The Lunar Society

Share this post
❤️ + Musings
www.dwarkeshpatel.com

❤️ + Musings

Girard at Jamestown, VC bubbles, Elon’s ambition, big tech R&D as F1 cars, med-school mania, and the entrepreneurial middle class.

Dwarkesh Patel
Apr 20
Comment13
Share

I went from 800 to 14,000 followers in 48 hours, and all I did was ask why Einstein was so productive that one year.

Wednesday evening: It’s been four months since I’ve graduated college, and I’m starting to feel a little embarrassed introducing myself to people as a “blogger and podcaster” (a non-starter with girls).

I released my new post about miracle years this morning, and it’s gotten a dozen or so likes on Twitter but lead to no new subscribers - just like my last post.

What am I doing with my time? I have a CS degree - why am I writing essays instead of coding? I message a friend who works at a tech company, “You guys hiring any developers?” “Constantly. You looking?” “Maybe, I’ll let you know in a few days.”

And then Marc Andreessen tweets this:

Twitter avatar for @pmarcaMarc Andreessen @pmarca
On great scientists and their miracle years. Side note, I believe none of the work of Einstein, Newton, Copernicus, Darwin, von Nuemann, or Gauss was peer reviewed. Has the method of science regressed? @dwarkesh_sp
The mystery of the miracle yearAn interesting pattern recurs across the career of great scientists: an annus mirabilis (miracle year) in which they make multiple, seemingly independent breakthroughs in the span of a single year or two. Einstein had his annus mirabilis in 1905. While he was still a patent clerk, he wrote four pape…dwarkeshpatel.com

April 14th 2022

197 Retweets1,696 Likes

My Twitter starts to blow up. The next morning, someone DM’s me, “How do you know Jeff Bezos?” I’m confused - why does this guy think that an unemployed 21 year old knows Jeff Bezos? And then I see it. @JeffBezos was following me.

So I tweet out a screenshot of my view of Bezos’s profile, where it says “42 Following” and “Follows You”. A mutual asks the only reasonable question: “yo WHAT”. And I respond with the only reasonable answer, “I have no idea what the fuck is happening lololol.” A few hours later, Bezos replies:

Twitter avatar for @JeffBezosJeff Bezos @JeffBezos
@dwarkesh_sp @joodaloop You’re thoughtful and thought-provoking. Gratitude. Please keep it up!

April 14th 2022

24 Retweets1,418 Likes

I’m still having trouble imagining the process by which Jeff Bezos might have discovered my content. Huh, why did Newton get so much done during the Bubonic plague? Click link. ‘opposite of modern PhD program’?! Damn, gotta follow. It makes no sense, and obviously I recognize how ridiculous and unmerited this all is.

I remember meeting this guy, who would later become my roommate, in a freshman entrepreneurship course. He told me that the class was bullshit and that if I really wanted to learn about about startups, I should read someone named Paul Graham instead. I spent the next few months binge reading all of PG’s essays and discussing them with anyone who was remotely interested (or overly polite). So it’s impossible for me to describe what it felt like when this happened:

Twitter avatar for @paulgPaul Graham @paulg
"Ken Thompson built the basic chassis of UNIX in a week when his wife went on vacation and he had a lot of time on his hands."
The mystery of the miracle yearAn interesting pattern recurs across the career of great scientists: an annus mirabilis (miracle year) in which they make multiple, seemingly independent breakthroughs in the span of a single year or two. Einstein had his annus mirabilis in 1905. While he was still a patent clerk, he wrote four pape…dwarkeshpatel.com

April 15th 2022

144 Retweets1,418 Likes

The kind words and followers I have received in the last few days are orders of magnitude more than I deserve, and over the next few months, I will attempt to be worthy of even a small portion of them.

I am thrilled to announce that the home for these new endeavors will be Substack1. I can’t wait to show you what I get up to!

