not, I'm not talking about the average Silicon Valley product manager or software developer. I'm talking about like the average person you think you meet, whether in China or the U S what do people.
Jasmine Sun (1:06:06)
Mm. I think overall, I would not underestimate how much that AI will get better. And I think one thing that I think is the wrong way to look at AI is like you look at what it can do right now and you say like, okay, that means that AI is really good at coding and it's really bad at, I don't know, like writing pros or accounting or something like that. And like that may very well be true about AI right now, but like to me, what that also says is like,
So many resources have been put into making AI extremely good at coding and also all the people building AI right now, they're very familiar with what it looks like to be good at coding. And my personal sense is that like they can point to that amount of resources at a range of other problems, whether that is law, whether that is services, whether that is writing and AI will.
probably get significantly better on those axes in the future. And so as we think about the future jobs or the economy or whatever, we should basically expect a lot of improvement and not get frozen in a mindset where it's like the things that AI is good at today are the things that AI will be good at forever. So I think that's one of the things. It's just like even just, I know writing, so I think a lot about how good they are at writing. They're definitely not as good as good human writers, but from when ChachiBT first came out until now, it is a world of difference. And if you look at it side by side, everyone will be able to see that.
But in the moment, I feel like sometimes people will just say, it's so bad at this. And like, that may be true, but you want to look at the rate of improvement. so like, think like just thinking about the rate of improvement is like something I would recommend folks do.
Grace Shao (1:07:38)
Yeah, definitely. And also, think more more focus, we're seeing more and more focus on vertical AI software and development. Okay, so I have another question for you. Given you are one of the first employees, well, actually 25th of Substack what are some tips for people? Because a lot of our listeners will be on SubSac as well, as we were sharing this there. What are some tips or takeaways you have for people who are... writing on a newsletter, trying to grow their following, whether they're actually professional writers or thought leaders or whatnot.
Jasmine Sun (1:08:07)
Yeah, you don't definitely don't have to be a professional writer. That's like one of my favorite things about Substack right? Is like you can have any expertise, any passion, ⁓ you can do it on the side. And if you write really well, that can become either very meaningful hobby or it can become a job. And so like, that's one of my favorite things. In terms of tips, I think that for me, one of the most valuable things is getting to know and connect with other writers in a meaningful way. Like obviously like,
do all the things that Substack tells you to like turn on recommendations and like blah, blah, blah. Like that stuff definitely works. But I think the thing that is both personally fulfilling and has had a big growth impact for my Substack is literally just meeting people, collaborating with them, like befriending other writers in my space, right? And I think like part of that is like at first you maybe just like reply to each other on notes and you kind of become internet mutuals. You can also like email a writer who you really admire and either send them a link to one of your posts in case you think it's relevant or you can just be like, hey, like I'm a...
early writer, I'm trying to build my knowledge, like, will you get on a 15 minute call with me? Like, I remember when I was still literally in college, Azim Azhar, like, and I got like a 45 minute coffee at like this London Soho house or something, or like Casey Newton made time for like a 20 minute call with me in like 2019 or 2018, when I was like much, much younger. And those I still remember, like the generosity that they expressed, and they gave me genuinely very good advice. And in the future, like,
Those people can either become friends, they can become mentors, they may be people who share your posts and then just suddenly one day like drive a ton of new subscribers. Like lot of the times you just get new subscribers because someone bigger than you recommends your work or like shouts you out. And like, think just like trying to connect with people in an authentic way, a lot of big writers are big readers. They love discovering smaller writers. They love helping out. And so as long as you're like respectful about it and you're not like, give me a recommendation right now. Like I think that's can be a very meaningful thing to do.
Grace Shao (1:09:54)
Yeah, I've been so pleasantly surprised by the community because like, I honestly, as much as like I mentioned, I'm outgoing and everything, I actually hate doing self promo. It's so cringe and like, but we all have to do it little bit. Right. And like, I feel like I just hate being that person who's on LinkedIn 24 seven, be like, my think SubSac has been amazing where obviously just been writing kind of with my head down and you'd be surprised people who are interested in your topics or niche will reach out. And like you are so generous with their time and they're willing to jump on up.
