2025-08-03 16:38:31
Here are ten thoughts on AI that I personally find economically useful. Let’s go.
First: there is no AGI, there are many AGIs. That is: we empirically observe polytheistic AI (many strong models) rather than monotheistic AI (a single all-powerful model). We have many models from many factions that have all converged on similar capabilities, rather than a huge lead between the best model and the rest. So we should expect a balance of power between various human/AI fusions rather than a single dominant AGI that will turn us all into paperclips/pillars of salt.
Next, AI moves all costs to prompting and verifying. Basically, today’s AI only does tasks middle-to-middle, not end-to-end. So all the business expenditure migrates towards the edges of prompting and verifying, even as AI speeds up the middle.
AI is amplified intelligence, not artificial intelligence. Today’s AI is not truly agentic because it’s not truly independent of you. The current crop of agents can’t set complex goals, or properly verify outputs. You have to spend a lot of effort on prompting, verifying, and system integrating. That just means the smarter you are, the smarter the AI is. It’s really amplified intelligence, more than agentic intelligence.
AI doesn’t take your job, it lets you do any job. Because it allows you to be a passable UX designer, a decent SFX animator, and so on. But it doesn’t necessarily mean you can do that job well, as a specialist is often needed for polish.
AI doesn’t take your job, it takes the job of the previous AI. For example: Midjourney took Stable Diffusion’s job, and GPT-4 took GPT-3’s job. Once you have a slot in your workflow for AI image generation, AI code generation, or the like, you just allocate that spend to the latest model. Hence, AI takes the job of the previous AI.
AI is better for visuals than verbals. That is, AI is better for the frontend than the backend, and better for images/video than for text. The reason is that user interfaces and images can easily be verified by human eye, whereas huge walls of AI-generated text or code are expensive for humans to verify. See discussion with Andrej Karpathy here and here.
Killer AI is already here, and it’s called drones. And every country is pursuing it. So it’s not the image generators and chatbots one needs to worry about.
AI is probabilistic while crypto is deterministic. So crypto can constrain AI. For example, AI can break captchas, but it can’t fake onchain balances. And it can solve some equations, but not cryptographic equations. Thus, crypto is roughly what AI can’t do. See also these talks (1, 2) on how AI makes everything fake, but crypto makes it real again.
AI is empirically decentralizing rather than centralizing. Right now, AI is arguably having a decentralizing effect, because there are (a) so many AI companies and (b) there is so much more a small team can do with the right tooling, and (c) because so many high quality open source models are coming.
The optimal amount of AI is not 100%. After all: 0% AI is slow, but 100% AI is slop. So the optimal amount of AI is actually between 0-100%. Yes, the exact figure varies by situation, but just the idea that 0% and 100% are both suboptimal is useful. It's the Laffer Curve, but for AI:
Finally, here’s an a16z podcast with Martin Casado and Erik Torenberg where we go through all these ideas in depth!
Putting it all together, fundamentally, this is a model of constrained AI rather than omnipotent AI.
AI is economically constrained, because every API call is expensive and because there are so many competing models.
AI is mathematically constrained, because it (provably) can’t solve chaotic, turbulent, or cryptographic equations.
AI is practically constrained, because it has to be prompted and verified, and because it does things middle-to-middle rather than end-to-end.
AI is physically constrained, because it currently requires humans to sense context and type that in via prompts, rather than gathering all that for itself.
To be clear, it’s possible that these limitations are overcome in the future. It’s possible someone could unify the probabilistic System 1 thinking of AIs with the deterministic/logical System 2 thinking that computers have historically been very good at. But that’s an open research problem.
Some of these ideas previously appeared on X and YouTube. See comments here, here, and here. And the talks here, here, and here.
2025-07-28 09:24:26
All property becomes cryptography.
Let me explain why.
First, right now, trillions of dollars worth of digital gold is secured onchain. Bitcoin is now valued everywhere there is an internet connection. And no matter what political faction you're in, everyone agrees on the raw fact of who owns what amount of BTC.
Next, right now, the full legalization of stablecoins means that every other asset goes onchain. Because if there is a legal status for onchain currency, of course there's a legal path for onchain stocks, onchain bonds, and every other type of financial asset.
So, that's a lot right there. But there's actually a next step. As you can see from the video below, this door can also be secured with an onchain smart lock. That means the door to your house can be secured by crypto. So can the door to your car.
From Hideyoshi Moriya: My home also can be unlocked with Ethereum, by verifying a NFT (ENS) ownership. In this example, the system verifies if I have an ownership of piyo.eth @ensdomains.
Indeed, any door can be secured in this way. The door to a plane, to a train, to a boat, to a building. Any door can be secured onchain.
But this is really more than the keys to your car door. It's really the keys to your car itself. The digital signature starts the engine. And that means any piece of capital equipment, from cranes to drones can be similarly secured.
That includes the humanoid robots, the sidewalk robots, the self-driving cars, and just about anything that's controlled electronically. Which is almost everything.
There are exceptions. The food on your plate, the shirt on your back, those can't and won't be secured onchain. But that's actually a negligible fraction of value in the world.
For everything else, for 99%+ of what's valuable, for every financial asset and every capital asset, we will secure it onchain.
And the reason we'll have to do that is that the blockchain is the only truly secure backend. The Pentagon gets repeatedly hacked, as do many web2 services, but scaled public blockchains do not.
So: even the control plane for the drones goes onchain. The blockchain is the basis by which we can build a code-based order on the Internet, a new kind of global economic union that allows anyone with an internet connection to access world-class monetary policy and contractual equality.
All property becomes cryptography.
This post originally appeared on X. See comments here.
2025-04-19 00:35:39
Global meritocracy is finally here.
We’re launching a new fellowship for founders and creators, open to anyone from any country. You can apply online in ten minutes. If accepted, you will receive $100k in funding for a new or existing venture. Our main requirement is that you relocate to the Network School campus near Singapore with other fellows for one year to lock in.
