Balaji Srinivasan joins Sam and Shaan to discuss his project 1729.com — a newsletter that pays readers in bitcoin for completing micro-tasks — and the long-term vision behind it: a decentralized media ecosystem built on crypto oracles, cryptographic truth, and portable credentials. The conversation ranges widely across the history of tech vs. media, the concept of the network state, startup cities, the idea maze, and why Balaji believes math is more reliable than institutional science.
Speakers: Shaan Puri (host), Sam Parr (host), Balaji Srinivasan (guest, entrepreneur/investor)
Introduction and Guest Welcome [00:00:00]
Shaan: Before I introduce our guest, I’m going to butter them up a little bit — because I’m genuinely excited for this. There are probably three or four podcasts so far where I wake up in the morning excited to do it, and I know that if we do a good job this is going to be one of people’s favorite episodes.
Sam: It’s not morning — it’s going to be midnight soon where Balaji is.
Shaan: Yeah, we got up late, or stayed up late. That’s all right. It’s my birthday today, by the way. My wife was like, “What do you want to do for your birthday?” I said, “Don’t worry, it’s already planned. I’m doing a podcast.”
Sam: I love it.
Shaan: The reason I think this episode is going to be good is because our podcast is all about ideas. People listen because we get the wheels turning — we present new and interesting ideas, or talk about spaces in ways you haven’t heard before. One of the ways I get a lot of my ideas is from our guest, Balaji Srinivasan, who is here. Mark Andreessen at one point called you the highest ideas per minute of anyone he’s met. Do you know that quote?
Balaji: He said they were useful ideas too, which is good. People don’t usually call them useful, but that’s okay.
Shaan: Entertaining is fine too. We teach people how to launch apps. You warned the world of a pandemic. Your podcast is actually very popular — I’ve heard a lot of people say good things about it.
Balaji: Thank you.
Shaan: You’re one of my answers if I started a new Twitter account and had to follow five people to get information flowing in — you’d definitely be in there. Sam, I know you’ve been following as well. So we’ll give a quick rundown and then get into topics. We’re going to talk about crypto, some non-crypto things, and try to steer clear of too much controversy. Let’s play ball.
Balaji’s Background: Genomics, Coinbase, and Investing [00:02:00]
Shaan: So — at one point you were the founder of a genetics company. How would you describe that first company?
Balaji: Clinical genomics. Basically, one of the first companies to use genomics in medicine. If you’ve heard of Tay-Sachs or sickle cell — those simple Mendelian diseases you probably learned about in high school — we took all of them and tested at once with a universal genetic test. That was one of the first applications of genomics in the clinic. The scalability of it, the fact that you could assess many different markers at the same time, allowed you to cut costs relative to the traditional one-gene-at-a-time blood test.
Sam: Which I actually just used — my wife and I used it for freezing embryos. I saw the packaging in the trash and thought, “Oh wait, I know this name.”
Balaji: That’s what it was called — Counsyl. So yeah, Sean, that’s what you used it for.
Shaan: Genomics and biomedicine has sort of evolved on its own. It hasn’t really linked up with the main body of internet and consumer technology — yet. I think we’re going to start seeing more computation on large numbers of genomes this decade. That’s a whole separate topic, but happy to dive in if you’re interested.
Sam: What did you do after that? That was like a $300 million exit or something?
Balaji: Yeah, it was a good company. After that, I was an early investor in cryptocurrencies — bitcoin, ethereum, Zcash, and others. Early investor in Superhuman, Soylent, Cameo, Lambda School, Replit, Mighty, and a bunch of others that have done fairly well. I invest in things that help improve quality of life in the future I envision. You can view it through the lens of the unbundling and fracturing of all pre-internet institutions, and then the rebundling into new internet-native forms. I think very few institutions that predated the internet will survive it in the fullness of time.
Introducing 1729.com [00:07:00]
Sam: Before we really get into it, I want to call out your latest project right away rather than at the end. Sean, you want to take it?
Shaan: I’ll let you introduce it — but it’s a super interesting project. The URL is 1729.com. Simple idea, which I love: it’s a newsletter that pays you. You get a newsletter with these little micro-tasks. I got one recently — basically you can earn $3,000 in bitcoin if you create memes that promote crypto adoption in India. Little tasks come out and you can do them and earn.
Balaji: That’s right. And it’s not just interesting or fun for its own sake — the long-term plan is that you’re building up crypto credentials, not just cryptocurrencies. Badges in your wallet that reflect your ability to do certain things. Complete five Python problems, get a badge. Complete ten questions on Russian history, get that badge. And why does that matter? It totally disrupts the concept of a college diploma.