Musings

Girard at Jamestown

Why did British investors keep pumping money and people into the Virginia Company (and why did the monarch sanction it) despite the fact that Jamestown was an obvious failure? It was a hotbed for malaria and other diseases, an infertile backwater, and a death sentence for the majority of colonists.

Because Britain was in mimetic competition with Spain2. It superficially copied Spain (by investing in a colony in the new world) but it did not replicate Spain’s underlying cause of success at the time (extract a precious metal which is stupendously valuable in China as a substitute for the Ming dynasty’s unreliable currencies).

From Charles Mann's book, 1493:

When the initial hope of discovering precious metals and a route to Asia didn’t pan out, the company tried wine making, shipbuilding, iron monging, silk weaving, salt panning, and even glassblowing. All failed, at dreadful cost in money and lives. Nonetheless, the firm kept dumping money and people into Virginia.

We have a strong impulse to emulate the surface-level details of the successful without analyzing how exactly they put themselves in the right place and time. During the Dot-com bubble, hundreds of CEOs and millions of investors noticed that internet companies were incredibly highly valued, and too many figured that all they had to do to match the success of Yahoo and Amazon was to suffix their brandname with a “.com”.

You see this in modern bubbles too: people assume that by jumping to Substack (👀), they’ll become the next Glenn Greenwald or Jonah Goldberg, but fail to notice the decades of experience and accumulated followings these writers bring with them. To be clear, I think you should move to Substack, even if you’re starting a new blog (I wish I had done so much earlier). But it would be a mistake to assume that Substack makes the writer and not the other way around.

Same goes with founding a remote work startup after the virus has already leaked from the lab or getting into crypto after the entire world has heard about it. I’m not saying there’s no alpha left in these categories, but the obvious gimmicks are ran through.

Why venture capital causes bubbles

In my post on talent as leverage, I write:

Technology, more than any other sector, seems to have this strong pattern of bubbles, where one hype train follows another. Perhaps this is because the smartest, most talented people go to work in tech, and just as credit produces bubbles in financial markets, talent contributes to bubbles in technology, because that is where it is most concentrated.

But another more direct reason for why technology has so many bubbles is that before a startup goes public, there is no way to short it.

In his book on the history of venture capital, The Power Law, Sebastian Mallaby explains why the Dot-com bubble spiraled so out of control:

Unlike hedge funds, which can bet against a bubble by using derivatives and other tricks, venture capitalists can only bet on values going up.

If someone knows that a startup is worth less that the marginal venture capitalist is willing to bid for it, he can’t make money by correcting the foolishness.

What compounds this problem is that nobody wants to be guy who declined to invest in Uber, and the peace of mind which comes from having a stake in everything probably leaves you with net negative expected value.

Eliezer Yudkowsky explains how the same concept applies to housing markets in his book Inadequate Equilibria:

Some smarter agents might decline to buy, and so somewhat reduce demand. But the smarter agents can’t actually visit Boomville and make hundreds of thousands of dollars off of the overpriced houses. The price is too high and will predictably decline, relative to public information, but there’s no way you can make a profit on knowing that. An individual who owns an existing house can exploit the inefficiency by selling that house, but rational market actors can’t crowd around the inefficiency and exploit it until it’s all gone.

What raised Elon’s ambitions

In my post on how scouting talent is like buying options , I explain why you should do cool shit when you’re young, as it signals you’re a higher volatility asset. But there’s another reason why it’s import to pull some crazy impressive stunts in your 20s - it sets a minimum bar for future success which prevents you from becoming too complacent.

In his excellent book on the founding of PayPal, Jimmy Soni discusses how the success of Elon Musk’s first startup (Zip2) raised his ambitions so much that Musk was determined to revolutionize the entire financial industry. He saw PayPal’s main business of facilitating email payments as simply a stepping stone to this larger goal:

“Musk set a big goal because he sought a big win—bigger than Zip2. “I built a company four years out of college that sold for three hundred million dollars. So an outcome like that would have been, like, I already did that,” Musk admitted.

And after that, he needed a bigger win that PayPal - perhaps setting up a human colony on Mars would do.