If that already sounds amazing, go to ns.com/apply. Then read on for details on who should (and should not!) apply, what the process is like, and why we’re launching this.
You should apply if you’re a global tech founder, AI-first creator, or community builder.
Our ideal founder recognizes that the Internet is rising, that Asia is rising, and that technology has now fully decentralized1 out of Silicon Valley. We welcome applicants from any country, so long as you have English language fluency. As for the specific area2, what you’re working on matters less than how intensely you’re working on it. Some of our founders will be early in their careers (like recent high school grads or newly minted PhDs), while others will have more experience (like open source developers, research scientists, early stage startup alumni, or engineering professors). But we’re flexible, so long as you’re technical.
Our AI-first creator category is newer and deserves a bit more explanation. In short, we believe 20th century media is over. And we think every niche filled by the formerly mainstream media — Hollywood, newspapers, magazines, broadcast television, and the like — is up for grabs as never before, with production shifting from traditional venues to the global Internet. So, we’re looking for filmmakers, artists, and storytellers who want to make tasteful3 use of AI to hyperdeflate production costs and write new stories for the 21st century. We’re open to “traditional” creators too4, but we think the economic leverage from AI will be substantial5, so we want creators who think about AI as an integral part6 of their workflow.
Our community builder category is likewise new and also merits explanation. Briefly, we think the third kind of “thing” after tech companies and cryptocurrencies will be startup societies. That means intentional communities, pop-up cities, and network states, where digital tribes assemble offline in person. I’ve written about this at length in The Network State, but see this video and dashboard to see why we we think startup societies will be a megatrend.
OK, so that’s the type of person we’re looking for professionally — a founder or creator. But what are you like personally?
You’re Internet First, in that you broadly believe in the open web, decentralization, cryptocurrency, and encryption.
You’re classically liberal, centrist, or otherwise politically pragmatic7 — rather than ultranationalist or socialist.
You’d be in tech even if markets crashed for a decade, so long as you had time to study, learn, and explore.
You’re excited to relocate to our solarpunk campus near Singapore for a year to level up among others of like mind.
And you’re likely what we call dark talent — undiscovered greatness from the middle of nowhere8, whether that’s India or Indiana, Eastern Europe or the Middle East — who just needs a shot to prove yourself. And this is your shot.
So: you should apply9 for the Network School Fellowship if you’re looking for a sunny, green, and calm environment to build your dream amidst a community of builders. We’ve created a peaceful place with minimal distraction where funding, coffee, food, gym, office space, laundry, cleaning, and all the rest is provided so you can focus on work.
Now, let’s focus on the application.
Because we’re open to the whole world, we expect many to apply. So, we have a simple multi-stage process for obtaining a Network School Fellowship.
Form. First, fill out the form at ns.com/apply. We’ll review to see if you have either potential (as demonstrated by test scores and academic accomplishment) or results (as demonstrated by tech portfolio or startup traction). We’ll also administer a short screening test and ask for references.
Zoom. For those we deem competitive in the first round, we’ll do a structured video interview where we get to know you and ask a number of interactive questions. We seek people with high integrity and high agency, a combination that’s hard to assess through forms alone.
Room. Finally, we ask candidates who make it to the third round to fly out10 to our Network School campus near Singapore11 for an in-person, proctored, pencil and paper objective examination similar to the IIT JEE, the GaoKao, or the SAT. This offline component allows us to AI-proof the online talent search, thereby confirming your skills. During your trip you’ll also meet other Network Schoolers, see the facilities, and determine if you’re ready to build.
Boom. Then, if you’re accepted, you’ll relocate to Network School for one year to build. We’ll either fund12 your existing company or set up a new one. Existing companies will be backed in their current jurisdictions, while new companies will be incorporated by default in Singapore.13
So that’s the idea: a global startup draft, like the NBA draft, where we source talent from around the world and colocate it in person. We’re starting with 100 global founders in our first batch, and plan to scale up over time if things go well.
But why are we doing this?
We believe the old path for finding global tech talent — which went through Stanford14, Silicon Valley, San Francisco, California, and ultimately America — is ending. So we’re going to need a new path, based on pure global meritocracy. And we think the Network School Fellowship could be a starting point for this new path.
To motivate this, let’s talk about (a) what the path for tech talent used to look like, (b) why it’s now broken, and (c) how we plan to replace it.
Let’s break this up into the American and global paths.
The American path once started with a smart young kid in the middle of (say) Indiana like Scott McNealy. He’d take the SAT, get a great score, attend a top engineering school at a reasonable price, come to Stanford for his graduate degree, found a Silicon Valley tech company, scale it up, and eventually go public.
The global path once started with a smart young kid in the middle of (say) India like Vinod Khosla. He’d take the IIT-JEE, get a great score, attend a top engineering school at a reasonable price, come to Stanford for his graduate degree, found a Silicon Valley tech company, scale it up, go public, and eventually become a US citizen.15
Now, if you know anything about those two men, you’d know that they cofounded a little multibillion dollar company called Sun Microsystems. That is: they came from very different places, but thanks to the magic of global meritocracy, thanks to international capitalism, they could build something great together that pushed tech forward and benefited the world.
Yet their story is much less possible in today’s US. I won’t rehash every headline, but today’s Scott McNealy might well be blocked from matriculating by DEI, and today’s Vinod Khosla might well be blocked from immigrating on an H-1B. And that’s really just the start of it. Every single piece of the old path for tech talent is now either fully broken or under assault. To review:
Stanford. With anti-merit admissions, unaffordable tuition, student visa denials, grade inflation, ideological zealotry, AI fakery, lack of reproducibility, and budget cuts, US academia (including Stanford, Berkeley, MIT, etc) is in steep decline.