A Harvard or Stanford degree is a very coarse-grained illustration of skills. But if every single problem somebody solved was in a folder on GitHub — 25 courses, 10 problems each, 2,500 problems organized in folders — you could drill down and see their actual skills. This person knows operating systems. She knows compilers. This person knows machine learning, as evidenced by the 15 problems they’ve solved with source code online.
Once those badges are in your crypto wallet, you could log into a job board and just start working because you’re automatically qualified based on your credentials. Right now you can credibly declare your interests but not your actual skills. People can put words in their CV, but — as you guys know from hiring — that’s an unreliable narrator problem.
Shaan: So right now in its current form, it’s much simpler — it’s just, here’s something funny or useful, I’ll pay you money if you can pull it off.
Balaji: That’s right. And that’s interesting and fun and neat in itself. But the long-term vision is that if you’re building a company, anyone who meets the requirements can log in based on their task history and contribute, and get paid for it.
Sam: Everything you say sounds crazy but you’re right a fair bit. I’d like life to be easier — if we could just write you off as crazy, we could move on. But the guy’s right. So now we have to take each one of these ideas and actually take it seriously.
The Entropic Internet vs. Deliberate Learning [00:14:00]
Balaji: One big thing is that 1729 is a reaction against what I think of as the entropic internet. If you look at Hacker News or Twitter, you can’t unsee it once it’s been pointed out — those 30 links on the page are like a Jackson Pollock. It’s just guaranteed to fragment your brain into a thousand pieces.
The opposite of that is what Michael Nielsen has been doing with spaced repetition: you learn something, then you get reminders at one minute, one hour, one day — and with a little bit of spaced repetition it gets stored in permanent memory. You actually remember it.
Right now what we’re doing with the entropic internet is eating junk food. It feels good because of novelty-seeking — cool, that’s cool, that’s cool — but if instead you can be much more deliberate, and make one high-level decision to come to 1729, then we basically serve you content that strengthens you.
Shaan: I want to talk about media. You said something about East Coast media where the writers are the stars, and Silicon Valley where the engineers are the gods. Both can stink, both can rock. But you’re saying you’re paying attention to content now.
History of Tech vs. Media [00:19:00]
Sam: You built a genetics company, you were CTO of Coinbase, and the current product in its form is a newsletter — relatively low tech but high reach. Why did you go into a space where we debate the business model all the time?
Balaji: Let me give you the history. You can date the modern era of tech starting at 1995 with the Netscape IPO. From ‘95 to roughly 2008, tech and media basically ignored each other. Tech was happening but wasn’t front-page news. Google was just a search engine. Facebook was just a college social network. The iPhone hadn’t yet come out.
Tech also ignored media for a different reason: Terry Semel was at Yahoo and he tried to turn Yahoo into a media company, and for an entire generation of tech founders that seemed like a very bad model. Yahoo failed relative to Google. And so that whole generation of companies — Facebook, YouTube, Twitter, Dropbox — basically said, “We’re a neutral pipe. We’re just building technology.” User-generated content. That was pretty hard enough.
Then the financial crisis happened and the iPhone happened, and from 2008 to 2012, Facebook, Google, and Apple started going totally vertical. There’s a graph called “Print Media Disruption” — after the financial crisis, print media revenue dropped from something like $65 billion down to $17 billion while Google and Facebook were going straight up. Craigslist was also a factor.
Sam: When we started The Hustle we started out of Craigslist’s office. There were negative pieces on Craigslist. I never saw any stats on whether their classifieds were actually worse than newspaper classifieds, but it seemed motivated in part by the economic threat.
Balaji: Right. Craig Newmark effectively bought some peace off by funding journalism through the Newmark Foundation. I have nothing against Craig — I’m just noting the observation.
So from 2008 to 2012, tech suddenly came out of its box and came for all the segments that were previously East Coast: Madison Avenue, entertainment, all of it. And tech and media were basically allied through 2012 to get Obama reelected. If you look at the late 2012 coverage — “the nerds go marching” — that was sort of the high water mark of the tech-media alliance.
Then in 2013, after the election, the knives came out. Media was like, “You guys are getting too big. We thought of you as part of the coalition, but you’re coming for the entire enchilada.” And all the negative articles on “tech bros” — you can date it exactly to 2013.
Sam: And after that, for the next several years, tech got its head caved in.
