This interview focuses on Eugene Wei's concept of 'invisible asymptotes'—the real reason growth slows isn't competition but that a product's design naturally excludes some users (e.g., Twitter's style appeals only to a certain crowd). Wei argues social networks are like 'proof-of-work' systems where users must perform specific tasks (like posting a nice photo) to earn social capital (likes, followers); this task determines the network's ultimate size. He cites Amazon as breaking its asymptote by entering new businesses, and WeChat as evolving from a status game into an indispensable utility. He also warns that digital networks amplify inequality, with early users' accumulated follower advantage (like 'old money') making it hard for newcomers to catch up.
At a Glance This episode of Invest Like the Best features Eugene Wei, former product lead at Amazon, Hulu, and Oculus, exploring the intersection of technology, media, and culture. The core concept discussed is "invisible asymptotes"—the fundamental reasons behind slowing growth for companies or ind
Eugene Wei, former early product lead at Amazon, product head at Hulu and Flipboard, and head of video at Oculus, centers this discussion around his core concept of "invisible asymptotes"—the fundamental reason for slowing growth in companies or individuals is often not competition, but rather the inherent product form that naturally sets an upper limit on the addressable user base; the key to breaking through lies in identifying and altering that "invisible boundary," rather than optimizing within the existing framework.
Eugene Wei argues that most companies attribute slowing growth to competition or market saturation, but the real cause is often the "invisible asymptote"—the product's inherent design naturally excludes certain user groups.
During his early days at Amazon, Wei participated in customer research and found that the company surveyed two types of customers: those who "bought but didn't buy more" and those who "never bought." The "never bought" group provided far more valuable insights than "loyal users" because the latter only tell you what they like, which is precisely the reason the former dislike it.
> "The very reasons might be exactly the reasons why other people don't like your product."
Twitter is a classic example: its interface style and content presentation are highly appealing to one type of user but completely off-putting to another. Wei points out that the more precisely a product serves a niche group, the lower its "invisible asymptote"—because the product is "tailor-made" for specific users and naturally unsuitable for others.
Path to breakthrough: either enter multiple new business lines like Amazon (initiating new S-curves), or "generalize" product features over time like Facebook/WeChat to attract a broader user base. However, generalization comes at a cost—when a network becomes "useful to everyone," its "buzz" and intensity of "status competition" diminish.
Wei argues that every successful social network inherently has a "proof of work" mechanism—users must perform specific actions to "mine" social capital (follows, likes, followers).
He draws an analogy with cryptocurrency:
Key insight: This "proof of work" mechanism not only determines who wins on the network, but also determines the network's ultimate scale. Musical.ly's "proof of work" only suits American teenage girls, so it can never cover everyone like Facebook does.
Facebook's success lies in continuously generalizing its "proof of work": from restricting registration to Harvard email addresses (natural status stratification), to opening up to everyone, then adding photos, videos, live streaming, etc.—each generalization expands the user base, but also dilutes the "heat" of status competition.
Wei highlights a core contradiction: Status competition requires "scarcity" to generate energy, but growth requires "generalization" to expand scale. All large social networks must eventually transition from "status games" to "utility tools"—for example, WeChat in China has become an "infrastructure" for payments, bill payments, and social interaction. Although it is no longer "hot," users cannot do without it.
Falsification condition: If a social network, after generalization, both loses its "heat" and fails to become a "utility tool," it will decline.
Wei argues that digital networks inherently tend toward "winner-takes-all" dynamics and amplify social inequality—whether in wealth, power, or social capital.
Historical analogy: The "equality-inequality" pendulum described in Will Durant's The Lessons of History is accelerated in the digital age. Network effects + smartphone penetration enable information, capital, and attention to concentrate at the top at an unprecedented speed.
The "old money" problem in social capital: Early Twitter adopters (e.g., those included in the recommended follow list) accumulated millions of followers, making it nearly impossible for later entrants—even with superior content—to catch up. Wei believes social networks need some form of "reset mechanism" to clear out "old money", so new users feel "the game isn't rigged, and I have a chance to win."
Conflict with market consensus: The market typically views network effects as a "moat," but Wei points out that if network effects + first-mover advantage + luck together create a monopoly, this may stifle future innovation and competition. He questions: "Do we want that? And how do we control that to encourage more entrepreneurs?"
Reader's note: Wei is expressing concern about the "winner-takes-all" structure here, not offering a solution. This is a reflection from the perspective of a position holder; readers should assess the policy implications on their own.
Wei argues that the potential of video as a communication medium is severely underestimated, particularly in the fields of education and knowledge dissemination.
Core Judgment: Most people do not enjoy reading. Among the sales of New York Times bestsellers, the actual completion rate is "shockingly low." In contrast, consumption of video (movies, TV series, documentaries) far exceeds that of text.
"Media Arbitrage": Converting text-based content into video can reach audiences who would otherwise never read it. Wei cites the documentary The Defiant Ones as an example—if it were a book, its impact would be far less than that of a four-hour documentary on HBO.
