Breaking Up the Internet

This was originally published on Substack, August 8th

The fragmentation of the internet is continuing. Top of mind for most Americans is the possible nationalization of TikTok. The Trump administration is also saber-rattling with a “Clean Network”. This behavior is dangerous for the future of American companies. We are losing the moral high ground on open markets, and fewer markets will be open to American companies. Fragmentation will accelerate.

There was already some natural fragmentation

Natural fragmentation. In some internet-enabled businesses, locals have a strong home-field advantage. They tend to not need any protection. There are millions of small complexities not translatable across borders. Or there is a need for a strong ground game. Examples include ridesharing/food delivery, accounting/tax, logistics, non-card payments, e-commerce, SME services.

Natural oligopoly/monopoly. Other businesses are much more global. They don’t need much local customization. This means providers can scale, bottoms-up, even without an office or local entity. Returns on scale mean winners take most. Examples include social media, communications, search, most SaaS, card-based payments, and digital infrastructure.

The current global providers of naturally oligopolic (love that I can do this in English) services are generally American. AWS (but not, and MS Azure will continue to dominate. WhatsApp is everywhere outside of the US and East Asia. Instagram is everywhere except China. Google search and Android will be near-impossible to unseat. However, America has burned through the goodwill generated by programs like and Next Billion Users.

Regulatory fragmentation is coming for the natural oligopolists

“How Asia Works” suggests a few key differences between successful and unsuccessful modernization. Successful countries protected and nurtured domestic manufacturing, but with mandated export discipline. Locals were protected at home, and got a lot of help, but had to compete globally. Unsuccessful countries opened access to domestic markets, but did not up-skill. Their economies are still dominated by multinationals. Oil, mining, auto, electronics.

China and Russia applied this framework to the internet. They’ve built protected local internets. India is starting, too. Governments will try to reduce the influence of Silicon Valley/Washington, and political stability / economic sovereignty will take priority.

Facebook has seen the fragmentation accelerating, and is scrambling to find local allies. It recently invested in Reliance Jio (India) and Gojek (Indonesia), both national champions. These also happen to contain 22% of the world’s population. Reliance Jio and Gojek will then encourage local lawmakers to remain open to FB.

What’s next?

Real answer is ¯\_(ツ)_/¯, but that’s never stopped thinkbois like me from speculating. Generalizing from the FB example:

  • Politicians, regulators, state media (especially large, nationalistic, or authoritarian ones) intensify targeting American companies for: 
    • Taxation
    • Nationalization / divestments
    • Shakedowns
    • Jobs programs
    • Censorship
    • Scoring points
  • Favoritism toward local champions become more overt
  • “Regulatory consulting”, campaign donations, direct bribery grow
  • Multinational (esp. American) corp dev teams get busier yet face more uncertainty
  • Development dollars used as currency for US/Chinese alignment

Who benefits?

  • Local competitors to natural oligopolies in large markets; in the short term. If markets open up again and they can’t get legal barriers, they’ll get stomped
  • Politically-connected entrepreneurs, “regulatory consultants” and lawyers
  • Consumers in small, still-open, or previously-ignored markets
  • VCs who can’t get into X, but can get into X for Region
  • State censors

Who loses?

  • Local competitors in small, still-open markets, who get stomped in the short term
  • Consumers in restricted markets, who get less choice
  • Companies with X for Everywhere baked into valuation
  • American multinationals on average. $FB? $MTCH? $NFLX? Someone else can tell you how much international growth is baked into valuations. 

Some other examples that helped form this thesis:

List is definitely incomplete. I’ve ignored Africa and most of Central/South America, which I know little about. I would love to learn more.

Natural fragmentation:

  • Ridesharing – Uber burned billions then sold local ops to Didi, Grab, Yandex
  • Ecommerce – losing in Asia (seems ok in Europe)
  • Food delivery wars are not settled, but mostly local players
  • Stripe has given up on India

Regulatory fragmentation:

  • TikTok nationalization in the US
  • Huawei’s troubles
  • GDPR
  • Reliance Jio’s fundraising bonanza. It included Facebook, Google, a bunch of American private equity, the UAE and Saudi sovereign wealth funds
  • India bans on Chinese apps, bounty for a local Zoom clone
  • Lark, Bytedance’s Slack competitor
  • Whatsapp Pay blocked in India and Brazil
  • Indonesia’s “foreign digital goods” tax
  • Facebook censoring in Singapore

Thanks to Turner Novak for reading a draft of this essay. 

Career advice for employees at struggling companies

This piece was originally shared on Substack on April 22, 2020

Most career advice on the internet is from people who had some sort of meteoric success. Why read advice from someone who’s had a mediocre career? But there’s massive sampling bias. All this advice will try to draw grand, sweeping narratives and also typically fails to sufficiently factor in luck.