Balaji: The tech lash. I warned against it because I could see it brewing and going viral in 2013. You could watch the radicalization pipeline of certain journalists in real time — their world was collapsing and they blamed it on tech. Tech didn’t try to do it, but Google, Facebook, and Twitter took away not just ads — they took away a lot of influence. What media corporations could do was write stories and shape narratives.
Sam: The difference between a story and a product announcement?
Shaan: One has a hero and a villain.
Balaji: Exactly. And as people have said, you can build a religion without God but not without Satan. While tech was busy doing pastel-colored product announcements — “Dropbox offers 10 more gigs!” — media corporations were busy lining up to paint tech as bad and evil. A lot of the older tech companies that didn’t have any ideological defenses got mentally colonized.
Sam: You said something on the Tim Ferriss podcast that I’ve thought about a lot: “Media scripts humans just as code scripts machines.” Can you unpack that?
Balaji: The way to get people to do something is sometimes to tell them to do it, but often it’s just to show it being done. That’s Peter Thiel’s citation of René Girard on mimesis. Humans are memetic — that’s how we acquire language, how we align on objectives.
Whether you want to call it contagious mental illness — I think that’s an understudied concept. We know there are contagious physical illnesses. Well, we’ve all stuck our brains into Twitter. Imagine putting your brain in a vat with 300 million other people and they’re all sending their brain states to you. It feels a little unhygienic if you think about it that way. We social-distanced from each other offline but packed close online.
Shaan: So the dumb version of what you said: I can convince people to believe certain things by making them feel certain emotions, through sticky viral content — which boils down to memes — and we’re all racing to stake our ground, and email is one great empire to build?
Balaji: Email is one piece of it. But within five to ten years, I think your ENS name — your crypto domain name — will become probably the third most popular communications channel after email and phone number.
The Vision for Decentralized Media [00:35:00]
Shaan: Let me summarize some of the interesting ideas. First: the pendulum — media ignored tech, tech ignored media, and now tech is going to realize the power of narrative. To do that they’ll need to flex different muscles: storytelling, memes, packaging. That’s one interesting trend.
Second: empirical truth enforced by economics. Media companies claim to be the truth — it’s corporate propaganda. But there are all kinds of hypocrisy underneath. Criticizing lack of diversity while not being diverse. Criticizing billionaires while being owned by generational billionaire families. And how do you get away from that? Through crypto oracles — the journalists and media companies of the future.
Balaji: I’d offer a few asterisks. The vision is oracles and advocates. Redefine a reporter as a sensor — a sensor that writes data on chain. That could literally be a machine taking a temperature reading and writing it on chain. In aggregate, if you had lots of them, you’d have a cryptographically provable record you could use in discussions about climate change.
But the first layer is the raw facts — cryptographic attestations from sensors. Those sensors could be lying to you. But they have a digital signature, so you can see their track record, and others can audit them.
Then the second layer is advocates on top of those oracles. These are humans with editorial judgment. Not just what they write, but what they write about — and the selection function is now visible. Editorial judgment is quantifiable because you have a layer of raw facts on chain, and then narrative becomes tangible because it’s which facts you actually cite in your story. And crucially, what was omitted.
Sam: So you’re going to hire 200 engineers to build this ledger of record, and 200 advocates to editorialize the facts?
Balaji: No, it’s decentralized media. What I want to do is invest in lots of individual diverse influencers — pseudonymous voices around the world. Everything that’s not the Brooklyn book contingent: voices that are underrepresented ideologically, demographically, internationally, and scientifically. Those are the folks I want to fund.
The tools I use for selection — the YC-style SAFE equivalent — I want those to be public. I want people to pick them up and copy them. I’m not proprietary about that.
The second component is to bootstrap the technology: figure out the right version of using crypto oracles in content. Maybe that requires a few years of compounding. Maybe I just write the articles on what that vision looks like and one of my portfolio companies figures out how to build it.
And third is writing the story. I have a book — The Network State — which I’ve been editing very carefully. I put the first chapter up at 1729.com: “How to Start a New Country.” The idea is essentially: when there’s no land, rather than try to take over a peninsula or declare yourself Sealand, you build a community in the cloud, and that community goes and takes pieces of land all around the world — cul-de-sacs, apartment complexes, networked together — until it has the scale of hundreds of thousands or millions of people and the footprint of a state, but discontiguous and networked on the internet.
The Network State and Startup Cities [00:48:00]
Shaan: Is there a project out there doing that — creating a community in the cloud and then having decentralized pods around the world?