Why Video Outperforms Text:
1. Passive Consumption: Video plays automatically, whereas a book does not read itself aloud to you.
2. Richer Sensory Experience: Visuals combined with audio capture attention more easily than text alone.
3. The "Native Language" of the Younger Generation: For young people raised with smartphones and cameras, "selfies are their second language," making video communication far more efficient than typing.
Specific Opportunity: Educational content. Wei believes that the first wave of MOOCs (Massive Open Online Courses) merely "recorded lectures" without leveraging the visual advantages of video. The next generation of educational content should employ "cinematic language" and production value akin to movies to truly unlock the mass market.
Falsification Condition: If the production cost of video-based educational content cannot be reduced to a scalable level, or if users' consumption habits for "deep content" remain anchored to text, then this judgment may not hold.
Wei extends the concept of the "invisible asymptote" to personal growth — most people fail to break through their own bottlenecks because their feedback loops are too slow.
Core mechanism: Why are exercise and healthy eating so difficult? Because feedback loops are too slow — running three miles won't instantly shed five pounds, but eating a bag of chips provides immediate taste pleasure. "Almost everything in the world could be fixed by tightening feedback loops in general."
Solutions:
1. Seek external feedback: Like Atul Gawande, invite peers to observe one's surgery and offer advice. "We will always be the most biased observer of ourselves."
2. Turn goals into habits: Instead of fixating on the distant goal of "losing five pounds," fall in love with the habit of "doing 100 push-ups every day." "All success is just habitual."
3. Set arbitrary goals: Goals themselves are a form of "gamification." Wei cites his own monthly goal of doing 2,000 push-ups — once recorded in a Google Sheet, he remembers to complete it every night before bed.
Implications for companies: In the modern workplace, the average employee tenure is only about two years, leaving companies with little incentive to invest in training. Individuals must take responsibility for their own learning, and while the internet offers unprecedented free resources (YouTube, etc.), self-discipline remains a scarce commodity.
| Position | Guest Attitude | Key Data |
|---|---|---|
| Amazon | Positive case (a model of breaking through the asymptote) | Early customer research found that users who "never bought" had higher information value; Jeff Bezos is the "guardian deity" of Chinese tech companies |
| Risk warning (interface naturally repels some users) | Early third-party sites like "favstar" created global tweet rankings, activating status competition | |
| Neutral (analyzing its "proof of work" mechanism) | "Instagram coin" = taking a nice square photo → earning likes and followers | |
| Musical.ly | Neutral (as a differentiated case of "proof of work") | Its "proof of work" only suited American teenage girls, limiting network scale |
| Neutral (as a case of "generalization") | From Harvard-only email to open to everyone, then adding photos/videos/live streaming | |
| Positive (as a case of "utility tool" replacing "status game") | Became infrastructure in China for payments, bill payments, and social networking | |
| Netflix | Positive (as a case of "Silicon Valley's first wave of disrupting Hollywood") | Directly gave David Fincher $100 million to produce House of Cards, bypassing traditional studios |
| Apple | Positive (as a case of "Silicon Valley's first wave of disrupting Hollywood") | Specific data not disclosed |
| Google Maps | Positive (as a product with "the most potential but underutilized") | Underlying data can be repurposed for e-commerce, social networking, VR, and other scenarios |
| Beats | Positive (as a case of "cultural scarcity + marketing") | Banned by sports leagues → best PR event |
| Hulu | Positive (as an attempt to "reform the industry from within") | Specific data not disclosed |
1. The “invisible asymptote” is the true cause of slowing growth, not competition (Eugene Wei) — Product design inherently excludes certain user groups; the key to breakthrough is identifying and changing that “invisible boundary,” rather than optimizing within the existing framework.
2. Every social network is essentially a set of “proof-of-work” mechanisms (Eugene Wei) — Users “mine” social capital through specific behaviors; this mechanism determines who wins on the network and also dictates the network’s ultimate scale.
3. All large social networks must eventually shift from “status games” to “utility tools” (Eugene Wei) — Status competition requires scarcity, while growth requires generalization; WeChat is a typical example — no longer “hot,” but users cannot live without it.
4. “People who have never bought” provide more valuable information than “loyal users” (Eugene Wei) — Amazon’s early research found that the reasons users like you are exactly the reasons others don’t use your product; “The very reasons might be exactly the reasons why other people don't like your product.”
5. Video is a more efficient communication medium than text, and educational content will undergo a “cinematic” revolution (Eugene Wei) — Most people dislike reading, but video plays automatically; the next generation of educational content should have “film language” and production value.
6. Digital networks inherently amplify inequality, and social capital has an “old money” problem (Eugene Wei) — The follower advantage accumulated by early adopters is hard to catch up; networks need a “reset mechanism” to make new users feel “this game isn’t rigged.”
7. Almost all problems can be solved by tightening feedback loops (Eugene Wei) — Exercise is hard because feedback is too slow (running three miles doesn’t immediately shed five pounds), while eating fries brings instant pleasure; “Almost everything in the world could be fixed by tightening feedback loops in general.”
8. “High conversational entropy” is a hallmark of “interesting people” (Eugene Wei) — Google’s autocomplete feature reveals the predictability of most people’s email content; truly interesting conversations are those “a computer cannot autocomplete.”