Why immigrants and the middle class are so entrepreneurial

When Peter Thiel gets asked about why the PayPal diaspora has gone on to found some of the top companies in the Valley (LinkedIn, YouTube, SpaceX, Tesla…), he responds that building PayPal was the exact right difficulty problem. If you work for a company like Facebook or Google and oil springs into the air as soon you crack the ground, then you learn the lesson that startups are easy, and you don’t know how to respond to serious challenges. If you work for a company that fails, then you learn that startups are impossible, and you avoid building your own. PayPal taught its early employees that startups are possible but difficult.

This is also an excellent explanation for why most entrepreneurs come from the upwardly-mobile middle class. If you grow up with a silver spoon, you expect easy success, and you don’t perform the grueling and often menial tasks that are required to build something new. If you grow up in poverty, you worry that a stable job is the most you can hope for, and any scheme that risks this prospect becomes not worth pursuing.

Byrne Hobart notes that the PayPal mafia’s success is the result of an extremely talented team that got a middling exit which left them hungry for more. They didn’t make enough to rest on their laurels, but they had enough to invest in new ventures.

The immigrant experience can be described in the same way. Some of the most talented people in India and China come to America, but because of the broken immigration system, they get chained to their mid-level careers as doctors or engineers - a respectable outcome, but not one their children can count on for support.

The reason these second generation immigrants are so often successful is the same reason that the PayPal diaspora went on to create half of Silicon Valley - the shift from third world to first world showed them that great things are possible, but it wasn’t big enough or easy enough to make them complacent.

How big tech R&D is like Formula 1

Car companies lose money by developing Formula 1 vehicles - they cost a tremendous amount to develop and improve (there’s only a 1.3% difference in performance between the best and worst car). The reason they keep doing it is that it makes customers associate their brand with a machine that is monstrously fast and expensive - but one which they will never actually get a chance to climb inside.

I think this is an excellent analogy for why big tech companies invest so much in crazy R&D projects that have no feasible payoff in the short to medium term. Google paid half a billion for DeepMind and has invested hundreds of millions on it since3. Microsoft invested a billion in OpenAI, and Amazon has announced that it will spend 10 billion dollars to build a satellite internet service similar to Starlink.

The bottom line revenues from these moonshot investments are distant and speculative4, but the reputational benefits are immediate and significant. You want to work for a company that is building AGI, yes or yes?

Don’t forget: your Civic has nothing in common with Honda’s F1 on the racetracks, and the median Google employee will be working on fine tuning the sub-sub-sub settings menu in the new version of Android, not on building a deus ex machina.

By the way, I am glad that these companies invest so much money on R&D signalling. The externality is a higher rate of innovation and growth for the rest of us. My previous blog was made with React (released open source by Facebook), deployed for free on Netlify (basically a frontend for AWS), and featured code I executed in Google Colab. While I was working on the blog, I spent some of my free time teaching myself machine learning in PyTorch (Facebook again) running on a notebook in AWS.

Bell Labs was AT&T’s F1 racecar - it was an attempt to make a government monopoly seem beneficent. People were basically paying a small tax on every call they made, and 2.8% of that revenue was distributed to R&D, with only 0.3% going to basic research. Still, the result was 6 Nobel Prizes, the transistor, laser, Unix, C, and information theory.

Med school mania

Here’s a question I’ve had for a long time: suppose that going to med-school or law-school costs you $200,000 dollars and 4 years of time. The average medical or law student probably won’t enjoy being a doctor or lawyer (actually I can’t imagine anyone enjoying being a lawyer).

So why don’t they just decide to take that money and invest it in trying 4 businesses over 4 years5? The fact that they can get into a medical school (that is not in the Caribbean) shows that they have the intelligence and grit to start a company. So what is the lacking third trait which would allow them to do something more interesting and higher in expected value? Perhaps courage? Tolerance of the unknown? Self discipline and the ability to work without a boss or professor yelling at them? Let me know in the comments.