Balaji: I think it’s a 20-year project. My one-liner: 2000s were tech companies. 2010s were crypto protocols. 2020s are startup cities. 2030s are network states. Company, protocol, city, state. And startup cities right now are at the same phase as bitcoin in 2010.
Sam: So what is a startup city? Is it Sam leaving San Francisco and moving to Austin? The mayor of Miami actively recruiting talent like a startup CEO?
Balaji: That’s right. The term encompasses three definitions. First: a city where startups happen — which is what San Francisco used to be. Nowadays, choosing San Francisco for a startup is like choosing Oracle for a new company. You’re locked into a horrible cost structure with declining product when far better options are available.
Second: a city that acts like a startup. Miami under Suarez. That’s amazing.
Third: a city that is a startup itself — like Prospera in Honduras, or the various startup cities that Pronomos VC is investing in.
Sam: Bitcoin in 2010 was a pretty good buy — and startup cities in 2021 is the analog?
Balaji: Exactly.
How Balaji Got Into Crypto [00:53:00]
Sam: Where were you when you first heard about bitcoin? How did the mind virus infect you?
Balaji: After the financial crisis, everything broke. You start thinking more deeply about the system you thought was functional, and it just shattered. I’d been thinking about the true definition of money as energy — in the sense of joules. Visualize an iPhone coming together from its rare earth elements, the glass, the aluminum, the silicon — how much energy would it actually take to assemble that from its constituent parts? If you could reduce the energy cost over time, you’d genuinely reduced the price. Joule deflation.
A friend of mine around 2010 or 2011 pointed me to bitcoin. I thought it was really interesting because it was also using energy to support value. But the most counterintuitive thing was that the extra computation was not going into increasing transactions per second but rather increasing the security of the system. That was a design decision totally foreign to everything we’d learned about scaling systems. And the fact that all blockchain addresses were public was just outside the box of what I thought a distributed system should be.
In retrospect these are genius moves. With the pseudonymity it worked, and then with Zcash and others we figured out how to add back that privacy.
It was actually the 2011 crash and recovery that convinced me. The coin dropped from $32 to $2 then recovered to around $10. That convinced me it had staying power — it’s extremely rare that something endures a 90-plus percent drop and then comes back.
In late 2012 I gave a course at Stanford where I talked about bitcoin. I said, “I’m not sure if bitcoin is the one that wins, but something in this class will be a big deal.” And that was around when Coinbase got going. Brian and Fred and I first met in January 2013, very early in Coinbase’s history.
Post-Economic Thinking and Pragmatic Idealism [00:59:00]
Sam: The Tim Ferriss podcast — Tim used the phrase “post-economic” to describe you. Money is beyond your concern. Where has most of your financial success come from — early crypto investing or building companies?
Balaji: I’ve done well from both, fortunately. As an investor you have 1% of the work, 10–20% of the return, and zero percent of the control. I’m not really a “tech guy” in the sense of having had a Commodore 64 early. I was a math guy growing up — Feynman lectures, physics. Computers were always a tool, and business was always a tool.
The reason I got into clinical genomics was that in academia, if you raise a $10 million grant and each row of data costs $1,000 to collect, you can only do 10,000 rows. Whereas if you’re making even $10 per row in a business, you can collect a much larger sample and make a much bigger difference.
Sam: What I was really asking is: did you take a practical path first, get the bag, and now you’re doing all the crazy stuff?
Balaji: I wasn’t capable of doing a job purely for the sake of the money. I really believed in both Counsyl and crypto and Coinbase. That’s how I could put in that level of energy. Buffett has this concept — if you’re waiting to do what you want, it’s like waiting to make love at 70. Just doesn’t work like that.
What I try to do is advance short-term goals on a winding path, but with a long-term vector I’m always moving toward.
Shaan: You embody that — extreme long-term north star, then today’s building block. 1729 is a good example: the north star is disrupting credentials and decentralizing learning, but the v1 crawl is just, “send a newsletter with a micro-task every day.”
Balaji: That is absolutely an acquired language. Moritz talks about this half-jokingly: every app needs one of the seven deadly sins. Tinder is lust. Twitter is wrath. Arguably early Google is sloth. You need to deliver a visceral instant benefit that can be explained in a second. “A newsletter that pays you” — earn bitcoin for submitting a form in a minute. Instant benefit. If you can’t give that instantly, it’s too complicated.
The Idea Maze [01:08:00]
Shaan: You’ve helped popularize the idea maze concept. Was that you who coined it?
