Whose Golden Age Is the AI Era? | Chatting with Zhang Zala: Liberal Arts Majors, Taking Action, the Patterns of Viral Hits, and How Ordinary People Can Win

It's your golden age, my golden age, everyone's golden age.

👦🏻 Podcast interview: Koji

🥷 Edited by: Crossing

🧑‍🎨 Layout: NCon

🚥 For the first episode of "Crossing" in the new year, our guest is Zara Zhang.

Over the past year, she grew from 20,000 to 180,000 followers on Xiaohongshu, published hundreds of posts, completed a career transition from non-technical background to AI product manager, and launched her first AI product as a side project — LongCut[1].

In this episode, we discussed how humanities majors can find their place in the AI era, how to start creating content from scratch, how to discover viral patterns, how to achieve growth with "relaxed energy," and her 2025 AI product recommendations and usage tips.

I also talked with Zara about her insights on AI-era marketing trends: when products are abundant and attention is scarce, "human-ness" becomes the key to standing out; founders are becoming the best CMOs, and "marketing for all" is no longer just a slogan but a strategy; pop-up cases like Cursor Cafe and Anthropic Thinking Cap are making AI product marketing increasingly resemble consumer goods.

One thread running through the entire conversation: you don't need anyone's permission. Want to build a product? Open Cursor and start building. Want to create content? Pick up your phone and start recording. In the AI era, code and media are two levers that don't require anyone's permission — and now, AI has lowered the barrier to both to unprecedented lows.

Listen on WeChat:

Listen on Xiaoyuzhou:

🎬 The video podcast is also available on @Koji's WeChat Channels, Xiaohongshu, Bilibili, YouTube, and other platforms.

Since the full interview is quite long (24,141 words), you may refer to the table of contents first:

🟢 2025: A Year of Action

Watching videos and studying is actually a form of laziness.

🟢 AI Makes Building Products Cheaper, But Marketing Harder

Code used to be the scarce resource; now attention is the scarcer one.

🟢 The Future of Marketing: Human-ness

Social media is flooded with AI-generated content that clearly reads as AI-written — deeply inauthentic.

🟢 Marketing for All: The New Trend in Silicon Valley

🟢 Xiaohongshu: From 20,000 to 180,000

People see the follower count, but what's more noteworthy: I published nearly 500 posts in total.

🟢 Follow Builders, Not Influencers

I can have free video calls with the top practitioners in Silicon Valley.

🟢 The Best Time to Be a Humanities Major

When a new technology wave arrives, everything resets to zero — everyone starts from the same starting line.

🟢 Idea is Expensive Again

🟢 Go-to-Market Case Studies: AI Company Marketing Increasingly Resembles Consumer Goods

The highest level of marketing: you're not promoting yourself, you're promoting your users.

🟢 Poke: Turning Onboarding into a Negotiation Game

🟢 LongCut.ai[2]: My YouTube Learning Tool

I really dislike summaries — once something becomes a summary, the original essence is lost.

🟢 What Building Products and Creating Content Have in Common

🟢 Levers That Don't Require Anyone's Permission

Ask yourself: is anyone actually stopping you? No one is.

🟢 Learn in Public: Why Public Learning Matters

🟢 Growth with Relaxed Energy

Many people ask how I have so much time. I ask back: how do you have time for Douyin, gaming, and watching shows?

🟢 The Long Game and Karma Theory

Every follower I gain today is marketing cost I save in the future.

🟢 AI Product Recommendations of the Year

• Claude
• NotebookLM
• Replit
• DesignFaces

🟢 AI Usage Tip: Let AI Ask You Questions

🟢 Spiritual Mentors of the Year

• Chris Pedregal (Founder of Granola)
• Josh Woodward (Google VP)
• Lulu Cheng Meservey (Silicon Valley PR Guru)

🟢 Introducing Yourself with 10 "I Ams"

……The remaining 5 are even more fascinating — keep reading

2025: A Year of Action

👦🏻 Koji

I'm very happy today to have Zara with us. Say hello to everyone.

👩🏻 Zara

Hello everyone, I'm Zara.

👦🏻 Koji

Zara's Xiaohongshu is like a treasure trove, with massive output and incredibly rich content. Today I really hope you can share your reflections on a year of creating content and building products. We're recording this podcast at year-end, so my first question is: if you had to pick one word to summarize your past year, what would it be?

👩🏻 Zara

2025 has been a year of action for me. The year before last, I started learning about AI and realized this was the era-defining opportunity I needed to seize, but at the time I mostly just watched videos. Later I discovered that watching videos and studying is actually a form of laziness — you study so much because you're afraid to take action.

So what I'm proud of this year is that I truly took action, both in content creation and in building products. I found that the best way to learn is learn by doing. Watching more videos, listening to more podcasts, hearing more founder interviews — none of it compares to actually getting your hands dirty.

👦🏻 Koji

Speaking of which, should everyone just turn off this podcast right now and go take action? Haha!

👩🏻 Zara

Haha, actually you should still listen to podcasts — just change the order. Many people stop at the "watching" stage and never move forward. I found that just reading AI news, studying the latest papers, and watching interviews is useless; you have to actually start building, whether products or content.

This year I broke through in both Chinese and English content creation, switched roles to become a product manager, launched my first side project, gathered user feedback, started Vibe Coding, and turned ideas into reality. What I learned through taking action far exceeded what I gained from reading books and watching videos before.

And after taking action, when you revisit those videos with specific questions in mind, you'll find: oh, now I finally understand what they were talking about.

👦🏻 Koji

Hearing you say "this year is a year of action," I feel like you could literally be the spokesperson for "Crossing." Our opening line for every episode is: to seek out, interview, and bring together a new generation of entrepreneurs and active builders in the AI era.

AI Makes Building Products Cheaper, But Marketing Harder

👦🏻 Koji

Zara, have you noticed any new trends this year that others might not have paid much attention to?

👩🏻 Zara

People talk more about technical and model breakthroughs, but I've been paying closer attention to go-to-market and marketing trends. One significant change in the AI era is that AI coding tools have dramatically reduced development costs and accelerated speed, but the consequence is that marketing and distribution have become much harder. Many people haven't realized this is a massive trend.

The first trend is that because AI has lowered development costs, products have become extremely abundant.

👦🏻 Koji

Extremely abundant.

👩🏻 Zara

There's an overwhelming number of products. We're bombarded with various AI products every day — too many to even try.

👦🏻 Koji

Because the cost of building a product has dropped, whether for design or coding.

👩🏻 Zara

Right. Many people are starting to build products, and the number of startups has exploded. In this environment, capturing users' attention and getting them to use your product instead of someone else's has become extremely difficult. We used to think of code as a scarce resource because R&D costs were so high; now I think attention is the scarcer resource.

👦🏻 Koji

What changes does this bring?

👩🏻 Zara

The first change is that marketing, distribution, and brand are becoming moats. Previously people believed product and technology were the moats — you'd build the product first and only then think about how to launch it, with distribution as an afterthought.

👦🏻 Koji

Before people would say everything is ready except the CTO. But now it feels like the CMO has become more important.

👩🏻 Zara

Storytelling ability, marketing, and building user communities — these have all become critically important.

Additionally, AI has brought another shift: social media is now flooded with AI-generated content. You can tell at a glance it was written by AI — it feels deeply inauthentic. Using AI to assist your creative process is fine, but if what you produce lacks genuine feeling, original perspective, or even the basic effort to sound human, that's a problem.

The Future of Marketing: Human Presence

👩🏻 Zara

So in this environment, what kind of content actually stands out? Content with human presence.

So the key term for future marketing is "human presence."

👦🏻 Koji

Human presence?

👩🏻 Zara

Human presence means I need to feel that there's a real, living person behind your product or content, and that I can connect with that person. Podcasts, for example, are an incredibly human medium — when people listen to our voices, they immediately know we're not AI.

👦🏻 Koji

Hahaha, because our voices still have some imperfections.

👩🏻 Zara

Or take appearing on camera — that's why I often like recording videos of myself. It's the fastest way, and also because I want people to feel there's a real person behind it, someone they can connect with.