Most content is also produced by VCs for founders, which is two layers of incentive structures away from regular working stiffs. I’ve spent a lot of hours with friends and and colleagues examining what we would’ve told our past selves. Below is a quick distillation of wisdom if:

Your company is struggling

  1. The company is not your family. Some of the people in the company are your friends in the current context. It’s like your dorm in college. Hopefully some of them will still be your friends after. But don’t stay because you’re comfortable.
  2. Don’t think that there won’t be politics because there wasn’t politics before. Politics emerge most strongly when the players believe the game is zero sum, which happens when growth slows, stops, or reverses.
  3. Maybe the turnaround will work. Maybe it won’t. You have to decide whether it’s worth waiting for. Your existing time is a sunk cost.
  4. Retention offers are negative signals for the company, even though they’re positive signals for you, personally. If you take a retention offer, DO NOT act like you’ve already earned the money. Those 6-12 months could destroy your soul.

You’re not sure of how much risk you want to take

  1. It’s okay to go for the safe job. Sure, Airbnb was founded during a recession. 50% of the time, you’ll survive the charge across no-man’s-land and reach the enemy trench.
  2. It’s also okay to take risks. Staying at a company that’s slowly dying has its costs too. Stick around too long and you’ll lose your belief that you can build, that change is possible. Try not to learn the wrong habits.
  3. Just… try not to fuck up the ratio. You can take the risk but if you do so, demand some of the upside. 

You’re finding a new job

“When a management team with a reputation for brilliance tackles a business with a reputation for bad economics, it is the reputation of the business that remains intact. – Warren Buffett, maybe

  1. When picking a job, yes, your manager matters. But if you have an amazing manager at a shit company you’ll still have a shit time. In some ways, it’ll actually be worse. If they’re good at their job (including retaining you), they’ll keep you at a bad company for too long. And then they’ll leave, because they’re smart and competent. Maybe they’ll take you with them.
  2. Your equity package is a lottery ticket with expected value of zero.
  3. VCs are salespeople for a living, are constantly talking their own book, and only have to be right 1/10 of the time. Even the ones you trust to have your best interests at heart are often wrong. Don’t ever join a company solely on the recommendation of a VC. Even if they’re a board member. It’s the nature of boards that they ignore (externally) problems until their hand is forced.
  4. “Take any role, at any pay, on a rocketship and everything will work out” is only sort of true. It’s the Hollywood model… do it for the exposure and it might work out great! Technically true, but it’s also a way for people with more information to offload the risk of bad companies onto… you.
  5. Seriously, you have to form your own opinion. Hyped = hiring good PR != successful. Every time I’ve outsourced my thinking for a job change, I’ve missed out on 5-6 figures of equity appreciation. But, who the fuck knows. Get lucky picking the right person to follow and it’ll probably work out too.

Lastly, it all tends to work out, on average, eventually. Don’t beat yourself up too much.

2018 Books

Favorites – nonfiction

  • 7 Powers: The Foundations of Business Strategy
  • Born a Crime
  • The Elephant in the Brain: Hidden Motives in Everyday Life
  • Bad Blood
  • Boyd: The Fighter Pilot who Changed the Art of War
  • Between the World and Me
  • Thinking in Bets
  • Algorithms to Live By
  • Poor Charlie’s Almanack
  • How Not To Be Wrong

Others (still liked) – nonfiction

  • Atomic Habits
  • Barbarians at the Gate: The Fall of RJR Nabisco
  • Game Change
  • High Growth Handbook
  • Measure What Matters
  • Only the Paranoid Survive
  • Principles: Life and Work
  • Reminisces of a Stock Operator
  • Sam Walton: Made in America
  • The Antidote: Happiness for People Who Can’t Stand Positive Thinking
  • The Art of Being Unreasonable
  • The Inner Game of Tennis
  • The Smartest Guys in the Room
  • Titan: The Life of John D. Rockefeller, Sr.
  • The House of Morgan
  • Born Standing Up
  • The Upside of Stress
  • Getting it Done: How to Lead When You’re Not in Charge

How to Kill Meetings and Demotivate People

I offer this to you in the hopes that it will help you identify and stop unproductive behavior from yourself or from your colleagues.

Back in World War II, the OSS (predecessor to the CIA) created a document called the Simple Sabotage Field Manual for ordinary citizens in occupied territory. The first half focuses on physical sabotage, like putting sand in the gears at a factory. The second, in my opinion, is more interesting.

“A second type of simple sabotage requires no destructive tools whatsoever… It is based on universal opportunities to make faulty decisions, to adopt a non-cooperative attitude, and to induce others to follow suit.