Balaji: I coined the term. You never know what’s going to resonate.
Shaan: The Steven Spielberg version of startups is: founder wakes up, eureka, straight line to success. My reality was: I’m interested in this problem, it’s all foggy, I sprint, I hit a dead end, I backtrack three steps, I go down another fork, I hit another dead end, I wander around, and eventually I come out the other side. And that’s actually a normal process.
Balaji: That’s quite good. A few additional thoughts. One: it’s similar to mining for gold. There’s a quote about San Francisco during the Gold Rush that sounds exactly like startup culture — everybody had crazy plans, unlimited financing for any gold mining operation. The idea maze is: which path leads to the pot of gold?
Two: the maze is time-varying. Pandora pre-iPhone was only an okay idea. Post-iPhone it became a billion-dollar company — the iPhone gave it the lift it needed. Very small-seeming permutations can radically change outcomes. Twitter versus Facebook: Twitter was default public and asymmetric. That meant it was one connected graph. Fundamentally different experience.
Three: the specific idea maze you decide to run really does matter — it can be a 1,000x multiplier. Genomics was a good idea and I had all the right arguments and the declining sequencing cost curve, but that generation of companies topped out somewhere between a few hundred million and a billion. It was the Zipcar era when the Uber era was still coming.
Substack two or three years ago seemed like “all right, just another blog, who cares.” But adding two features — monetizability and newsletter delivery by default — those changed things completely. Ben Thompson of Stratechery had been building his own thing for years when nobody cared, and then it actually worked. Part of it is also the decline of the existing East Coast media environment. That made this thing that was okay here into an amazing idea, because suddenly there were all these nails that needed a hammer.
Teleport, Digital Nomadism, and Relocation [01:20:00]
Sam: You built Teleport around 2014 — it was going to aggregate tons of data and tell you the best place to live. What happened?
Balaji: The concept was that every quality-of-life metric could be implicitly assigned a dollar value. How much more is a place worth because it’s warmer? Out of millions of choices like that, you can back out the function of how much people value that extra degree of warmth.
You layer on top: cost of living, how much you’d make in that geography, bars and nightlife, schools, airports, rule of law, what’s legal — and you have a complicated mathematical optimization problem. The best place to live would basically be a search engine that tells you the maximum value. By changing your X/Y location you might gain $50,000 hedonically — maybe $30,000 economic and the rest in quality of life.
It was a fine outcome. But it was a little too early. It’s similar to the idea maze — Teleport was the Zipcar era. If it had been started in 2019 or 2020, during the pandemic, it’d be in exactly the right moment. Nomad List by Levels.io was working on the same theory at nearly the same time, and he’s done great with it.
Post-vaccination, the world is going to see a huge wave of international migration. When you’re cooped up in a room, you want to burst free. Tools for international relocation are still early. If anyone wants to build Teleport 2.0 or Nomad List 2.0, DM me.
Shaan: Where do you most want to live right now?
Balaji: Have you heard the term entrepôt? A commercial center. I think those are waxing in importance. Dubai, Monaco, Singapore — and Miami, which is actually the Singapore of Latin America. If you’ve got a Mexican national and a guy from Paraguay and they want to do a deal, they’ll do it in Miami. Good banking, rule of law, relative currency stability. It’s an international business capital that many Americans just don’t know about.
Also worth noting: covert has made longitude rather than latitude the organizing principle. If someone’s not in the office, the next question is: what time zone are they in? So longitudinal arbitrage — you can go to Argentina, have low cost of living, and still be aligned with US time zones.
Geographic Proximity vs. Cultural Proximity [01:29:00]
Balaji: There’s a concept: geographic proximity no longer leads to cultural proximity. Somebody in an apartment building often doesn’t know their next-door neighbor — 50 feet away, closed doors, completely siloed. But who do you recognize? The person 3,000 miles away you’re laughing with on video chat. Your social network is scattered around the world.
Shaan: I buy into that completely. I’ve got so many friends I consider dear close friends who are just internet friends.
Sam: Soon we’ll just drop the “internet” part.
Shaan: At this point the majority of my friends are internet friends. That’s just called “my friends” now.
Sam: But I want to challenge that belief, because we’re all in our bubble. I’m from Missouri. My family lives there. One’s a teacher, her husband works at a bar, their kids play softball on Saturdays. Do you ever ask yourself whether the ideas you’re coming up with really only impact a small sliver of people? How are they going to impact the random person in Nebraska, or someone in Kenya?