👦🏻 Koji

Actually, OpenAI has very strong human presence. I can barely think of anything influential ever posted from OpenAI's official account. Most of the time it's Sam Altman himself constantly posting on Twitter.

👩🏻 Zara

Right. In the AI era, personal brands and product brands will increasingly become indistinguishable. Founders have to be their own spokespeople; founders are the CMO. We're seeing this trend quite broadly. For instance, when you learned about OpenAI's new features, it was definitely from Sam Altman's Twitter, not OpenAI's official account. People prefer following people, not brands.

👦🏻 Koji

And OpenAI's product launches — they're not held in some grand auditorium, elevated and distant. They're like this, us sitting here, intimate and energetic.

Everyone Does Marketing: The New Silicon Valley Trend

👩🏻 Zara

So if you're going to run an account, rather than using an official brand account, you're better off investing in the founder's personal account. Or even getting other employees out there to become part of marketing.

I'm seeing this as a trend in Silicon Valley: everyone does marketing. Marketing isn't just the marketing department's job.

👦🏻 Koji

Everyone's a CMO.

👩🏻 Zara

Exactly. Say a feature was built by a particular engineer or PM — they should be the ones posting on Twitter to promote it. This creates direct, close connection with users, and the feedback and complaints in the replies help with product iteration.

👦🏻 Koji

I personally love the browser Arc. Their weekly changelog has such human presence. Most companies would say "we updated 10 features, one two three four five six seven eight nine ten." Arc expands on each feature, even with images — doodle-style. And they'll put the name of the programmer who built that feature right there.

👩🏻 Zara

Put their profile picture right on it.

👦🏻 Koji

Yes, and you develop this incredibly strong sense of connection and community with the product.

👩🏻 Zara

It also makes employees prouder, more fulfilled — they feel "I built this." Small companies have a huge advantage here. We often talk about building in public, which is very hard for large companies to pull off because of all the PR policies preventing people from speaking externally.

But for small companies, this is a massive advantage — you can have everyone doing marketing. Small companies necessarily trust their employees more; if you don't trust them, you probably hired the wrong people. Everyone at the company should feel ownership of their work and be able to go out and explain it clearly.

👦🏻 Koji

You build a product, and if you don't talk about it, nobody knows at all — that's the real tragedy.

👩🏻 Zara

Right, the scariest thing is nobody cares. The default state is nobody cares.

👦🏻 Koji

Can you share more examples of startups where everyone's acting as CMO?

👩🏻 Zara

For example, Lovable's head of growth shared that they put heavy emphasis on "Employee Social."

👦🏻 Koji

Employee Social?

👩🏻 Zara

Not just Founder Social, but Employee Social. It's not only the founder posting constantly — employees are encouraged to post too. Every time a new feature ships, the whole team retweets and builds momentum.

👦🏻 Koji

That reminds me of Jumei International Holding Limited. When we did big sales events, every time, the entire company would change their WeChat profile pictures to "301" or "901" — our anniversary sale icons. Even back then we had this awareness of everyone doing social.

👩🏻 Zara

And I've seen many overseas product managers or operators directly collecting user feedback on Twitter, or before shipping a feature, going out to ask if people even want it.

👦🏻 Koji

Overseas it's Twitter; domestically it's Xiaohongshu.

👩🏻 Zara

Right.

From 20,000 to 180,000: The Truth Behind Xiaohongshu Growth

👦🏻 Koji

You've been very active on Xiaohongshu. Have a lot of people drawn inspiration from your account and started wanting to post themselves, to build in public or learn in public? How does someone find their positioning on Xiaohongshu?

👩🏻 Zara

At the beginning of this year I had about 20,000 followers on Xiaohongshu; now I'm at 180,000. People see that number and think it's impressive, but here's another number: I've posted nearly 500 pieces in total.

👦🏻 Koji

500 pieces!?

👩🏻 Zara

Since starting on Xiaohongshu, nearly 500 posts.

👦🏻 Koji

That's an average of 1.5 per day.

👩🏻 Zara

More like one every 1.5 days, over roughly two years. There's no need to set a goal from day one about how many followers or how much traffic you want. The only goal I set for myself at the start was "post daily for six months straight." I posted every single day for six months without missing one. I think that's the more meaningful number.

👦🏻 Koji

Looking back at your earliest content now, how do you feel?

👩🏻 Zara

I don't think I had a clear positioning early on. I've realized that positioning isn't something you think up — it's something you stumble into.

👦🏻 Koji

Stumble into?

👩🏻 Zara

Yes. A lot of people ask how to find their positioning, and I can't tell them. Stop overthinking it and just start posting. After dozens or a hundred posts, you'll develop a sense through feedback about what people want to see. Post a few hundred times, and your positioning naturally becomes clear.

Besides, positioning isn't what you decide — it's who the audience thinks you are, and that's your positioning. My eventual positioning ended up quite different from what I initially imagined.

👦🏻 Koji

So how would you describe your positioning now?

👩🏻 Zara

On Xiaohongshu, my positioning leans toward sharing AI trend insights and personal learning growth from a non-technical background perspective.

👦🏻 Koji

That's very clear. Can you share some "failure" stories to help people get over their fear? Like content you posted that seems unwatchable now, but gradually got better?

👩🏻 Zara

There were a few types of content that surprised me by finding an audience. Things I considered completely ordinary that just blew up.

One category was "how humanities majors can get into AI," or how to learn AI, how to participate in this wave. I hadn't expected so many people to care, but I later realized this might be my unique perspective in this conversation. Because the vast majority of people talking about AI come from technical backgrounds — researchers, engineers, or technical founders. Very few speak from a non-technical perspective.

But I realized that the vast majority of people in the world are non-technical, so they need this perspective. My Xiaohongshu growth has mainly been about riding the wave of the times — lots of people outside AI want to know how to catch this wave, how to enter this field and learn. So you have to tap into the trends of the era.

👦🏻 Koji

So one thing you didn't expect to blow up was "how humanities majors can learn AI." Any others?

👩🏻 Zara

Another category was resource recommendations.

👦🏻 Koji

Resource recommendations?

👩🏻 Zara

Yes, I normally watch long interviews on YouTube and assumed everyone did too. Turns out nobody does. Or rather, everyone agrees primary sources matter, but they don't know where to find them, or haven't even realized the information they're consuming is already second- or third-hand when higher-quality original sources exist.

I thought I'd need to write beautifully crafted summaries with deep insights. Turns out, I could just screenshot the five YouTube videos I watched and that was enough.

👦🏻 Koji

Hahaha, sounds like a pretty simple operation.

👩🏻 Zara

That's also a form of "human presence." Who you are, what you're talking about, why you think it's good.

Follow Builders, Not Influencers

👦🏻 Koji

But I've seen others post similar recommendations, like "5 great podcasts," and they usually get no traction. How did you make it work?

👩🏻 Zara

It fundamentally comes down to what content you're recommending. I have my own filtering criteria for finding content. For example, I have one principle: Follow builders, not influencers.

👦🏻 Koji

Can you expand on that?

👩🏻 Zara

Before last year, when I was just starting to understand AI trends, I didn't know much either, so I followed a lot of influencers. Like some marketing accounts on Twitter — someone releases a new model, they explain it, but with no real perspective or insight. Later I realized the highest-quality information all comes from actual practitioners in the industry. People working at AI research labs like OpenAI or Anthropic, or founders and team members at startups like Cursor. And these people are incredibly generous about sharing — there's this open culture where they frequently post on Twitter, go on various podcasts to talk, and the most substantive resources are all completely free and public.

So I developed a habit of watching YouTube long-form videos from start to finish. I found it incredibly helpful. Some creators pack their content with insights, not a single wasted sentence, and they're practitioners I deeply respect in the industry.

Watching the whole thing feels like having a video call with them. I'm basically getting free video calls with the top people in Silicon Valley. A lot of people never considered this could be such a good learning method. That's my approach to finding videos.

👦🏻 Koji

And today, while we're recording this podcast, Xiaoyuzhou happened to release its annual summary report. I noticed that for Crossing's most popular episode this year, the top episode for comments, saves, and shares was the same one.

Our guest was Xiaopai Liu, who's also a builder. We found him because a third-party agency published a Claude Code token consumption ranking, and it turned out he was the guy who had quietly pushed Claude Code API usage to number one globally.