Silicon Valley. Still great people here, but the proverbial garage startup may need $250k just for compliance. High cost, high regulation, and high ideological conformity now means fewer crazy ideas — and much more consensus (and hence fundable, but perhaps not exceptional) work.
San Francisco. Even more extraordinary cost, rampant crime, out-of-control filth, anti-tech laws, and ideological anti-tech sentiment that’s escalated all the way to ritual destruction of self-driving cars and violent attacks on founders.16
California. A bankrupt state that can’t control fires, stop theft, or build public infrastructure — but that will still raise taxes and impose anti-tech regulations.17
America. Now institutionally hostile to skilled immigration, foreign investors, student visas, exports, imports, free trade, and foreigners in general, with unpredictable corporate law.
In short: the combination of surging anti-capitalism and anti-internationalism means it’s going to be much harder for global tech talent to come to Silicon Valley to build international tech companies. That path is now closing.
Fortunately, we still have the Internet.
Remember, there’s no silicon being mined out of the hills of Silicon Valley. There’s no natural resource keeping tech there. So in theory, given that you just need a laptop to build things on the Internet, we should be able to materialize the global tech community anywhere. But in practice, someone needs to actually do that.
And constructive proofs are always more satisfying than existence proofs. So, rather than simply assert that a better path exists, we must build it. And from a practical standpoint, we must also recognize that we can’t replace the old path for global tech talent in one fell swoop. After all, literally millions traveled that path, and will continue to do so for some time, even if it’s breaking. So we’ll have to start small, with perhaps just one hundred builders, to prove a new path for tech talent.
And thus the Network School Fellowship.
In a sense, our fellowship just does the “obvious” thing. It starts with the observation that tech has already decentralized out of Silicon Valley, and it gives a new coordinating point for global tech talent to assemble.
We put the first node of the Network School in the new Singapore-Johor Special Economic Zone because (a) thousands of millionaires have moved to Singapore, (b) >50% of the Earth’s population is within a short plane flight, (c) ~50% of GDP is now within Asia, (d) crypto has already decentralized to Asia, (e) AI is in the middle of similarly decentralizing, and (f) the zone has plenty of visas for global talent and a stated goal of recruiting 20,000 skilled workers.
In other words: we have capital here to fund tech companies, visas for tech talent, and the will to build. And with that will, we’ve already built our initial facilities and seeded the community. Which means we have everything set up for you.
And all we need is you. So, apply for the Network School Fellowship at ns.com.
For example, you should be aware of not just Tesla but BYD, not just SpaceX but ISRO, not just ChatGPT but Mistral, not just Anduril drones but also Bayraktars. That is: you should know about the best of Silicon Valley, but also the best of the world.
We have a list of areas we’re interested in at ns.com/fellowship. But in brief, applicants will usually be in software, hardware, crypto, AI, robotics, genomics, or really any other tech-enabled vertical.
What is tasteful use of AI? It’s not slop. It’s not text in ChatGPT voice, or humans with twelve fingers. It’s using AI to create something cool. This is a good example. Arguably this is as well. And much of the work from accounts like this, this, and this. For example, one of the best near-term targets for AI will be disrupting Hollywood by telling new stories they can’t or won’t.
In general, we think the workflow of an AI-first creator is highly technical and similar to a programmer, so in a sense you can think of this as a subclass of tech founder.
As one example: a traditional creator often builds an audience on the basis of their looks or charisma, which is fine. But an AI creator is more like a screenwriter, who can cast any actor or actress in any role at any time. This makes them far more flexible, and capable of generating many different kinds of content.
As another example, a creator like Mr. Beast spends millions on video production. Could some of that eventually get turned into AI, with tools like Runway ML? I’d be surprised if the Beast team isn’t looking at it. Of course, any given AI tool today won’t necessarily hit the bar to replace any given live shoot, but they keep improving. And we want pragmatic creators who’ll make smart decisions on when to shoot live and when to do it on the computer. Pragmatism means you don’t generate AI slop (at one extreme), but you also aren’t irrationally anti-AI (at another extreme).
For example, AI is already extremely useful for cleaning up raw audio transcripts, generating timestamps, and tasks of this nature. It usually requires polish, but it saves time.
As a first cut: if you like Lee Kuan Yew and Singapore, and believe in results over ideology, then we will probably get along.
Of course, if you come from “somewhere”, like New York or Tokyo or London or the like, feel free to apply. Dark talent is about your current position, not so much your current location.
Just as important is who should not apply. You should not apply for the fellowship if you’re seeking something purely recreational. Of course, we do have fun, but our events calendar is more like a research university, fitness bootcamp, or tech conference than a tropical resort. That is: the typical event is organized around learning technology, burning calories, and/or earning revenue — around self-improvement of some kind. So, if your idea of fun is working with robot dogs, running a 5K, hearing guest lectures from founders like Vitalik and Bryan Johnson, and meeting people of like mind — you’ll definitely have fun. And Singapore is right next door, so you can hit a global city whenever you want. But the focus is not Bali, but technology. Meaning: not eat/pray/love, but truth/health/wealth.
We may reimburse travel for applicants with true financial need. Just indicate this when you apply.
You can fly in via Singapore’s Changi Airport (SIN) or one of Malaysia’s airports like KUL or JHB.
A certain kind of extremely finance-focused guy will immediately ask “OMG on what terms!?!”
(a) The short answer is that this depends on the stage of your startup. If we're the first investor, we'll propose a competitive seed valuation. If you have existing investors, we can discuss.
(b) But the longer answer is that I have hundreds of startup investments and don't necessarily think a formula is useful. Network Schoolers already come from 90+ countries, and Network School Fellows will be at widely varying stages in life. Some will be 18 year olds from the formerly “third world” and some will be 48 year olds who’re founding their third startup. Startups are by their nature heterogenous, and the template of “Delaware C Corp operating in Silicon Valley, California with X check at Y valuation in Z round” is already broken in several ways, not least because Delaware, California, and Silicon Valley are broken. Moreover, international founders and web3 add all kinds of new complexity as Sam Lessin has also observed. Apply with your startup and we can figure it out.