Balaji: I absolutely do. The v1 goal with 1729 is to build a community of technological progressives — not just tech founders but media founders, influencers, writers, activists for technological progress. People like pro-nuclear activists or folks pushing for faster drug and device approvals. Build consensus among that group that technology is good, that we can work across borders toward a better future.
Then v2 is: how do we build positive-sum user interfaces for folks who may not be interested in the world of ideas? I mean “user interface” in the broadest sense — how your house works, how your car works, how your community works. How do we build things where the founder or CEO gains or loses just as the average citizen does? Quantifiable alignment.
Crypto Social Networks and Decentralized Social Graphs [01:38:00]
Shaan: Let me transition to a topic that’s a little controversial but exciting: crypto social networks. We talked about BitClout on here and got a lot of flack. The interesting things I saw: number one, the growth hack — they used economic incentive to bring high-status individuals to a new network. Number two, you can buy someone’s coin and benefit from the rise of their reputation — a curator gets rewarded alongside the creator. Number three, they took Facebook’s most valuable thing — the social graph — and put it out there as the product, with economics built in, so any social network can be built on top. What do you see in these crypto social networks?
Balaji: Disclosure: I’m a small early investor in BitClout. And I do think it’s a good idea overall.
First: crypto is the sequel to open source. It’s not just open source — it’s open state and open execution. Not only do you have the source code, you also have the database, and you can track every single op code when it’s executing. With Twitter and Facebook, they have the API and they can shut it off. Meerkat, Teespring, Zynga — so many companies that built on top of those APIs had their access throttled or their business model destroyed overnight.
The issue was: if Twitter let the API be completely open, somebody could clone the entire Twitter, sell ads against it, and Twitter would just be an API provider with no way to monetize. So they closed it. Crypto enables a third model: open state and open execution. Web3. We’re moving into a phase where the database itself is open and anyone can build on top.
How Balaji Thinks and Works [01:47:00]
Sam: How much of your day is spent on this crazy thinking, and how much is actually building?
Balaji: What I have to spend energy on — what’s actually mentally taxing for me — is small things like remembering calendar appointments. It’s like a sports car on a normal highway. You can do it, but you have to watch out. I’ve just built technological workarounds to help with that.
As for the ideas — I just can’t stop. I walk around thinking, “Oh, that’d be cool, this would be cool,” and tap it into my notes or tweets. I’m not trying to humble-brag. Feynman had this concept I think about a lot: he had a bunch of active problems at any one time. How do decentralized social networks work? How do you build alternatives to media corporations? How do you establish decentralized truth? And then against those active problems, I match solutions. I read something and I go, “Okay, this maps to that.” A hammer finds a nail.
Shaan: You also helped popularize the idea of hiring from Twitter rather than universities.
Balaji: I don’t really hire from Stanford, Harvard, or MIT anymore. I hire from Twitter. First, it’s international. Second, you can see what these folks are actually about — their writing, their interests, their values and character. There are super smart people in Nigeria, in the Middle East, who shine through. And you’ve got an index on the world that you didn’t have before. Whereas a lot of graduates from certain elite schools are coming out with what I’d call a religious education — that’s what they’re prioritizing — and a set of priorities and entitlements that are often adverse to what you need to build a business.
What Balaji Got Wrong [01:54:00]
Sam: What predictions did you make ten years ago about 2020 that were wildly wrong?
Balaji: I very much underestimated social networks. I only became a public figure in 2013. Before that I was cloistered in academia and in genomics.
I understood the appeal of putting a “facebook” — the physical booklet given to freshmen — online. But I only really understood the power of social networks when a friend of mine was live-tweeting a genomics conference. I realized: that saved me a flight. No journalist would have written up the material as well as this scientist did. And I was able to download five hours of presentations without a plane flight, in 10 minutes. That was a lightbulb for me — I only underestimated social networks once I realized they weren’t just time wasters but could actually spread useful knowledge.
I also probably overestimated genomics in that decade. I had all the technology arguments right — the declining sequencing cost curve looked like Moore’s Law — but what I didn’t understand as an academic, and only understood as an entrepreneur, was what the FDA really was. Basically the primary blocker of medical innovation, not just in the US but in the world. People are now catching up to this because of the pandemic — the FDA holding back authorization on COVID testing, holding back the J&J vaccine because literally one in a million people were getting blood clots. People are understanding that their degree of risk aversion is not actually optimal. There are both type one and type two errors. If you’re rejecting good drugs, good vaccines, or good tests, you may be doing as much or more damage than approving bad ones.