So when he came on to share his building experience, that episode 《揭秘 Claude Code 榜一大哥：一个 AI 创业者如何把工具用到极致｜对话刘小排》 got a ton of attention — which aligns perfectly with what you're describing.

👩🏻 Zara

Because I've found that builders speak from hands-on experience, while influencers are often just parroting what others have said, or simply telling you "OpenAI released a new model, Google released a new model." So what? That doesn't fundamentally change how you think.

Plus many influencers take sponsorships and ads, so the products they promote aren't necessarily ones they've actually used. But when builders talk about something, it's not how they make money, so it's more genuine.

👦🏻 Koji

I've always believed content should be "fun, useful, relatable." The humanities-major topic you mentioned is relatable, and resource recommendations are useful. Do you also make fun content on Xiaohongshu?

👩🏻 Zara

Some of my Vibe Coding posts are probably pretty fun, but not funny-fun — more like people going "oh, you can do it like that" or "that's a pretty cool idea." It mostly helps people see that technology isn't so mysterious.

Vibe Coding Case Study: Using AI to Make Recording Videos Less Awkward

👦🏻 Koji

Can you give an example?

👩🏻 Zara

I built a Vibe Coding product on Xiaohongshu and Twitter — a video recording tool. I noticed a lot of people feel very unnatural and awkward speaking on camera. Humans are actually pretty good at two-way communication; the most comfortable state is dialogue, like what we're doing now. When you're just monologuing at a camera, it's hard to be natural.

So I wondered: could AI turn one-way output into two-way communication?

I had AI play the role of a podcast host. Since Gemini has real-time voice capabilities, it asks follow-up questions based on what you're saying, making sure you can keep talking. It also has a fun feature called "one-click remove silence" — when the AI is speaking and you're not, that part gets automatically cut out, exporting a clean video with just your talking.

👦🏻 Koji

Wow, sounds amazing!

👩🏻 Zara

Another variant simulates a livestream interface. I realized I'm super natural during livestreams, very "alive," because I'm interacting with viewers in real time, talking about what interests them. So I made an "AI hype squad" — while you're recording, there's a bunch of AI-simulated viewers beside you, asking questions in real time, saying "you're amazing," giving positive feedback. This also lowers the barrier to recording videos.

👦🏻 Koji

That's such a great insight. When a person is speaking and getting no response, as a social animal, you just freeze up, get awkward, feel bored.

👩🏻 Zara

Exactly. Whether it's one-on-one with a person or simulating a livestream, the human brain is pretty hackable — but this is a benign hack.

👦🏻 Koji

Feedback is so important.

👩🏻 Zara

I became an AI Native person this year, but mainly because people outside perceived me as one, so I started seeing myself that way too.

👦🏻 Koji

How does that work? A bit abstract.

👩🏻 Zara

It's also a hack. Last year I actually felt AI was far from me, but I started learning in public, sharing my learnings and experiments externally. Because of my sharing, people started seeing me as AI Native, and that external perception reshaped my self-identity in return.

If you want to change your positioning, break free from past constraints, one hack is to deliberately create an impression outwardly. People tend to define themselves by how others evaluate them.

👦🏻 Koji

I remember getting carsick as a kid. My aunt taught me a trick: when you feel carsick, pinch the web between your thumb and index finger, and tell yourself "I'm the best, I'm not going to throw up today, I'm not a carsick person." That trick helped me overcome carsickness — eventually, I genuinely stopped getting carsick. The power of suggestion is incredibly strong.

👩🏻 Zara

Yes. I think in the AI era, people can shed all the labels they've put on themselves. I'm someone who doesn't like being defined. When I did content before, people wondered if I wanted to go into marketing; when I did VC, people said maybe I always wanted to do VC.

But I actually switch directions every few years — I don't want to do the same thing forever. Including my Xiaohongshu, there's no niche, because I think life is rich.

Previously people divided everyone into two categories: humanities and STEM, technical and non-technical. Studied CS in college and can code? You're technical. Didn't study it? You're non-technical. I think this binary is outdated in the AI era.

The Best Era for Humanities Majors

👦🏻 Koji

Right on this topic. You said many people follow you to see how humanities majors can learn AI in this era, but you also mentioned the boundary between humanities and STEM is blurring. I completely agree — I oppose any form of binary thinking; it boxes us in.

So I want to ask: as someone without a STEM background, how do you blur that boundary, create value in the AI era, and become AI Native?

👩🏻 Zara

"Non-technical" isn't a great label, but I've recently come to like a term: "technically curious" — having curiosity about and willingness to understand new technology. I find this has nothing to do with whether you studied CS. I know plenty of strong engineers who don't care about the latest AI developments at all.

Meanwhile, I'm not from a technical background, but I'm super enthusiastic — I'm on Twitter every day looking at new models, new APIs, and I try them as soon as they come out.

So this isn't just about ability, it's more about mindset and willingness.

When a new technology wave arrives, everything resets to zero — everyone starts from the same starting line.

Even if you majored in AI for undergrad, you have to relearn now. So there's no difference between humanities and STEM majors — everyone has to relearn. The key is what to learn, and finding your personal positioning in this era.

👦🏻 Koji

You said earlier that you think this is "the best era for humanities majors"?

👩🏻 Zara

First, because everyone's on a new starting line. Second, "technical curiosity" is an acquirable skill — it's not determined by innate qualities or past labels.

👦🏻 Koji

In Chinese, that would be "keji haoqi zhe" — tech-curious people.

👩🏻 Zara

Right, the curious and the doers. Even if you couldn't code before, now you can use natural language with AI coding tools to bring your ideas to life — even surpassing many programmers at the idea and distribution level.

Because humanities majors should inherently have deeper insight into human nature and users, and stronger communication and distribution skills.

I mentioned earlier that distribution and marketing will become a moat — that's one reason it's a good era for humanities majors. I heard something recently: idea is expensive again.

Idea is Expensive Again

👦🏻 Koji

People used to say idea is cheap. Since Vibe Coding has become so easy today, theoretically ideas should be easier to execute — so why have they become expensive again?

👩🏻 Zara

I think they're both expensive and cheap.

"Expensive" means a good idea comes from unique insight into a specific group of people. You can't get this from just doing a few interviews — it often comes from being that kind of person yourself, or having extremely deep understanding of a specific scenario. Take my video recording product: if you haven't done Xiaohongshu livestreams, haven't posted hundreds of videos, you'd never think of it. It comes from massive practice and feedback.

"Idea is cheap" because implementation costs and copying costs have both dropped dramatically. Having just an idea isn't enough — you have to execute, acquire users, iterate based on feedback, and generate new ideas. The faster this cycle spins and the more iterations you get, the stronger your competitive edge.

👦🏻 Koji

Since it's the best era for humanities majors, what roles can they play?

👩🏻 Zara

A year or two ago, as a non-technical practitioner in the internet industry, I'd sometimes feel imposter syndrome. I used to feel the industry was tech-dominated, especially when AI first emerged — all the discourse power was with researchers and engineers, and I felt left out, unable to participate. Later I realized technology only creates value when it becomes a product, gets released to users, and gets used. Technology alone isn't enough.

👦🏻 Koji

At what moment did that imposter syndrome disappear?

👩🏻 Zara

We used to say "talk is cheap, show me the code." I posted a tweet the other day: "code is cheap, show me the talk."

👦🏻 Koji

Hahaha!

👩🏻 Zara

Coding used to feel expensive and mysterious. You used to need to hire a lot of engineers; now you can build something in days.

In this environment, two questions become critical. First: what to build? Humanities majors are better equipped to answer this — they have deeper understanding of specific communities, sharper user insights, and the product thinking to translate those insights into products. You can even prototype directly with Vibe Coding tools.

Second: how to distribute? How do you help users understand the value? How do you package it into a compelling story? How do you build community, so users want to co-create and give feedback? "Head of Storytelling" has become a hot hire in Silicon Valley lately. Humanities majors are in demand.

Many Chinese founders have exceptional product and engineering chops, but they often ask me about Go-to-market. In fierce competition, branding and storytelling tend to be their weak spots. This is where humanities majors add value — telling the product's story well.