(c) The still longer answer is: prices are just fundamentally lower in Asia than America, which makes comparison across time and space difficult. I have plenty of friends in Silicon Valley, but that place is expensive. By contrast, you can live for a year at Network School with room and board and gym and everything for just $1500 per month. A $100k check then gives 5+ years of personal runway, more than enough time to build. That makes the Silicon Valley garage startup viable once again…by going 10000 miles away from Silicon Valley.
(d) In particular, a $100k startup check on pure potential is absolutely life changing for the dark talent. For all those billions who could never go to Silicon Valley, not least because it’s increasingly impossible to get a visa to get there. This idea of helping under-appreciated talent rise is something I've written and spoken about for 10+ years. Anyway…I am now in a position to devote capital towards global equality of opportunity — and that is a big part of what Network School is about. We are for makers from the Midwest and the Middle East, for Chinese liberals and Latin American libertarians, for Southeast Asia’s rising technologists and Europe’s remaining capitalists, for talent overlooked everywhere from Indiana to India. If that resonates with you, and you want to be part of the pro-tech startup society we’re building, apply.
Unless there’s regulatory justification for another locale.
I don’t mean literally Stanford alone. I mean Stanford in the sense of American STEM-focused universities like Stanford, MIT, Harvard, and the like that graduate tech talent. Though of course Stanford by itself was a large part of that.
For decades, many talented immigrants essentially applied to college, company, and country as one package at age 21. For the US, they applied to do their graduate degree on an F-1 or J-1 visa, they then worked at a company on an H-1B or O-1 visa, and then they got membership in the country via permanent residency and ultimately citizenship.
Some people I like say SF crime is on the mend, but I’m skeptical. In any case, it’s in a jurisdiction where extreme anti-tech, anti-immigrant, and anti-free-trade sentiment is indisputably rising.
Some of these measures, like the anti-sharing economy and anti-AI bills were narrowly defeated in California, but only after hundreds of millions had to be wasted on politics. The point is that many California legislators dislike tech and want it to leave. Which means we need to find a new location. Think of it as Silicon Valley’s Ultimate Exit.
2025-02-02 23:17:34
The curriculum for Network School v2 revolves around the broad area of building startup societies. Sample topics covered include:
Early American history and the settling of the Wild West
The fall and rise of modern China, India, Russia, and Eastern Europe
The creation of modern Singapore and Dubai
Biographies of political figures from Deng Xiaoping to Robert Moses
The decline of Argentina and South Africa
The case-control studies of East/West Germany, North/South Korea, and Taiwan/PRC
Edge cases like San Marino, Northern Cyprus, Transnistria
The technology-fueled ascendance of El Salvador and Bhutan
National turnarounds like the Meiji Restoration and Boluan Fanzheng
The unification of France, Italy, Germany, India, America, and China
Ancient history from Gobekli Tepe to the collapse of Rome
The history of tech companies on the English and Chinese Internet
The general theme is building, collapse, and rebirth — but at the scale of cities and countries, not just companies. This is the next step for technology, because heads of network like Elon, Satya, and Zuck are now on par with heads of state like Trump, Modi, and Xi.
Here’s a draft reading list.
Macrohistory: Foundation, The Lessons of History, Unqualified Reservations, A Short History of Nearly Everything, The Selfish Gene, Albion's Seed, War and Peace and War, Stalin's War, Energy and Civilization, Who We Are and How We Got Here, The Ancient City, The Fourth Turning, The Grey Lady Winked
New Countries: Imagined Communities, Invisible Countries, The Origins of Political Order, How to Hide an Empire, The House of Government
Political Case Studies: A New Idea of India, China's Political Model, Seeing Like a State, Tomorrow the World, From Third World to First, When Money Dies, Deng Xiaoping and the Transformation of China, Collapse: The Fall of the Soviet Union
Regulation: Where is My Flying Car?, Three Felonies a Day, Reputation and Power, The Power Broker
Rebuilding: The Knowledge: How to Rebuild Our World From Scratch, How Innovation Works: And Why It Flourishes in Freedom, How to Invent Everything: A Survival Guide for the Stranded Time Traveler, A Connecticut Yankee in King Arthur's Court, Principles for a Changing World Order
Self-Improvement: Atomic Habits, Tools For Titans, The Almanack of Naval Ravikant
Tech: Zero To One, Hard Thing about Hard Things, High Growth Handbook, AI Superpowers
Crypto: The Bitcoin Standard, Broken Money, The Sovereign Individual, The Truth Machine, Daemon
Startup Societies: Communistic Societies of the United States, The Significance of the Frontier in American History, The Network State
So, the general idea is to begin with macrohistorical context, zoom in to the details of how societies rise and fall, and then move to the specifics of how you’d (a) first build up yourself and then (b) build up your new society.
We’ll also have in-person and remote speakers of the quality we’ve had at the Network State Conference 2023 and 2024, and at Network School itself. Our past speakers include world-famous founders and investors like:
So: the topics, the reading, and the speakers give a common vocabulary for Network Schoolers on why we should build new societies and how others have done so in the past. With that foundation, we will add more content over time.
2025-02-02 23:16:43
The first Network School went amazingly well, as you can see from the video above! More than 4,000 applications came in from 80+ countries for 128 slots, showing global demand for our vision of a startup society. So after the quarter ended, we took a beat to scale up our capacity.
And now we’re back.
Network School v2 reopens on March 1, 2025 on an island near Singapore with double the capacity for a year-long term. Our goal is nothing less than bootstrapping a startup society capable of bootstrapping other startup societies.