Shaan: And that’s why you didn’t take the FDA position?
Balaji: The paradox is obvious. Google didn’t reform Microsoft by becoming Bill Gates. Google reformed it by building a parallel system, making billions of dollars, and eventually Microsoft reformed under Satya and embraced open source. I don’t believe one can abolish the FDA any more than one can abolish the Fed. But you can exit it — the same way we exited the Fed with bitcoin.
Crypto as Decentralized Truth [02:04:00]
Balaji: What is to our existing institutions what crypto was to finance? You can already see crypto law with smart contracts for big chunks of law. You can start to see crypto media in the sense of decentralized cryptographic truth — the crypto oracles I was talking about.
Can I dive into the quote-unquote “science” issue for a second?
Shaan: Let’s do it.
Balaji: You’ve seen all these people saying, “Masks don’t work because of science,” or “Masks work because of science.” What they’re actually saying is: some authority figure has told me this is true, therefore it is true. But science is not about peer review — science is about independent replication. Peer review is at best a proxy for independent replication. Most of the time peer review is a few emailed comments and “make a new figure.”
Science is the ostensible basis of our civilization — you don’t cite God or the Ten Commandments for policy, you cite science. But when that upstream scientific premise is subject to institutional capture — when public health officials admit that saying “masks don’t work” was a noble lie, which wasn’t even noble, it was just stupid — you can’t do that too many times before people just stop trusting it.
What’s the alternative? The answer isn’t astrology or woo. There’s only one thing more prestigious than science. What is it?
Sam: Math?
Balaji: Math. Every scientific paper is based on data and generated figures and tables. There’s a concept called reproducible research that says a paper should be fully reproducible from the underlying data and code — you should be able to hit enter and have it run. Sites like Distill.pub, or what OpenAI does with interactive paper websites, operate this way.
Here’s the point: every single scientific paper rests on calculations. The more of that independent replication process you can turn into math, the more people who don’t trust a central authority can check the claims. We’ve already seen a huge victory of this: all those Nobel laureates who denounced bitcoin — it can’t work, it shouldn’t work, it should be banned — mathematical economics defeated quote-unquote scientific economics.
What’s awesome about crypto is this: when I look at blockchain.com and see a transaction confirmed on chain, there’s a thing that says “number of independent confirmations” — how many different miners have confirmed that A did B at time Z. Independent confirmations are very similar to independent replications. They’re all doing mathematical calculations to confirm a fact.
If the code and data set for every scientific paper were on chain, not only could you independently replicate at least their post-upload conclusions, but you could check every calculation, every figure, every citation. You could map the trail of citations all the way back. That’s actually very difficult today — only Google can build Google Scholar because they have access to all these journals behind paywalls. Crypto unifies it in a universal address format.
That’s a recipe for what crypto science, or decentralized science, might look like.
The World Becoming More “Biology-Like” [02:20:00]
Shaan: You said the world is becoming more “biology-like.” What does that mean?
Balaji: From nine-to-five work to tweets and surges of activity. From day people to night people and async work. From status quo bias to disruption bias. From the physical world to the informational world — ideas matter more, ideas can shape the world more.
Soylent is a good example. I like the company. I prefer a keto version. But the concept of not having to think about food so I can just do more math or more programming — that’s very biology-like. And I recognize there are better foods than Soylent. But many people who are busy don’t eat well, and if the alternative is a candy bar, Soylent improves baseline health.
The things I invest in fit my preferences and the world is sort of adopting them. That’s hard to clone.
One asterisk: a lot of businesses do not fit my preferences. Instagram and selfies? That’s something you do, bro.
Sam: You’re supposed to be this big shot crypto entrepreneur and you have zero lighting, a super old camera —
Balaji: You’re right. I’m solving that. I’ve got a studio I’m building out. I go zero or a hundred on something, so I use the defaults until I actually engage. The reason I’m going to improve audio and video dramatically is that I need the content to tell the story to build a future. That’s the dot product with my north star. Before that, I just felt like an exhibitionist and it wasn’t my thing.
Advice for 21-Year-Olds [02:27:00]
Shaan: If you were 21 today, what would you focus on?
Balaji: Skills-wise: get really good at computer science and statistics. The reason is that every domain has algorithms and data structures. Computer science and stats are valuable anywhere — you can walk into Walmart, American Airlines, or Pfizer and start writing code for their specific problems. CS and stats are the universal language. I don’t mean just learning to invoke library calls — I mean actually understanding the concepts.