👦🏻 Koji

When you said "code is cheap," I flashed to oracle bones carved on turtle shells. Back then, carving characters into shells was something only elite engineers could do. That was "code is expensive."

Later came printing, then computers. Everyone could write, could type. It was no longer a scarce skill. Thousands of years ago, only a tiny handful knew how to write. Today's programmers are like those ancient scribes. It's kind of fascinating.

Go-to-Market Case Studies: AI Marketing Is Starting to Look Like Consumer Goods

👦🏻 Koji

You mentioned that finding insights and going to market both matter now. Have you seen any go-to-market cases that stuck with you?

👩🏻 Zara

A few interesting ones. First, I've noticed Silicon Valley AI companies are suddenly doing offline pop-ups.

👦🏻 Koji

Kind of like consumer brands?

👩🏻 Zara

For example, Cursor rented out a San Francisco café for a day — called it Cafe Cursor — and invited users to come code for the day. Core team members were there all day answering questions, just hanging out and talking with people.

Anthropic also did a hugely successful pop-up in New York. They gave away a hat with just one word on it: "thinking" — a thinking cap. People lined up for hours; some even flew in for it. The case blew up on social media.

👦🏻 Koji

I remember the context — the idea was that AI-generated content is mostly slop now, so they wanted to encourage independent thinking. Put on this "thinking cap."

👩🏻 Zara

Right. Wearing the "Thinking Cap" means you're an independent thinker. Everyone wants to be that person. It's clever: first, the cost is basically zero. Second, it's not saying Claude is amazing — it's saying "people who use Claude are cool."

This is a trend. Many of my non-technical friends, after using Vibe Coding to build something, their first instinct is to post on Xiaohongshu. They feel like "this makes me look cool, I'm someone who can code, who uses Cursor." I want to be that person, so I use this product.

Eventually, using the product becomes an identity. The highest form of marketing isn't promoting yourself — it's promoting your users. Using your product becomes a status symbol.

It's a bit like buying a luxury bag. Not necessarily because it's practical, but because of what you want others to think of you. From this angle, AI product marketing is increasingly like consumer goods.

Essentially, AI products are hard to differentiate on features now. There are at least ten AI coding products; ordinary people can't explain the functional differences, but they can feel the vibe. As a casual user, I won't spend time comparing. I'll pick whichever people are talking about, whichever vibe fits me, or whichever website looks better.

That's the nature of consumer goods. What distinguishes them usually isn't features — it's brand, design, storytelling, and overall feel.

👦🏻 Koji

Exactly. Like shampoo — incredibly homogenous, basically selling liquid. Mineral water is the extreme case. Why do people drink Evian, drink VOSS? They're drinking an identity.

I have a thesis: right now is an incredible arbitrage window for tech companies.

To consumer marketers, Cursor Cafe or the Anthropic pop-up are fairly basic campaign formats. But in tech, people go "you can do that?" and it instantly becomes a global sensation.

Of course this window may not last long. Once attention is activated, you need to keep delivering stronger stimuli. But at least for now — the next three to six months — doing a clever pop-up is a fantastic marketing lever.

👩🏻 Zara

Offline matters more now for another reason: "alive-ness." Offline is real people. You can see the creators behind the product face-to-face, chat with them directly, give feedback. That's good for both users and builders.

Traditional offline events reach few people. But the beauty of pop-ups is that everyone who comes posts on Twitter, posts on Moments — the experience itself becomes shareable content. A good product should have many built-in shareable moments.

The experience itself has to be interesting. First it's an interesting experience; the product pitch comes second.

Poke: Turning Onboarding Into a Negotiation Game

👩🏻 Zara

Another example — there was a buzzy product in the Bay Area called Poke. It's an assistant that connects to your email and calendar, and the big thing is you interact with it via text. It has a completely wild onboarding experience.

👦🏻 Koji

What kind of onboarding?

👩🏻 Zara

The onboarding is you texting with the AI, and you have to convince it to let you into its beta program.

👦🏻 Koji

Not everyone gets in?

👩🏻 Zara

Not everyone. The convincing process is a back-and-forth negotiation.

👦🏻 Koji

Negotiation?

👩🏻 Zara

And you haggle — there's no fixed price. First the AI challenges you: "Why do you want to use us? What's so special about you?" It doesn't flatter you like ChatGPT. It actually challenges you, even "roasts" you a little: "Who are you? You think you can use us?" That personality is striking.

And the price is genuinely flexible — whatever you negotiate is what you pay.

👦🏻 Koji

I remember you shared on Xiaohongshu that it initially quoted you $996.

👩🏻 Zara

Because I had done a podcast called 996, so it said "therefore you must pay 996 dollars."

👦🏻 Koji

One reason it creates an aha moment is that during negotiation, it already has all your personal context.

👩🏻 Zara

Right, because it read your Gmail. Privacy concerns aside, when I saw it knew where I used to work, what I normally do — I was shocked.

👦🏻 Koji

And you screenshot your whole haggling process and shared it. I saw lots of people doing the same.

👩🏻 Zara

That's also great marketing, but fundamentally it's a product decision. This isn't a marketing thing — it's making the product itself shareable and fun enough that people can't help screenshotting while using it. The best marketing is still the product experience itself.

👦🏻 Koji

Exactly. Talking about pop-ups reminded me — when I was CEO of Tangdao, we did an annual pop-up in Shanghai to celebrate the birthday of our flagship product, the "cat belly pillow." One year, we built a temple on Yuyuan Road. Passersby could go in and pray. Why pray? Because our cat belly pillow's mascot is a cat — the "meow god" who protects your sleep.

👩🏻 Zara

The sleep god.

👦🏻 Koji

So you pray, and you sleep well that night. Like you said, maybe only a few thousand or twenty thousand people came offline. But if the experience is fun, photogenic, makes you seem like "someone with an interesting soul," then the Xiaohongshu sharing rate is very high. The ultimate reach far exceeds those few thousand offline.

So if anyone wants to do a pop-up today, I'd love to brainstorm together — figuring out how to take your product to market and connect with users offline is genuinely joyful.

We talked about everyone being a CMO. In the process of taking product to market, besides what we've discussed, are there other small techniques you find especially important?

👩🏻 Zara

One crucial principle of marketing and communication: repeat, repeat, repeat.

Often people feel they've said something so many times they stop. But your audience has barely heard or absorbed it. When you're about to vomit from repeating yourself, that's when people are just starting to listen, just starting to notice. Because communication has loss. You think you said 100; they maybe received 50. Half probably didn't even notice.

For example, posting on Xiaohongshu — if you keep posting similar topics, you might worry followers will complain about repetition. But 99% of people see you through homepage recommendations; no one goes to your profile to scroll through everything. People consistently overestimate how much others pay attention to them. Modern attention is incredibly fragmented. People don't read.

You write a long article — most people will just have AI generate a summary. If they even read the summary, that's good. You can't expect them to read start to finish. You make a video — most people watch the first few seconds. That's reality.

Capturing attention is incredibly hard, and getting harder, because information keeps exploding.

In this environment, you need enough touchpoints. Sometimes I see a new AI product on Twitter, think it looks good, want to try it, but I'm busy, save it, forget. A few days later see it again, think it looks good, forget again. Then a friend recommends it. Maybe on the fifth time I'm nudged, I finally download.

From hearing about it to actually acting — there might be five touchpoints in between.

👦🏻 Koji

Yes, that's classic consumer marketing theory. Don't get annoyed by the repetition. To this day, McDonald's and Coca-Cola are still delivering their brand messages decade after decade: I'm about happiness. I'm about courage.

That framing you used was great — "the moment you feel like you're about to vomit from saying it." That's a good benchmark. Then tell yourself: "people are just starting to listen."

👩🏻 Zara

Exactly. So if you're building a personal brand — posting on Xiaohongshu, on Twitter — you need to post a lot. Volume matters as much as quality. Frequency has to go up. You'll notice that founders and investors who do really well on Twitter often post ten times a day.

These people aren't idly scrolling and tweeting. They have incredibly busy jobs. They still carve out time because it's a strategy.

To get traffic on Twitter, you have to engage heavily. That's how the algorithm works. Post casually. Lower the psychological barrier. Repeating and hammering on one thing over and over — it's absolutely necessary. Never assume that once you've said something, everyone's got it. Not even close. Maybe 1% of people caught 10% of your message.