That means we’re seeking frontiersmen. We’re seeking dark talent. We’re seeking people who want to create win-and-help-win societies, focused on both individual and collective self-improvement. We’re seeking remote workers, digital nomads, online creators, personal trainers, event planners, self-improvers and technologists of all stripes. And we’re specifically seeking those who want to help our nascent community learn technology, earn cryptocurrency, burn calories, and have fun.
You should come if you want to build up yourself while also building a cloud community, formed from scratch from the global Internet. If that sounds awesome to you, go apply at ns.com.
Then read more below.
We’re building the Network School in stages.
Startup society (v1). First, we ran an experimental Network School cohort for 128 people in Q4 2024. For most attendees, it was a smashing success, but we did learn some lessons that we’re addressing in our v2 curriculum.
Society-as-a-service (v2). Now, we’re beginning a year-long Network School for 256 members starting on March 1. The most important difference from v1 is that we now know we have enough demand to build a permanent cloud community. So our #1 criteria for admission is finding pioneers1 who want to build with us over the long run.
Society from scratch (v3). In parallel with running v2, we are building a permanent Network School campus nearby to house thousands of technologists from all around the world to learn technologies, burn calories, earn currencies, and have fun.
Scale the school (v4). As we build the permanent campus, we’re using prefab construction to create a portable template that we can open source and recreate anywhere in the world. Think of it like clone-stamping a Hilton or Starbucks, all around the world, except with the blueprints in the public domain.
So, by the end of stage 4, if all goes well, we’ll have a distributed network of Network School (NS) nodes around the world:
Start your own society (v5). But that’s not the final stage. Many Network Schoolers will attend with the intent of eventually starting their own societies, using the social, digital, and physical templates we develop at Network School. We intend to fund the best. And when those networks are overlaid on the same map, we’ll get a community of friendly startup societies that looks like this:
So, that’s how we’ll bootstrap a startup society that helps bootstrap other startup societies. Now let’s review Network School v1, and what we plan for v2 and v3.
More than a decade ago, I created a visualization of cloud formations, shown below. The lower right corner refers to the theoretical concept of a cloud community: a startup society, materialized from the cloud, without the baggage of the old world.
The idea was to build an Internet First community, which every member joins of their own free will. A city in the cloud, the 21st century version of the city on a hill.
The first Network School was an experiment towards making that vision a reality. I knew we could bring thousands together for the Network State Conferences in 2023 and 2024, but that was only one day. Could we bring together 100+ people for almost 100 days?
Answer: yes! We could, and we did.
As the video at the top of this post shows, we pulled together accommodation, gym, coworking, auditorium, cafe, laundry, wifi, office pods, and hangout spaces for a 128+ person community, along with spaces for self-organized events. Some of it we built, some of it we rented, and some of it we refurbished to create a whole that was greater than the sum of the parts. Our overall retention was very high, with the vast majority of members staying through their booked period.
What specifically went well in v1, and what could we improve? Well, we set out to make it easy to learn technology, burn calories, earn cryptocurrency, and have fun, while also building our cloud community. And the parts we really nailed were burning calories, having fun, and building community. Here’s a video with members working out together:
Given that Network Schoolers didn’t know each other before the cohort, and came in entirely through the Internet, I do think it’s awesome that we were able to build a cloud community and thereby establish the fundamental question we set out to test in v1. Here’s a little gif of everyone having fun, including the Christmas Party towards the end of the cohort:
That said, we didn’t get everything right in v1, and in v2 we want to improve the learning and earning experience. Basically, in v1 we’d set out to do daily tutorials and bounties. But we’d also set out to build a real community, with men and women from different stages of life, with different skills, and different time commitments. These two goals conflicted. If we made the tutorials/bounties too easy, then it turned off the hardcore engineers. Yet if we made them too hard, it turned off the casuals who just wanted to drop in and have fun.
To be clear: we did have guest lectures where members learned from world-class founders and investors like Vitalik Buterin (Ethereum), Bryan Johnson (Don’t Die), Ryan Petersen (Flexport), Olaf Carson-Wee (Polychain), Haseeb Qureshi (Dragonfly), Eli-Ben Sasson (Zcash/Starkware), Jesse Pollak (Coinbase Base), and Peak XV / Sequoia India (Shailesh Lakhani). And as expected, many members earned remotely, worked on their companies, or even built startups together.
But we encountered the Karpathy bimodality: either you’re learning just for fun by watching guest lectures, or you’re heads down with pen and paper studying a specific subject.2 So, we’re fixing this in the next version by staging our curriculum for different skill levels and focusing it on building a new society.
And that brings us to v2.
As mentioned, the second version of the Network School opens on March 1, 2025 with capacity for 256 members and a year-long duration. This group will be the seed for a permanent campus we’re building nearby, which will house 1024+ members.
So, you should apply if you want to bootstrap a startup society that itself bootstraps startup societies. And the next version of the Network School is more explicitly oriented around this goal of society building.
Because that’s actually how many great universities were first founded. For example, Harvard was started with just a single building and a single instructor, to train future leaders of the Massachusetts Bay Colony.3 And the land grant colleges were built to teach frontiersmen how to run farms, dig mines, and erect buildings.4
Similarly, our v2 curriculum is oriented around the idea that the Network School is a mix of town and gown, both a startup society in its own right and a school that trains others to build startup societies. With that in mind, here’s what we’re planning for curriculum progression:
Lectures and reading. First, we’ll have guest lectures on startup societies with no prerequisites that the whole community can attend. These will be both in-person and remote speakers of the quality we’ve invited to Network State Conference 2023 and 2024. In addition, we’ll have a reading list and regular book club for discussion. You can see our draft curriculum here.
Workshops and hackathons. Next, we’ll organize workshops for two tracks of Network Schoolers: power users and programmers. Power users will learn how to use (and combine) the constantly shifting list of online tools, particularly in AI and crypto. And programmers will get hackathons on specific APIs and tools like Solana, Ollama, and Midjourney, with the goal of building something for their GitHub or Replit portfolios.