Simple Sabotage Field Manual

The manual is a quick read and fairly interesting – but here’s my favorite excerpt:

Narrator / New Yorker cartoon caption: “The second Netflix culture deck was much less successful than the first.”

I can think of specific meetings in the last month where any (and all) of these have happened… but for the modern reader and my own amusement, I’ve updated the list. 

Disclaimer: I am NOT advocating you do any of these things! I am advocating that you do the OPPOSITE. 

  1. Insist on strict process. Never permit short-cuts to be taken in order to expedite decisions. Require that a JIRA ticket be filed before any action can be taken.
  2. Talk as frequently as possible and at great length, especially over the women. Never hesitate to make a few appropriate comments about “disruption”, “innovation labs” or “Google does it this way”.
  3. When possible, refer all matters to steering committees for “further review from key stakeholders”. Attempt to make the committees as large and senior as possible – never less than five. Never set an agenda. Bonus points if you spend 5 minutes messing with the projector before showing an illegible dashboard screenshot.
  4. Bring up irrelevant issues as frequently as possible. “What if we put this on the blockchain?”
  5. Haggle over precise wordings on PowerPoint slides. “Is this really a guiding principle? Or is it a strategic pillar?”
  6. Refer back to matters decided upon at the last meeting and attempt to re-open that decision. Also known among practitioners as the “disagree and un-commit”.
  7. Create analysis paralysis wherever possible. Debate whether an action is “on strategy”. Insist on additional data, litigate assumptions, and add additional scenarios to “stress test”. Then say the model is “garbage in garbage out”. “What if the recession hits two quarters earlier?”
  8. Let legal, compliance, and risk have everything they want.
  9. Send Slack GIFs all day.

Reducing analysis paralysis

There’s a beautiful scene in Master of None where Dev (Aziz Ansari) can’t decide where to go for lunch. Even after he decides on tacos, he needs to know who has the best tacos. He whips out two devices and consults Google, Eater and Yelp. He sorts through dozens of lists, eventually settling on a food truck that is the most delicious taco in the city. Even when he’s there, he can’t decide what taco to get – he asks “What are the most popular tacos? Between carnitas and chicken, which do you think customers are enjoying the most – like from their faces?” It turns out the truck was sold out – he took too long in deciding. “What am I supposed to do now?” he complains. “Go eat the world’s second-best taco?”

I faced a similar situation when my roommate hosted a movie night a few months ago – we watched the first ten minutes of four different movies, with about five times as many suggestions thrown around, before settling on 22 Jump Street, which almost everyone had seen already. Every movie we looked at was objectively good – between seven smartphones on IMDB, Netflix’s recommendation engine and a laptop, we were bound to get good candidates.

We live in an era of endless optimization. Recommendation engines and algorithms and unpaid interns and random people on the internet are constantly churning out ways in which we can have the best. We think endlessly about opportunity cost and are paralyzed by choice – am I getting the best food on the menu? Am I following the right meditation program? Is my boyfriend of the right race to generate the cutest babies? (Not joking, I’ve heard this before.) Am I writing about the best blog topic I can? My favorite: Am I waiting long enough to text my crush back? (The blog is pretty entertaining as well.)

There is this aura of complexity that’s created by our endless need to make a well-informed choice, and this aura obscures what our original goals were – to the point where we forget what we started out to do. We spend so much time focused on opportunity cost that we forget about search cost.

Over the last three months, I’ve attempted to apply the concept time is money to reduce analysis paralysis in my life. I don’t implement it strictly, but the general concept is as follows:

Would I pay someone (me) $10 an hour to make this decision? Conversely, would I be willing to do this for $10 an hour? (BTW if anyone makes an app out of this, I want to be listed as a “technically cofounder”)

Would I pay someone $10 to spend an hour on Yelp deciding between taco trucks? Fuck no, that’s like 5 tacos. I’ll go to both 🙂

Would I spend an hour optimizing a Coachella schedule for $10? Absolutely. I do that for fun.

It’s worked out pretty well so far, but if you have better ideas I would love to hear them. I’d pay $10 to work on better allocating my time!

practice, practice, practice

I’ve come across some version of this anecdote from the book Art and Fear too many times to count.

The ceramics teacher announced he was dividing his class into two groups. All those on the left side of the studio would be graded solely on the quantity of work they produced, all those on the right graded solely on its quality.

His procedure was simple: on the final day of class he would weigh the work of the “quantity” group: 50 pounds of pots rated an A, 40 pounds a B, and so on. Those being graded on “quality”, however, needed to produce only one pot – albeit a perfect one – to get an A.

Well, come grading time and a curious fact emerged: the works of highest quality were all produced by the group being graded for quantity!