LongCut.ai: My YouTube Learning Tool

👦🏻 Koji

You also build products yourself. Recently you launched something that helps people digest long YouTube videos — LongCut.ai.

👩🏻 Zara

Right. LongCut is the opposite of Shortcut.

👦🏻 Koji

Hahaha!

👩🏻 Zara

Because I believe learning shouldn't take shortcuts — you should take the longcut. It's a web app. You paste in a link to a long YouTube video.

The problem we're solving: most people want to efficiently absorb long YouTube videos, but people like me who watch them start-to-finish are a tiny minority. The traditional approach is dumping the link into Gemini or NotebookLM to get a text summary. But I hate summaries. Once something's summarized, you lose the raw stories, examples, and punchlines — and those are often the most valuable parts.

So I wondered: how can people consume content efficiently while preserving the original essence, without over-compressing it?

Our interaction works like this: the LLM reads the full transcript, extracts the most essential and uniquely insightful parts, highlights them on the timeline, and gives each section a title. You click a title to jump straight to that point in the original video.

My philosophy: you must watch the original video, even if you're skipping around. Don't assume that reading a text summary means you've absorbed the original argument.

Other features came from my own pain points. For example, watching AI researcher interviews — you hit terminology you don't understand, but looking it up loses context. So we built "One-Click Explain" — select a phrase in the transcript, and the AI explains it with surrounding context. There's also note-taking and a filler-word removal feature, making it easier to collect punchlines and reflections.

What Building Products and Creating Content Have in Common

👦🏻 Koji

Those three features all show deep insight. I feel like whether you're building products or creating content, you start from the user. You're a user yourself. Have you distilled any shared mental models?

👩🏻 Zara

Doing both simultaneously, I've found they have a lot in common. At their core, they're both "prediction systems for human behavior."

👦🏻 Koji

Prediction systems for human behavior?

👩🏻 Zara

Right. With content, I'm predicting: if I post this, will people watch it? Will they think it's good? Will it spread? Or will no one care? With product, it's: if I ship this feature, will people use it? What will the feedback be? You don't know any of this before you ship. Truth comes from practice.

Whether product or content, you need to run this feedback loop countless times: ship something, get user feedback or data, update your understanding and insight based on that, then ship the next thing. Infinite loop. The more cycles you run, the sharper your sense gets — content sense or product sense.

So there's no shortcut. No one is born with great sense. You just do enough, see enough data, and like machine learning, you gradually learn what users need.

👦🏻 Koji

I'm reminded of something you posted on Xiaohongshu: with AI so powerful, what value do humans still have? You mentioned three things: taste, distribution, and agency. What you just said — encouraging people to post more, iterate more, feel out their users — that seems like a concrete expression of agency.

Leverage That Requires No One's Permission

👩🏻 Zara

Naval has a famous line: there are two kinds of leverage that don't require anyone's permission — code and media. Both can make you money while you sleep. They're compounding over the long term. And you don't need anyone's permission to start. He said this before large language models existed. Today, with AI, it's even more true.

Growing up, especially in our education system, you're conditioned to think you need permission before doing anything. Parents say no, so you don't. Teacher says no, or your boss says no — you can't proceed. Or you think you need the right labels — elite school, big tech company — to do something. Or because you didn't study CS, you can't build something.

These all assume you need permission. But ask yourself: who's actually stopping you? No one. No one is saying you can't build a product. No one is saying you can't download an AI coding tool and turn your idea into reality. You can just post on Xiaohongshu, record a video. No one is stopping you.

👦🏻 Koji

This is easier said than done. Everyone knows posting on Xiaohongshu or vibe-coding is simple, but they can't take the first step.

Psychologist Li Songwei has a column called 5% Change. His view: ask people to change more than 5%, and they won't do it. So his advice is always about nudging you just slightly — 5%. You move, get positive feedback, then move more. For people who want to act, what's your 5% advice?

👩🏻 Zara

I hesitated for a long time before acting too. I wasn't a day-one action-taker. Before I started posting regularly on Xiaohongshu, I probably deliberated for a year. I bought lights, a tripod, said I'd film — the equipment gathered dust for a year. I couldn't get over the mental hurdle of not knowing what to say.

👦🏻 Koji

How did you get past it?

👩🏻 Zara

One day I just thought: screw it, post first. I stopped caring whether it was polished enough. I took questions people constantly asked me and explained them in plain language. I didn't even use that equipment — just held my phone and started talking.

Here's my cold-start advice: if you're creating content for the first time and don't know what to talk about, think about the three questions people ask you most often. If people keep asking you something, that means it's what others find most interesting about you. If you find yourself repeating certain things, they're perfect for video or content. That's exactly how I started.

Many people asked why I left VC for a tech company, or how as a humanities major I joined the internet industry. These answers seemed obvious to me — I was inside it, didn't feel special. But these were precisely what people were curious about. If you can step outside yourself, adopt an external perspective, and explain it in language a beginner understands — that's what people want to hear.

Don't have high expectations when you start. You won't blow up overnight. But lower the barrier enough. I have zero requirements for equipment or location — I just find somewhere and start talking. I've realized that what matters most now isn't being fancy. Not effects, not visuals, not flashy production.

People follow knowledge creators like us for the content itself — your insight. As long as the content is there and the topic is good, nothing else matters. I even think many viewers like my content precisely because it's unpolished. I'm not a professional creator. I don't spend time on surface-level stuff. The content is very raw, very human.

👦🏻 Koji

Love it!

👩🏻 Zara

Koji, what about you? What advice do you have for people starting out?

👦🏻 Koji

While you were talking, I had some realizations. I remember filming my first video — huge barrier. Being on camera, watching it back — couldn't stand it. My solution was finding a friend I really enjoyed chatting with, having them sit across from me holding the phone and asking questions. My first video came out of that impulse.

Second, it's a bit like losing weight or quitting smoking — you need to plant a flag and announce it widely: "I'm posting one video every day for the next month" or "I'm starting my diet next month." Once it's out there, it creates interesting social pressure.

I'm not encouraging everyone to be this intense. But for me, getting Crossing to a weekly release schedule this year was partly because early on, Ronghui and I kept telling everyone on social media: "We're a weekly podcast." At that point we were barely managing it. But the more we said it, the more people responded: "Wow, weekly — that's impressive." And slowly, we actually made it happen.

👩🏻 Zara

Like I said earlier — when the outside world sees you that way, you start seeing yourself that way too.

👦🏻 Koji

So these past few weeks we've been doing two episodes a week, hahaha!

Learn in Public: Why Learning Publicly Matters

👩🏻 Zara

I'm a big fan of the "learn in public" concept. It means not just learning, but sharing what you've learned — your takeaways, things you found valuable.

This has several benefits. First, many people avoid social media because they worry about not being expert enough. But realize: you don't have to posture as "teaching others." That comes across as condescending — people don't like it. You can approach it as a student: "I'm learning alongside you, sharing my notes, let's grow together." This is genuinely helpful. People love it, and no one will judge you for not being an expert.

Second, learning in public drives your input. Sometimes I have nothing to post on Xiaohongshu, so I go watch YouTube. To maintain your posting rhythm, you have to improve the quality of what you consume. And explaining knowledge in your own words is the best digestion process.

Also, learning in public generates positive feedback, which makes learning more sustainable. When the outside world sees you as someone who loves learning, you'll see yourself that way too — and then you actually start studying every day.

👦🏻 Koji

Positive feedback is crucial.

👩🏻 Zara

Right. In school, learning had built-in feedback — exams. Your score told you whether you'd learned it.

But after entering the workforce, there are no exams. You study something on your own, but you don't know if you've actually learned it, or — so what? What can you do with it? So I feel like if you're not learning in public, you might as well not be learning at all.

Because if other people don't know you learned it, that knowledge is hard to put to use right away. But if you can immediately showcase what you've learned in some form — whether as content, a small project, or a demo product — the outside world will know you have that skill.

👦🏻 Koji

You also get a huge motivational boost yourself.

👩🏻 Zara

Right. You feel a strong sense of drive, like "this wasn't a waste, I actually produced something."