Jasmine Sun (1:10:01)
Yeah, totally.
Grace Shao (1:10:20)
I'll just chat through some ideas or provide some advice or even collaborate. for sure, subsets been amazing on that. And I think it's just such a different mentality writing on subsets compared to writing for like a traditional publication where you need to remove yourself. If anything, I've been pushing myself, bring myself and my personal life more and more into it because people want personality. They want the they think they, and I also think that like they want to know your own, who you are because
who you are really does shape your views and explains your views or how you're getting information and even your credibility, right? So that's super cool. Last question, but it's a question we ask everyone on the guest. So what is a belief that you hold that is unconventional or against consensus? I wouldn't say this is your differentiated understanding, but this could be something just a bit maybe different from people around you. could be anything about your work. It could be something about your personal life. And I sometimes joke it could be even about you believing the world is flat.
Jasmine Sun (1:11:21)
this one's so hard. It's hard because I think that the beliefs that I have that are weird for Silicon Valley are very different than the beliefs that I have that are weird for the world, for example, right? I think it's just like, the world is so different and man. Yeah, I'm very unprepared for this one.
I mean, I think relative to Americans broadly, I really like using AI. think it makes me, and also relative to writers, would say, relative to both the American public and relative to most of my writer peers. I enjoy using AI. I think it helps me a tremendous amount. I can feel that I am a faster and actually, I would say, a better writer by using AI. I'm certainly not having it out paragraphs for me. I don't think that is a good approach based on where the AI currently is at. But I think a lot of writers would benefit
And a lot of people normally would benefit from playing more with LLMs and trying to see how it can teach you new subjects. Like I love like trying to read a book or like a paper that's like a little bit too hard and then just like asking the AI like explain it like I'm really stupid. And like, this is so helpful for me or like I have all these like shortcuts and tools to like do research in different ways and like.
⁓ help me edit transcripts. like, I just do really feel sometimes that if people are too scared of using LLMs, they're missing out, which is again, not to say that you don't want you want to outsource everything to them. You have to like really deeply recognize where they help them or they don't. But yeah, most writers, Substack, journalists, etc. I feel like are very LLM skeptical and I just like have gotten a ton out of them and I don't think my ratings any worse for it.
Grace Shao (1:12:53)
actually kind of hold the same view. think, like, equally like a lot of writer or journalist peers are like really skeptical. They're like, oh, they're such crappy writers, blah, blah, blah. But I'm like, actually, I don't really use it for the writing bit, but it's more, sometimes it's like you said, like, transcripts or even when I'm working with different languages with like ease of translation, research, like preliminary research where they can guide you to certain sources. It's actually really, really efficient, especially when you're working on like alone on a piece. It's really great to kind of get like,
Jasmine Sun (1:13:05)
No. Yeah.
Grace Shao (1:13:20)
a sense check if you're missing something, if you're not thinking about a topic holistically or 360. And I think actually I also found it quite useful. I don't spit out paragraphs with it, but a little grammar check here and there or a typo check is also really good. But cool, thank you so much Jasmine. We've actually did a long interview today. I really, really enjoyed our conversation and I felt like I got to know you better as a professional writer, but also as a friend.
I'm really glad you kind of had the time to share with us today with your insights and everything. Is there anything else you want to touch on that we have not touched on?
Jasmine Sun (1:13:53)
I don't think so, but I'm so glad you're doing this series and yeah, very honored to be invited on. And yeah, I'm so glad also just like we got to meet, but like in Hong Kong in person, which is again, one of the, was thinking about that during the like part about, you know, making friends on Substack. Like one of the things that I love most now is that when I travel, there's almost always like a couple Substack writers who like, I may not have ever like met before, but who are down to show me around, tell me about their lives, tell me about their cultures. And like, I just feel really grateful for that. So.
Very happy to have had this conversation and thanks for your time, Grace.