CS and stats is to this century what physics was to the early 20th century. Physicists could walk into any discipline and just write down equations and change it. Now, because so much of the world is a constructed virtual world — cryptocurrency, social networks, video games — CS and stats becomes a general-purpose foundation in any area.
Then in terms of domains: genomics is big. Robotics is big. Crypto is obviously big. An underappreciated area is relocation — permanent international migration. Digital nomadism is going to be huge. There are a thousand pieces of the tool chain for allowing people to be more mobile. Post-vaccination, I think you’ll see a great migration.
I wrote about this in 2013 in an article called “Software Is Reorganizing the World.” Because the internet is breaking geography — you’re closer to someone 3,000 miles away than your next-door neighbor — the underlying premises of the city and nation-state are being invalidated. Geographical proximity no longer leads to cultural proximity. What’s going to result is cloud communities eventually migrating and materializing into the physical world and forming cloud cities, and eventually cloud countries.
Closing: Science vs. Math, and What Balaji Does for Fun [02:38:00]
Sam: You were a partner at a16z — you’re just a weirdo and I love it. You’re so interesting. You’re just sitting here talking about math and science and crypto and where to live.
Balaji: We can go deep on a specific topic if you want. Actually, I’d love you to teach me something about newsletters. You built a large one. Given that I’ve got a newsletter that pays you, maybe you can teach me something.
Sam: I could teach you everything about newsletters. You know what you are to crypto — I am to newsletters. It’s unfortunate that crypto is far more lucrative, but I’m doing okay.
Balaji: If you just stay in the game long enough and you take enough swings, ones you don’t expect will connect much harder. Investments I’ve made where I spent five minutes have returned more than five months of work. You’re in the game — just put some of that newsletter money into crypto or other things.
Sam: Warren Buffett is brilliant at investing and also drinks six Cokes a day and eats McDonald’s every morning. Do you have something like that — the dumb thing you know is dumb but you do anyway?
Shaan: We both watch shitty reality TV.
Sam: Big Kardashian fan.
Balaji: I’m on Twitter constantly. That’s certainly a huge waste of my time. It’s one of those things — every time you get out, they pull you back in. I also watch tons of movies. Though less so recently — so much of Hollywood has gone so woke that I just roll my eyes knowing exactly what it’s going to be. The quality has been dropping off. The characters are hackneyed, the messaging is on the nose, so much is off limits that the humor is edited out.
Sam: What’s your Hollywood replacement?
Balaji: Old books. Math textbooks. And I’ve actually gotten back to working out. One thing I did at 1729 is “proof of workout” — you submit proof of a workout, we give you $10 in BTC if it’s in the top 100 that day. What I realized is that for the next thing I do, I actually had to make fitness part of the job, not a sacrifice to the short-term. Spaghetti code is a short-term optimization where you’re taking on technical debt. In the same way, if you’re not working out and eating right, you’re taking on physical debt.
I think daily fitness is on a straight line with transhumanism and reversing aging. I want to set that direction vector.
Outro and Where to Find Balaji [02:47:00]
Shaan: I think we lost Sam to a bathroom break — we’ve broken the My First Million record for longest podcast. We should wrap up. Where can people find you?
Balaji: Go to 1729.com. Free newsletter that pays you in bitcoin. Regular bitcoin bounties for tasks and tutorials — learn about how to register a crypto domain name, do a tutorial, tweet about your experience, earn $250 if you write a good set of educational tweets. We’re giving out $5,000 in Ethereum for that task right now.
Sam: Could other people sponsor tasks — put up five grand for something?
Balaji: Absolutely. Go to 1729.com/create. I want to eventually have companies with 1729 accounts where every task they can outsource to the public at large, they do, and maybe pay in their corporate cryptocurrency.
Sam: I love this mini X Prize concept. I’ve been talking about this on the pod — the X Prize got a bunch of people to innovate and step up, and more investment came in than the prize itself was worth. I wish more companies would put things on their roadmap out there and say, “If anyone builds the best version of X, we’ll acquire it for this amount.”
Balaji: Absolutely. One big thing I want to do is prizes for startup ideas — $25K for the best one in 30 days, with the right to put in a SAFE or something like that. If you want to try something, go to 1729.com/create and submit a task. If it’s good we may even fund it.
Shaan: All right. Thank you for everything.
Balaji: Thank you. I appreciate the conversation. You guys are too self-deprecating — your show is great and your feed is often insightful. Mutual admiration. Great job.