Courses and credentials. Over time, we’ll roll out AI-generated custom exams and cryptocredentials for particular courses that help build our society. The AI exams will come from turning public tech job postings into lists of skills and then exams, while the cryptocredentials will be NFT certificates awarded by proctors who conduct offline pencil-and-paper tests of these skills. But it’ll take us time to develop these.
Departments and placements. Eventually we want to organize STEM and humanities departments on an “Internet First” basis, with continuous education integrated with job placement. That means taking a software-centric approach to every discipline. For example, civil engineering would be centered around simulations, art would begin with AI generation and NFT monetization, and history would be combined with data science.
With all that said, I want to make four points that you should keep in mind before you apply.
First: Harvard wasn’t built in a day. Indeed, it was a single classroom entity at first with less than a dozen students! It took years of effort, many intermediate steps, and thousands of people to build it into a global university. Similarly, it will take time to go from guest lectures and reading lists to full-blown departments. So you should come only if you want to build the next Harvard, not simply attend it.
Second: the Network School itself is just a facilitator for bringing together great people. After all, you attend a conference for the talks, but mainly the community. You attend a college for the courses, but mainly the community. And similarly, you should come to the Network School for the content, but mainly the community, as you can see in this video:
Third: we want to pursue minimum necessary innovation in terms of new content creation, because there’s a tremendous amount of free online courseware that simply hasn’t been indexed and organized. A non-obvious point is that this rebundling5 of online courses is valuable in its own right, given the sea of information online. To navigate this ocean, we’ll focus on content that helps build startup societies. This criterion allows us to filter6 the vast Internet to determine what we want to teach and learn, and in what order.
Fourth: the Network School core team will organize as much as we can on a “centralized” basis, but the community will always be larger than the core team. So, you should come if you also want to teach and learn from other community members on a “decentralized” basis. These types of community-organized events were quite popular in v1.7
In short, you should apply to Network School v2 if you want to join a tech community. Think of it like moving to early Massachusetts with a bunch of positive-sum people while the construction of Harvard is in process.
Oh, one more thing about Network School v2. I’ve talked about the practical details of how we’re building the campus, and what to expect. But do we have something we want to prove in v2, like we did in v1?
Yes. At an individual level, v2 is similar to v1 in that it’s about allowing members to learn, burn, earn, and have fun. But at a collective level, if the purpose of v1 was to prove we could instantiate a cloud community, the focus of v2 is to build a paid waitlist that establishes the demand to build ever-more-ambitious cloud communities.
That requires a bit of explanation.
Remember how Elon memed the Model 3 into existence? He created a paid waitlist where 400k Tesla fans put up $1000 and waited literally years for the product. They did this because they wanted an electric car, because they believed in Elon, and because they knew how ridiculously hard it was to build something in the physical world. They took a calculated risk with Elon, and many of them also bought TSLA stock, and boy did he deliver for them.
That’s how we plan to build the Network School:
Elon’s waitlist established provable demand, and allowed him to negotiate with vendors to make the Model 3 a reality. Similarly, the Network School waitlist establishes provable demand that allows us8 to fund the land to make the Network School campus — and all future startup societies — a reality.
That brings us to our plans for Network School v3-v5. As noted above, we seek to build a campus from scratch (v3), open source the blueprints to scale it around the world (v4), and along the way encourage our attendees to start their own societies (v5).
Returning to that original visual, you can see how Network School v1-v5 form a progression, with gradually increasing scale and duration:
We’re already underway with the v3 campus, and we expect people who come to v2 to be ready to move there once residential units open up. Please note that this is a major construction project, and it may take a year or more to build v3, so like the patient Tesla waitlisters you should come with a flexible mindset.
One way to think about it: with normal real estate, you just rent an apartment and move in. But with the Massachusetts Bay Colony, the pioneers built their cabins from scratch by chopping wood. The Network School is of course much closer to moving in than chopping wood, but there is some expectation that you help9 build the community — by holding events, recruiting new members, engaging existing members, and scouting future sites — while the physical campus is being built.
Finally: I really like the facilities we’ve built, and use them quite a lot myself, but before you apply you should realize what we’re doing is frontier, not fancy. We are optimizing for editability, not luxury. So, while the Network School core team will attempt to handle as many issues as possible, you shouldn’t treat your stay as a luxury resort where the staff is at your beck and call, but as a startup society which we’re building together.
Another way of putting it: there are tech companies, tech currencies, and now tech communities. You know what an early stage startup looks like, and the milestones it hits before it becomes Google. You know what an early stage cryptocurrency looks like, and the milestones it hits before it becomes Bitcoin. Now we’re establishing what an early stage startup society looks like, and the milestones it hits before becoming something bigger.
Because the Internet now allows us to build societies from scratch, to go not just founder mode but founding father mode. And the closer you are to being a high-agency, low-maintenance technologist who just wants to work, work out, and build awesome things all day — while also hanging out with others with the same growth mindset — the more you’ll like founding Network School with us.
If that describes you, please do apply at ns.com.
We are particularly interested in tutors, recruiters, trainers, entertainers, coders, and builders.
You could actually make this more than a metaphor. Treat any aspect of starting a new society as an input to AI search, and then find courses online that relate to it. It could be at a high level (like the peopling of America) or at a low level (like the plumbing of a house).
Yes, Harvard was once a startup! It was opened in 1636 when the Massachusetts Bay Colony founded “New College” in Cambridge to train clergy. At first, there was just one building and a single instructor overseeing fewer than a dozen students. It was renamed “Harvard College” in 1639 after its first benefactor, John Harvard. The entire operation had very humble beginnings: just a handful of students, a very small faculty, and minimal facilities compared to the global university it would later become.