It seems that while the “quantity” group was busily churning out piles of work – and learning from their mistakes – the “quality” group had sat theorizing about perfection, and in the end had little more to show for their efforts than grandiose theories and a pile of dead clay.

It seems there’s no substitute for making mistakes and then learning from them. I want to share three anecdotes and what I’ve taken from them:

Hand billing

When passing out handbills on the main plaza in college, I attempted to go for percentages instead of absolute numbers. I’d handbill to people who were already making eye contact, who weren’t wearing headphones, who were already carrying other flyers. I had a great hit rate – over 80%, and my ego was nicely cushioned by a low rejection rate, but I didn’t get any better at hand billing.

I sat around and theorized about perfection, then I made the ceramics equivalent of a found object. I definitely didn’t get better at cold approaches, or talking to unreceptive people, or at handling rejection. Only practicing on something easy means you don’t get better.

Cold calling

In the following summer’s unpaid internship, I cold-called potential multi-family residential real estate clients in Walnut Creek. This largely meant suggesting to senior citizens that they look into selling their homes while they 1) wondered how the hell I got their number, 2) who the hell I was, and 3) why the hell I thought they’d want to sell their homes. This was a rather traumatic experience for a shy 18-year-old. I didn’t improve much from this experience either.

I was trying to wingsuit without the skill to indoor skydive. This was the opposite of my previous experience; practicing something too difficult is both discouraging and unproductive.

Bro-ot force learning

This story is from a banker who decided to wax lyrical during an interview. He had a fraternity brother, a legacy admit, who wouldn’t have gotten in through the regular process. He wasn’t particularly handsome or charming, and struggled both to connect with his brothers and with courting women. However, he realized his weakness and set out with a goal: to learn three things about ten people at every party. Freshman year, he showed a marked improvement. Sophomore year, his social skills were on par with his brothers. Senior year, he was president of the fraternity and one of the most popular guys on campus.

I don’t know whether this story is true, but I swear I didn’t make it up for the sake of narrative cohesion. The guy recognized a shortcoming, set a specific, quantifiable goal, and practiced every week in the relatively safe environment of his frat parties.

This blog is my pottery studio, my indoor skydiving, my fraternity, so that I can practice, practice, practice. Let me know how I’m doing, bro.

You don’t have to beat yourself.

I read The Silicon Valley Suicides today and identified so strongly with the students that it was difficult to read.

“Luthar constructed a profile of elite American adolescents whose self-worth is tied to their achievements and who see themselves as catastrophically flawed if they don’t meet the highest standards of success. Because a certain kind of success seems well within reach, they feel they have to attain it at all costs—a phenomenon she refers to as “I can, therefore I must.””

I used to think of life as a marathon – a marathon with a starting line stretching from horizon to horizon, an endless sea of runners that looked exactly like me. Each runner was a different version of myself, a more successful, capable version that didn’t need water at mile 21, or didn’t trip at mile 17, or didn’t get a B+ in his first quarter of AP Chemistry. My eyes fixed on the front-runner, I believed I was at the end of the pack. The lowest branch of a decision tree.

The more people told that me I had great potential, the higher the praise, the faster everyone else ran, and the further I fell behind. I thought about giving up altogether. All my choices were oriented around preserving my place in this marathon, around not falling further behind. I was deathly afraid of going backward – it would feel like all of my previous hard work would be erased, that I had failed.

When I was thirteen, I was upset that taking AP classes as a freshman required lobbying administration, and more so that I didn’t think to – I felt as though I had been held back, artificially limited, Harrison Bergeron’ed by bureaucrats into mediocrity. I was endowed with the terrible burden of my parents’ gift – they fought and scraped their way across 30.5% of the earth’s surface area so that I could have a successful, achievement-filled life, and I felt this gift was ruined if I took a semester of non-AP physics.

My parents are actually quite good, as Asian parents come – although they seem to think that pointing out my pimples is constructive and like my sister’s Dartmouth gear a lot more than my Berkeley gear, they do tell me that all they want is for me to be a happy, contributing member of society.

In some ways, this is scarier. Chasing achievements is not by any means easy, but it is defined. Take these classes, say these things during the interviews, improve process by this much. Lately my quarter-life crisis has come from defining my own path; separate from what I interpret from my parents or my peers. I don’t know if I’m getting it right. I’m not always happy and only contribute a little. To continue the metaphor, the marathon is already over. I just have to decide where to eat after.

The good news is that we get a lot of tries. Malcolm Gladwell argues that it takes 10,000 hours to master a subject – the next fifty years have 438,000 hours in them. Even with three hours a day (1/8 of my time) to learn new, that’s almost six tries. We don’t have to get it right the first time.