There are so many people who I have to thank for the events of the last few days, and every one of you who has decided to follow me on this intellectual journey is on that list. But I want to express my special gratitude to these people:

  • Bryan Caplan for being the best friend and mentor I could have ever asked for. I started my podcast two years ago. COVID hit, my classes went remote, and I was bored. So I cold emailed Bryan asking him to be my inaugural guest. And everything since has been compounding returns on his kindness.

  • Anil and Sunil Varanasi for offering me a grant when I was close to graduation and didn’t know what to do next. Their support gave me the financial ability and self-confidence to pursue this hobby of mine.

  • Leopold Aschenbrenner and Nick Whitaker for offering me support from the Effective Ideas Blog Prize and inviting me into a community of bloggers where I could be a peer to my heroes.

  • Applied Divinity Studies, Stephen Malina, Fin Moorhouse, Brad Davis, Rohit Krishnan, Erik Hoel, Slime Mold Time Mold, and Basil Halperin for giving me invaluable feedback on my posts, advice on writing on the internet, and camaraderie on this weird little adventure.

  • My friends Andrew (the aforementioned roommate), Ibrahim, Alek, Asara, Chris, Luis, Ryan, the other Chris, Alexander, Mason, Ali, Jared, Kyle, Connor, Max, Grant, and Grey for convincing me that my pet projects had potential and encouraging me to pursue them further.

  • Tony Kulesa, Alexey Guzey, and Aaron Stupple for sharing and endorsing my content when I had fewer than 500 followers and for giving me faith that someone was reading and enjoying what I was producing.

  • Jeff Bezos, Marc Andreessen, and Paul Graham for being an inspiration to me and for completely changing my world in two days.

1

Though our video interviews will still be on YouTube.

2

This idea came to me from reading Byrne Hobart and Tobias Tuber’s excellent paper on how Girard’s mimetic theory applies to financial bubbles.

3

Deepmind did return profits starting in 2020, but that mostly came from what Google paid it for its services, which can be a completely arbitrary number. It’s like me paying myself 2000 bucks to mow the lawn and saying it was a really profitable week.

4

Even car companies can say that the expensive fine-tuned innovations they make on their flagship F1 trickle down over the years into their consumer vehicles.

5

It is probably a bad idea to explicitly plan to fail at your first 3 businesses, but you get the idea.

Comment13
ShareShare

Create your profile

0 subscriptions will be displayed on your profile (edit)

Skip for now

Only paid subscribers can comment on this post

Already a paid subscriber? Sign in

Check your email

For your security, we need to re-authenticate you.

Click the link we sent to , or click here to sign in.

jhoang
Apr 23Liked by Dwarkesh Patel

Med school (not Caribbean) drop out here.

I think most Med students are seeking safety or prestige. Starting a business is inherently risky and not safe, and so not what they would want. You don't earn the prestige in startups until you are successful. In Med school, just getting in guarantees that you will be a doctor, and so you get prestige much sooner.

Probably a minority perspective, but I don't think they lack any third trait. I think it's mainly a matter of perspective. For example, I don't think you need the trait of being able to tolerate the unknown if you know that everything is fallible and so unknown. Another example would be that I would rather do something I find valuable rather than what other people find valuable which is denoted as prestige.

Expand full comment
Reply
1 reply by Dwarkesh Patel
Will
Writes My Bookshelf Runneth Over ·Apr 22Liked by Dwarkesh Patel

Med school admissions selects for conformity and willingness to jump through hoops. For a good med school application, you need to accumulate a few dozen hours of clinical shadowing, volunteer work, leadership work, etc. which are mostly not valuable in themselves but as signals of willingness to jump through hoops.

Expand full comment
Reply
1 reply by Dwarkesh Patel
11 more comments…
TopNewCommunity

No posts

Ready for more?

© 2022 Dwarkesh Patel
Privacy ∙ Terms ∙ Collection notice
Publish on Substack Get the app
Substack is the home for great writing