From Britannica: “The intent [of the land grant colleges] was clearly to meet a rapidly industrializing nation’s need for scientifically trained technicians and agriculturalists.“
As the saying goes, the only way to create value is bundling and unbundling. You unbundle CDs into mp3s, and rebundle into Spotify playlists. You unbundle newspapers into online articles, then rebundle into feeds. Similarly, we can unbundle traditional universities into online courses, and then rebundle them under individual tutors at the Network School, each focused on different aspects of building startup societies.
A curriculum focused on building new societies is simultaneously broad and specific, and also carries with it an implicit staging, because the knowledge that’s useful for a 100 person society is different from a 100,000 person society.
You can see a partial list of Network School community-organized events in this sheet.
Building real estate is like building automobiles, a capital-intensive and time-consuming endeavor. The paid waitlist model allows us to meme the Network School into reality, with concrete midpoints every step of the way.
We do want to figure out more “attachment points” over time where community members can help. At some point we’d like members to be able to contribute towards building the v3 campus, such that they aren’t simply electrical engineers but electricians, and not simply mechanical engineers but also carpenters. However, we need to figure out the logistics of that, because (a) construction is a specialized and regulated skill and (b) the NS core team can’t directly employ every (or even many) members of the community. What we can do in the meantime is give digital points of engagement: NS members can pitch in as tutors, recruiters, trainers, and entertainers, to help us learn, earn, burn, and have fun respectively.
2024-08-17 03:29:25
We got an island.
That’s right. Through the power of Bitcoin, we now have a beautiful island near Singapore where we’re building the Network School. We’re starting with a 90-day popup that runs from Sep 23 to Dec 23, right after the Network State Conference. Rent is only $1000/month with roommates or $2000/month solo. And we have plenty of day passes for visitors.
So, go apply online at ns.com! Then read more below.
As motivation, I’ve always wanted to expand equality of opportunity around the world. Because my father was born in a desperately poor country, but with the right opportunity he was able to make something of himself. Like dark matter, he was dark talent.1 And for more than a decadeI’ve been thinking about how to give others who are similarly situated the chance to make something of themselves.2 That is: I’ve been thinking about how to empower the dark talent of the world.
US universities used to fill this role, even imperfectly, and I loved Stanford when I taught there years ago.3 But the data shows they’ve declined in recent days. And they’re just not affordable or accessible to most of the world. So, it’s time for a new approach. And thanks to Saraswati and Satoshi, I have the resources to endow a new Internet-first institution: the Network School.4
The purpose of the Network School is to articulate a vision of peace, trade, internationalism, and technology…even as the rest of the world talks about war, trade war, nationalism, and statism. To revitalize democracy for the internet era, with digital polities and verifiable votes. To train the next generation to be not just leaders of companies, but inspirations for their communities. And to pursue truth, health, and wealth by leveling up our attendees personally, physically, and professionally.
Let me now describe in more detail how the Network School works, who it’s for, and how to apply.
The Network School is for people of all ages, not just the youth. And it’s meant to be lifelong rather than one-off, with both a structured and and an unstructured component. The structured part is about continuous daily self-improvement: learning skills, burning calories, and earning currency. Meanwhile, the unstructured part is about having fun and hanging out with people of similar values.
For short: learn, burn, earn, and fun.
The first part of the Network School is about learning technologies and humanities.
As motivation, the existing model of US undergraduate education is broken. You pay $100k+ for a four year degree, and then budget nothing for maintenance over the course of your life. It’s like paying $100k+ for a new car and budgeting nothing for maintenance.
By contrast, the Network School is about continuous education. It’s for remote workers, engineers, creators and digital nomads who want to integrate learning into their lives, rather than stopping everything to be a full-time student.
Here’s how that works. We set up mini-classrooms where you can drop in to see the problem of the day.5 You solve that problem and a proctor awards you a cryptocredential, a free non-transferable NFT sent to your crypto wallet that establishes “proof-of-learn.” Often your solution will involve putting code on GitHub/Replit (to show you understand a concept), or posting content to your social media profile (to show you understand a new AI tool). And over time, these cryptocredentials actually build up a cryptoresume proving what you know.
Our initial material focuses on founding tech communities, as distinct from tech companies. As such it touches on everything from crypto, AI, and social media to history, politics, and filmmaking. It should be useful even if you’re just growing a traditional company or building a following. Over time, of course, every branch of the sciences and humanities becomes relevant when building a community. So if this initial experiment works, we can expand branch-by-branch to build a new kind of university.
But we’re intentionally starting with something simple. Our learning is about continuous education, about solving the problem-of-the-day.
The second part of the Network School is about burning calories.
Longevity is important, but 20th century communities just aren’t physically set up to maximize physical fitness. Quite the contrary: the default mode of Western society is sedentary and sugary. Those who want to escape this need to roll their own nutrition and workout program, which takes time, money, and energy.
We’re changing that. I’ve teamed up with my friend Bryan Johnson to set up Blueprint-inspired food and fitness for the entire Network School community. Bryan will be on campus to set up the program, and then his designates will maintain it on a daily basis. Like many of you, I’ve been both fit and fat at various times, so this is a product I want to use myself.
Every member of the Network School gets a daily workout slot with a semi-personal trainer, much like a group fitness class. You run and lift in the morning at your chosen time, getting a proof-of-workout from your trainer. Your group holds you accountable for showing up. Then you get a box with your Blueprint-optimized healthy meals and head to work. The whole point is to provide willpower-as-a-service, where the community6 provides the discipline.
We’re starting with the basics of running, lifting, eating, and sleeping properly. The initial goal is to hit the limits of your genetics. But if all goes well, the biotech founders that come to the Network School will eventually help us all surpass our genetic limits, to live much longer than we otherwise would.
But again, we’re starting small! And so the “burning calories” part is about lifelong health and the workout-of-the-day.
The third part of the Network School is about earning currency.