👦🏻 Koji

This reminds me a lot of Duolingo's product design philosophy. You learn vocabulary, maintain your daily streak, they give you a badge — you get that sense of accomplishment, and you're willing to share "I've hit 323 days in a row." That's another source of sustained motivation.

Growth with Ease

👩🏻 Zara

Another keyword for me this year is "growth with ease" — yǒu sōngchí gǎn de chéngzhǎng.

I feel like I've grown quite a bit this year, both in content and product. But it hasn't been that kind of bitter, grinding struggle. I didn't force myself into Vibe Coding or posting on Xiaohongshu — I just seamlessly wove them into my life.

A lot of people ask me: "You work full-time, post on Xiaohongshu, try all these AI products — how do you have so much time?" And I ask back: "How do you have time to scroll through Xiaohongshu, browse Douyin, play games, watch shows?" These things are entertainment for me. They're strongly positive-feedback, energy-giving activities, not energy drains. So they've become completely sustainable. I approach them with real ease. Recording Xiaohongshu videos, for instance, is relaxing for me because I enjoy self-expression.

👦🏻 Koji

Right. A lot of people say true relaxation isn't lying down or scrolling short videos, because you actually feel emptier afterward.

👩🏻 Zara

Exactly. So I think ease is important — don't put too much pressure on yourself. With AI products or the latest models, I mostly approach them with a playful mindset. "What happens if I try this prompt? What about that one?" I find it genuinely fun, like running experiments, driven by strong curiosity and a spirit of exploration.

👦🏻 Koji

A lot of people probably envy this kind of "easeful growth." Any advice on how to find it?

👩🏻 Zara

This thing won't come find you — you have to go find it. Same as before: whether it's positioning or ease, you don't figure it out sitting at home doing nothing, or waiting for it to magically appear. It takes effort and action.

For example, I was pretty anxious learning about AI last year. I felt like I lacked relevant background and was just watching videos. Models weren't that smart yet, the Vibe Coding experience was poor, lots of ideas couldn't materialize — bugs everywhere, no idea how to fix them. But I broke through this year. Some things I can actually build now. I got positive feedback.

The prerequisite is you have to keep trying, don't give up. It's not "I posted a few videos, no one watched, so I quit," or "I learned something, got no feedback, so I stopped." If you persist, feedback will come eventually. Once you have feedback, it becomes easier for this to turn into something you enjoy. When it becomes something you enjoy, that's when you can have ease.

👦🏻 Koji

Right, and the "easeful growth" people find doesn't have to be in AI.

👩🏻 Zara

It can be any field.

👦🏻 Koji

I'm thinking of Paul Graham — I shared an article of his on the podcast before, How to Do Great Work. His specific advice was to try lots of things, find something where "it feels effortless to you, but others find it hard."

👩🏻 Zara

Naval said it too: "Find what feels like play to you, but looks like work to others." What feels like work to others but like play to you.

Content creation is that for me. For many people, there's huge pressure and a high psychological barrier to making content. For me it's genuinely like play. So don't feel anxious after hearing this. Your easeful growth, your "feels like play" thing — it doesn't have to be content or AI.

👦🏻 Koji

Right, it's different for everyone.

👩🏻 Zara

But I do feel like people shouldn't think of personal positioning as some pie that falls from the sky. Behind it all, you need to put in a lot of effort and action to find it.

The Long Game and Karma Theory

👦🏻 Koji

Get your hands dirty. A lot of what you do seems to have long-term value to me — like making friends with time. Your content, your Vibe Coding products, as you said, generate compound returns even while you sleep. How do you think about long-term value?

👩🏻 Zara

Many friends ask me whether Xiaohongshu is monetized, whether those products make money. Actually, no — the products weren't built for profit either. These are just my side projects, hobbies. I still have my main job.

But I've noticed many people's mental model for building products or content is that it must pay off short-term.

I see it as a longer-term thing. With content, there's the short game and the long game.

The short game is: accumulate followers, immediately take ads, monetize through traffic.

The long game is: I don't do any monetization, but through this platform and distribution, I've met many excellent people and connected with lots of resources — outstanding founders, technical talent, investors in the industry, or future collaborators, people I could recruit, and future customers.

For example, if I launch a product in the future, the cold-start traffic problem is basically solved.

I look at it from this angle. Every follower I gain today is marketing cost I'm saving for the future.

You can't wait until product launch to think about distribution. All content serves my future career; it's not the end goal itself. I don't want to become a full-time creator. This is the long game.

To accomplish something, the resources you need to mobilize, and the reputation and influence you need in your industry, are far greater than people imagine.

Many entrepreneurs only realize after starting that they don't know investors, don't know users, can't think of anyone to beta test their product. By that day, it's too late.

👦🏻 Koji

Right.

👩🏻 Zara

I think you should start building your influence, your reach, your network early. These will all help you later.

In my mind, I picture this like Karma in Buddhism — somewhat like merit. You help people without seeking return. On Xiaohongshu I haven't earned a cent, but I've helped many people, influenced their thinking and behavior, or they've learned something from me. Many Xiaohongshu fans bring gifts when they meet me offline. It's our first time meeting — they bring gifts because they feel I've helped them, so they want to give back in some way.

👦🏻 Koji

Wow, that must feel amazing.

👩🏻 Zara

You'll find that the Karma you accumulate will come back eventually. It's like depositing money in a bank — you can withdraw it someday, just not necessarily now. The longer you save, the more compound interest. I want to leave that money in as long as possible, not withdraw it quickly.

👦🏻 Koji

Have you thought about whether short-term and long-term value can be balanced simultaneously?

👩🏻 Zara

I think if you want to turn content into a career, or make it your main monetization channel, maybe that's possible.

Annual AI Product Recommendations

👦🏻 Koji

You mentioned you'd recommend some AI products you actually use. At this year-end, year-start moment — want to recommend which AI products you've found genuinely great this year?

👩🏻 Zara

A few categories. First, daily chat AI — I recommend Claude.

Because most people use ChatGPT, Gemini, or some domestic models. Fewer people use Claude. I personally love Claude's personality. It comes through in how it talks to you, its tone — unlike ChatGPT, it doesn't flatter you or always act like you're right.

Chatting with Claude feels like a spring breeze. I often discuss personal issues with it, or use it as a therapist. Also, its writing ability is very strong. I suggest everyone try it if they can, to experience this personality difference.

Second, for learning — I really love NotebookLM. It introduced a great concept: content remix. Everyone absorbs content differently — some prefer video, some podcasts, some text, long or short.

👦🏻 Koji

And some have reading disabilities.

👩🏻 Zara

Right. NotebookLM can transform content into whatever form you need. For example, a new feature this year can directly generate video overviews, or generate slides.

I have a highly recommended use case: Most people use NotebookLM to learn information; I use it to understand myself.

You can upload very personal things — like your resume — generate a podcast where two AIs praise you for 20 minutes, and you'll better understand where your competitive advantages lie. I also upload transcripts of daily conversations and have it generate slides. You'll discover it can organize the messy ideas in your head with remarkable clarity.

👦🏻 Koji

Yes, language is intelligence. Often we can't express ideas because our intelligence hasn't reached that point yet. But AI can sometimes help us articulate those half-formed, not-yet-distilled thoughts ahead of time.

👩🏻 Zara

Right, I feel that intensely. NotebookLM's slogan is "understand anything" — I'd add to it: "understand anything, including yourself." You can use it to understand a document, a paper, a video, but also to understand yourself.

👦🏻 Koji

So one is Claude, one is NotebookLM, any others?

👩🏻 Zara

For Vibe Coding. If you're just getting started, I recommend two. One is Replit Design within Replit. Replit is a very beginner-friendly Vibe Coding tool because it handles the full pipeline — from your idea to final deployment and going live, even domain names.

👦🏻 Koji

One-stop solution.

👩🏻 Zara

Exactly. And you never have to touch a single line of code. Its Agent is really smart. The Replit Design feature they launched is perfectly positioned — everything you build just looks good. So I'd recommend that anyone without a technical background start by building a personal website. Upload your resume and turn it into a beautiful personal site. You'll get some positive feedback right away. It's simple — no databases, no backend.