We’ll have crypto prizes of the day for open source projects, AI content creation, and microtasks. There will be $1000 bounties every day for the duration of the program, similar to the various prizes I’ve posted on Twitter and Farcaster. And community members will post their own prizes.
Next, in keeping with the overall theme of self-improvement, we’ll have office hours to help with your job, your career, your visa status, and your funding. We’re much more invested in you than a typical college career center because our interests are aligned: the more you earn, and the stronger you are financially, the more you’ll eventually have to reinvest in the community.
Finally, we’ll have visitor hours with famous visiting technologists. As my friend Sriram noticed, when investors come through Singapore I typically do a podcast with them. Many will also visit the Network School to meet attendees, invest in them, or hire them. Others will give remote talks. And you can see the quality of visitors from the speakers in our conference and podcast.
So, earning is about constant career development and the prize-of-the-day.
All work and no play makes Jack a dull boy, of course. So the fourth part of the Network School is about fun.
This is the unstructured component. It’s most of what you’re here for. It’s just about assembling great people in one place: positive-sum people who believe in technological progress, internationalism, and capitalism. It’s your internet friends, coming from URL to IRL. Stanford introduced the concept of residential education, but this takes it to the next level.
In fact, our initial location is very similar to Stanford. It’s beautiful and sunny, and less than an hour from a major city (Singapore) with an international airport (Changi). That means you can be heads down during the week, head into the city on the weekends for fun, and get to just about anywhere in Asia within the same day. This is convenient for the >50% of the world that lives within the Valeriepieris circle.
We’ll do some group outings too, but most of the fun will be up to you.
Who is the Network School for? There are four lenses on this: demographical, ideological, professional, and personal.
As mentioned, our focus is the dark talent. The more respect you have for legacy institutions, and the more respect they have for you, the less suitable you’ll be as an applicant.
So: the Network School is for Indian engineers and African founders, for makers from the Midwest and the Middle East, for Chinese liberals and Latin American libertarians, for Southeast Asia’s rising technologists and Europe’s remaining capitalists.
It’s for everyone who doesn’t feel part of the establishment. But it’s definitely not only for tech, because a community does not run on tech alone.
Ideologically, the Network School is for people who admire Western values, but who also recognize that Asia is in ascendance, and that the next world order is more properly centered around the Internet — around neutral code — than around either declining Western institutions or a rising Chinese state.
For example, the Network School is for those who understand that Bitcoin succeeds the Federal Reserve, that encryption is the only true protection against unreasonable search and seizure, that AI can deliver better opinions than any Delaware magistrate, and that democracy can be rejuvenated with cryptography. It is for those who believe in technology, harmony, internationalism, and capitalism. It’s for those who want Silicon Valley without San Francisco. And for those who want to found, fund, and find not just new companies and currencies — but new cities and new communities.
Our ideal applicant is capable of remote work, or has enough savings to support themselves while at the Network School. For our initial cohort, we’re seeking three major groups of people in particular:
Writers, artists, influencers, and filmmakers
Trainers, athletes, coaches, and clinicians
Founders, engineers, designers, and investors
These are, roughly, the demographics focused on learning, burning, and earning respectively. Of course, if you fall outside those categories but still think you have something to contribute, you should still apply to the Network School.
I should mention that the Network School is a “product” that I built for the young version of myself — the aspiring young engineer. This is the community I want to live in: a technocapitalist college town, a Stanford 2.0 that’s globally affordable and genuinely meritocratic.
So, I’ll be on campus full time. Bryan Johnson and I are supervising the setup of everything from bench press to French press. And we’ll eventually be recruiting faculty in the form of content creators, fitness influencers, and angel investors for the learn, burn, and earn portions of our program respectively. But all that in due time.
Ok, so how do you apply to the Network School?
Just register at ns.com/apply, where we’ve set up a simple Luma page. That initial application takes a few minutes. Then, if you pass review, we’ll send a second application where you pay rent. As mentioned, our monthly rent is $1000 (with roommates) and $2000 (solo). We also have daily and weekly rates too, but short-term visitors still need to apply.
The rent gets you an air-conditioned room on a beautiful island, with internet, gym, and access to all courses and community services. You’ll still need to handle your flights and pay for your food, but we think the overall package is extremely affordable. And so the Network School could be an amazing option for individuals or small teams looking to save money, get fit, and level up while living in paradise.
I’m looking forward to seeing you there! Just fill out the application at ns.com/apply.
Join the discussion on your forum of choice:
We’ll answer questions and update the FAQ.
Dark matter is the undiscovered matter that tools like the Hubble Telescope let us see. By analogy, dark talent is the undiscovered talent that we need the “mobile telescope” to let us see. That “mobile telescope” is all the new mobile phones around the world, newly connecting people to the global internet and making talent visible.
See this 2014 interview where my collaborator Vijay Pande mentions our work on the dark talent. I’d been thinking about this since our online course with 250k students. It proved to us just how much untapped talent was out there.
As background, I once taught bioinformatics and statistics at Stanford, as well as an online MOOC in 2013 with 250k students. Before getting into startups, I thought I’d be a professor. I still admire what Stanford once was, but it’s in decline just like other US institutions. I do feel fortunate that I didn’t waste my life in academia like some of my peers. But I also want to build an Internet-first version.
Saraswati is the Hindu goddess of knowledge and Satoshi is of course the founder of Bitcoin. This is just meant to be fun, don’t read too much into it.
We recognize that you might need to focus on your remote work job for a few days in a row, so you can also catch up by doing several problems at a time.
In general, you shouldn’t need extraordinary willpower and discipline to be fit. But you currently do because (a) corporations have optimized the placement of hyper-palatable foods in every checkout line, grocery store, and strip mall and (b) modern dwellings aren’t built around the premise that everyone will be exercising every day. There’s more reasons too, but suffice to say that a subscription residential community with fitness as part of its premise should enable a different outcome.