There's another niche product called Faces. It's also a website builder, but focused on lightweight sites like personal resumes or portfolios. A lot of people used to ask me for AI learning resources, so I compiled podcasts, videos, and newsletters into a website — and I built that with Faces. The standout feature is that everything looks great out of the box. All the interactions and animations are handled automatically. You don't need to write long prompts.

These two are perfect for beginners to get started and build some momentum.

AI Pro Tip: Let the AI Ask You Questions

👦🏻 Koji

Zara just recommended three products. Do you have any special, year-end-recommended AI use cases?

👩🏻 Zara

I have one: flip the relationship between you and the AI. Most people ask AI questions, but I think AI should be asking you questions.

Specifically, this works great for writing. When I need to write a document, I don't start with a blank page. I open Claude or ChatGPT in voice mode and say: "I'm going to write a product proposal. Ask me one question at a time to help me think this through." The key phrase is "one question at a time." Then it starts asking: Who's the audience? What problem are you solving for them? How is this different from competitors? And I just start talking. Speaking has no barrier — you can ramble naturally, no need to worry about structure.

👦🏻 Koji

When you said that, it reminded me of a small project we did at The Fair last year called "AI Will." The idea was that people rarely think about death, especially at our age. But thinking about death is actually useful — it clarifies what matters most in life.

So what we did was have AI ask you questions, and through that questioning process, help you write your will. It forces you to confront: if you actually had to think through your will, what adjustments and changes would you make?

👩🏻 Zara

Letting AI ask people questions — there are so many use cases for this. In the writing workflow I described, after talking with AI for ten or fifteen minutes, I'll say: "Now organize everything I just said into a document." And the document comes out. I barely need to edit it. And it's completely mine — it doesn't read like AI wrote it.

👦🏻 Koji

That's a really interesting use case.

👩🏻 Zara

And with this approach, you can write when you can't actually write. Once I was stuck in traffic, and I voice-chatted with ChatGPT about a document. By the time I got home, the document was done. You can do this in the car, while walking, or doing other things — writing gets done in the background.

👦🏻 Koji

I think everyone listening should try this. It's the turn of the year — perfect timing to write your goals for next year. I bet most people stare at a blank Word doc and can't get started. But try having AI ask you questions: "I'm going to set my goals for 2026. Ask me questions, one at a time, and help me draw them out."

👩🏻 Zara

Another technique: I ask AI, "What information do you need from me?" A lot of the time, you give AI very limited context — maybe just a short prompt, and you can't think of what else to add on the spot. But it needs way more context than that. Most AIs just start working at this point. They don't know what they don't know, so the output is poor. Fundamentally, you haven't given enough context.

So sometimes I'll add: "What do you need from me to do this better?" The AI will then reflect on what it needs. A good assistant will proactively ask you questions and tell you what resources and help it needs to get this done.

👦🏻 Koji

That's also a great pro tip. In DeepResearch or Manus, when you give it a task, it usually asks you three questions first to help clarify and expand your thinking. With more context and background, it can do better.

👩🏻 Zara

This technique works in life too, or treating AI as a therapist. Say something's bothering you and you can't figure it out. Tell the AI your confusion, then have it ask you one question at a time. Before you know it, you've talked your way to the solution yourself.

You'll realize you didn't actually need someone else to tell you the answer — you already knew it.

👦🏻 Koji

Therapists don't give advice either. They guide you through questions.

Spiritual Mentors of the Year

👦🏻 Koji

You mentioned that when learning, you especially like to consume content from builders. Who are your "spiritual mentors" of the year — whose content, insights, and thinking have helped and inspired you most?

👩🏻 Zara

I highly recommend this: when you find a builder you like, go through all their content and interviews on social media. It's such a great way to learn from one person.

The ones who stood out to me most this year: first, for product, Chris Pedregal, founder of Granola. Granola is a great case study — it innovated at the application layer and not only survived in a competitive space but got really strong feedback.

👦🏻 Koji

Right, Granola is a product with great user insight.

👩🏻 Zara

Exactly. He has exceptional user insight. His UX innovation fundamentally comes from deep understanding of his user base.

A few things he said really stuck with me. For example, as a PM, of course you do user interviews — but don't just build whatever users say.

Everything users say is actually training your own intuition. In the end, you still have to rely on your own intuition and sense to make decisions — but that intuition comes from countless user interviews and conversations.

He also believes that at the application layer, don't compete with the model layer — leverage its strengths. For instance, he thinks context window issues will be solved by model companies, so he doesn't spend excessive time on them.

Second is Josh Woodward, VP at Google, who leads Gemini and Google Labs.

I think Google made huge breakthroughs with Gemini and Google Labs this year. The whole narrative shifted. Last year people thought Google was still catching up; this year it feels like they burst onto the scene. Josh led many products from 0 to 1, almost like incubating startups internally. I highly recommend checking out labs.google — there are lots of small products there that haven't been publicly promoted yet.

What struck me about Josh: first, he completely lacks the typical big-company slickness. Watch his interviews — you can see the love for technology in his eyes, like a kid. And despite being at a big company, he operates like an entrepreneur.

For example, when Google Labs launches new products, he tells PMs not to stare at dashboards, but to look into users' eyes — see if they light up, if there's an aha moment. He thinks this metric matters more than numbers. Early on, you don't have enough volume for those numbers to mean anything compared to Google's hundreds-of-millions-user products. This perspective is really inspiring for anyone trying to innovate inside a big company.

👦🏻 Koji

That's interesting. It reminds me of how Procter & Gamble trains PMs on two "magic moments."

The first magic moment: when a user is standing in front of the supermarket shelf, seeing your product among competitors for the first time — do they feel the urge to pick it up?

The second magic moment: when they get home and unwrap it, using it for the first time — is the experience good?

These are very emotional judgments. Today with internet or AI products, because everything's digitized, we tend to be more data-driven in our decisions. But we've forgotten that PMs used to make decisions with intuition and feeling. Feeling and data are equally important.

👩🏻 Zara

Especially early on, feeling might matter more — because you don't have much data yet.

👦🏻 Koji

Right. Andrej Karpathy said recently on a podcast that a key difference between large models and human intelligence is that large models don't have feelings. Human feelings — while sometimes making us uneasy, anxious, or frustrated — are often what enable our more remarkable intelligence.

👩🏻 Zara

Another spiritual mentor, on PR, marketing, and go-to-market.

👦🏻 Koji

Who?

👩🏻 Zara

Lulu Cheng Meservey. If you haven't heard of her, I highly recommend searching YouTube and watching a few of her interviews. She's basically the "PR guru" of Silicon Valley. She previously led Comms or PR at several well-known companies, and now runs her own firm helping Silicon Valley founders with external communications and PR.

She has a very influential concept called "go direct" — founders and CEOs should directly reach and influence their audience. In the past, you might try to get TechCrunch to cover you, or buy ads somewhere. She thinks in today's era, these aren't the best channels.

Founder-Led Marketing: This Isn't a Personality Issue, It's a Strategy Issue

👦🏻 Koji

So she recommends founders go on podcasts. It's a very direct way to have conversations with users.

👩🏻 Zara

Bay Area AI founders are on podcasts constantly, talking about their vision, building in public. Many Chinese founders tell me: "I'm introverted," or "I'm an I, I'm not good at expressing myself."

But this isn't a personality issue — it's a strategy issue. These people aren't out there talking because they're extroverted or love expressing themselves. It's their communications strategy, their marketing and go-to-market strategy, and it genuinely helps their business. So marketing isn't mysticism — it's not that some people are naturally good at it and others aren't because of personality. If you treat it as a strategy and iterate on it seriously, you'll do these things.

👦🏻 Koji

This is really well put.

👩🏻 Zara

Lulu also had a perspective that really influenced me. She has a theory: your personal positioning is the intersection of three circles.

The first circle is "what you want to say." The second is "what your audience wants to hear." The third is "what's good for you."

The intersection of these three circles is your personal positioning. She was specifically talking about founders and people running businesses. If you're not a founder, you sometimes don't need to worry as much about the third circle — you might go viral talking about gossip, but that doesn't help your business.

Many founders know the first circle, what they themselves want to say. But what they're missing is the second — and it's not "what everyone wants to hear," it's "what your audience wants to hear." You need to be extremely clear about who your audience is. For a founder, your goal isn't to become an internet celebrity. It's to reach the people you actually want to influence. "Who is more important than how many" — reaching the right people matters more than how much traffic you get.

👦🏻 Koji

Yeah, I've noticed this too. A lot of founders who come on Crossing also have to overcome some introvert barriers. But for many of them, a big motivation is recruiting.

👩🏻 Zara

Right.

👦🏻 Koji

A lot of people assume they're doing it for VCs or customer acquisition, but I've felt this past quarter that many founders coming on the podcast are doing it so that the talents they want to attract in the industry know "who I am, what I'm doing, and whether I can inspire you with what I'm building to join me."

👩🏻 Zara

It's the best recruiting channel. People genuinely buy into the ideas you spread, and then actively want to join.

Also, the reason people come on podcasts is that this long-form content format is incredibly helpful for explaining something in depth. Most founders start companies because they have some belief that others didn't buy into but they firmly hold. That belief can't be explained in a couple of sentences. So podcasting is the perfect medium — taking an hour or two to walk through it and make it clear.

And right now "long-to-short" is very popular abroad. Few people will watch your two-hour content from start to finish, but you can cut out the精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华精华

Once ritual kicks in, you feel like the thing has life and meaning — because it even has a logo now, and subconsciously you feel more accountable to it. That's my hack for overcoming inertia.

👩🏻 Zara

Yeah. I think AI is the best of times for some people, and the worst of times for others.

For high-agency people, it's the best of times, because AI becomes your superpower and lowers the barrier for doing so many things.

But for another group of people, they really might get displaced, or discover that abilities they used to have are no longer needed by society.

👦🏻 Koji

So how do we help those people in the "worst of times" gradually turn things around, so this era becomes good for them too?

👩🏻 Zara

First, everyone probably needs to consciously learn how to use AI well. Whether or not you work in the AI industry, ordinary people need to learn how to harness AI, how to work alongside it. Because whatever you do in the future, AI can help you do it better — especially when it comes to knowledge work.

On the other hand, people should probably make a habit of asking: what value can I still provide in this era? Or what value can I create together with AI? Make it something that empowers you, not something that replaces you.

Introducing Myself with 10 "I Am"s

👦🏻 Koji

Finally, we have a bonus round: please introduce yourself with 10 "I am"s. These need to have some substance — you can't say "I am Zara" or "I am a Xiaohongshu blogger with 170,000 followers." You need to use things people might not know, interesting things, the B-sides of your life. Up for the challenge?

👩🏻 Zara

Haha! Let me try.

First, I am the mom of a one-year-old.

👦🏻 Koji

Wow!

👩🏻 Zara

Second, I am a former journalist.

👦🏻 Koji

You were a journalist?

👩🏻 Zara

Yeah, at The Information in Silicon Valley. That was my first time in Silicon Valley — I was a junior in college doing an internship there, writing articles.

👦🏻 Koji

The Information is also a publication that takes no advertising whatsoever, purely subscription-based.

👩🏻 Zara

Right, it only reports exclusive news.

👦🏻 Koji

Very impressive.

👩🏻 Zara

Third, I am a Northeastern Chinese person.

Fourth, I have been studying Japanese all along and passed the N1 exam.

👦🏻 Koji

Wow, how did you learn it?

👩🏻 Zara

I studied it for a few years in school, then kept teaching myself after graduating. I'm someone who really loves learning languages.

👦🏻 Koji

Okay, that's four.

👩🏻 Zara

Fifth, I am someone who has loved self-expression since I was little.

I was thinking about it the other day — when I was young, I would bind my essays into books, and I'd record covers of pop songs and burn them onto CDs to give to friends.

👦🏻 Koji

Haha, you were a content creator from birth.

👩🏻 Zara

Sixth, I am someone who has lived in three countries: China, the US, and Singapore.

👦🏻 Koji

You lived in Singapore too?

👩🏻 Zara

Yeah, I went to high school in Singapore.

👦🏻 Koji

Interesting.

👩🏻 Zara

Seventh, I am someone who did my undergrad at Harvard, studying psychology.

👦🏻 Koji

Pretty rare, a Chinese person studying psychology at Harvard.

👩🏻 Zara

Right, there aren't many of us. At the time in Singapore, I was the only student in the humanities track whose first language wasn't English. Our humanities track was very language-heavy — we studied English literature and things like that. I loved languages, so I ended up choosing a major that most Chinese people wouldn't pick.

👦🏻 Koji

That's seven.

👩🏻 Zara

Eighth, I am someone who doesn't drink coffee.

👦🏻 Koji

No caffeine at all?

👩🏻 Zara

No coffee, no tea.

👦🏻 Koji

Is that a deliberate choice, or do you get really wired?

👩🏻 Zara

I just can't handle it. But because of that, I also don't rely on it to maintain my energy.

👦🏻 Koji

So what do you use to perk yourself up?

👩🏻 Zara

I don't.

👦🏻 Koji

Hahaha!

👩🏻 Zara

I think my baseline energy is pretty high.

👦🏻 Koji

True.

👩🏻 Zara

Because everything I do is something I like. When you're doing things you enjoy, you don't even think about time management and all that.

👦🏻 Koji

That's eight.

👩🏻 Zara

Ninth, I am someone who has been in China's venture capital circle for almost 10 years now.

👦🏻 Koji

Right, we met about 10 years ago.

👩🏻 Zara

I interned at ZhenFund when I was a freshman — that was 11 years ago.

👦🏻 Koji

Right, then you went to GGV. We met through the GGV Fellow program.

👩🏻 Zara

That was the first GGV Fellow cohort.

👦🏻 Koji

At the time I went in as a founder to give a brief sharing for the GGV Fellows. I made a lot of friends through that program, and I'm still in touch with many of them.

ZhenFund's intern program is also quite interesting — now that I'm a Venture Partner at ZhenFund, I've noticed that ZhenFund interns have reunions every year, and interns from over a decade ago still come back. I haven't seen that kind of pull and cohesion in any other organization, whether a company or a VC. It's pretty special.

👩🏻 Zara

Yeah.

👦🏻 Koji

ZhenFund's intern program for this year has also launched — if you're interested, check out ZhenFund's WeChat account.

Zara, what's your final, grand finale "I am"?

👩🏻 Zara

Tenth, I am someone who doesn't really like being defined.

I'm constantly refreshing my definition of myself. When I was in school, I wanted to get into tech — I thought tech was so cool — but since I didn't study CS, I didn't know how to break in. So I chose to become a journalist at The Information first, thinking at least I could combine my writing skills with tech. Then I went into VC, spending half my time looking at deals and half helping the fund with branding and marketing — which was also tech storytelling in a way, but I still felt like an observer. Later I joined an internet company and started doing marketing operations, getting closer to the business, but I was still promoting a product. Finally I became a builder.

I feel like step by step, I got closer to the core of things.

But every time I do something, the outside world thinks "you should just keep doing this, you can do this forever" — people define you that way, because you might actually be decent at that particular thing.

👦🏻 Koji

Right.

👩🏻 Zara

Most people would probably just keep doing that one thing. But I want to keep refreshing my definition of myself. I can not only do that thing above, I can also do the next thing. I think this is a trait I discovered about myself this year.

👦🏻 Koji

This really gives a lot of people a sense of empowerment — like the boxes others draw for you aren't a problem at all. In you I see the power of not setting boundaries, constantly breaking through, constantly challenging.

👩🏻 Zara

Life is an infinite game.

👦🏻 Koji

Right, actually what we ultimately build is our own lives.

👩🏻 Zara

Yeah.

👦🏻 Koji

Thank you, Zara. Happy New Year, everyone!

👩🏻 Zara

Happy New Year!

👦🏻 Koji

I hope in the new year, as we've kept saying in this podcast, you take active action. But you don't have to be so intense or anxious — the key is finding that point where you can "grow with relaxed ease." Your point might be completely different from ours.

I hope we can all continue to live each day with relaxed ease next year!

👩🏻 Zara

Great, thank you, bye!

References

[1] LongCut: https://www.longcut.ai/

[2] LongCut.ai: http://longcut.ai/

---