Papertrade Launch, 1000x Leverage & Onchain War Stories - blurr
Paper Trade replaces the traditional casino bankroll with a tokenized queue intended to return the house’s structural edge to participants. When the bankroll is insolvent, winning P&L becomes debt that later trader losses repay; when it exceeds its adjustable limit, the surplus can be distributed pro rata to stakers. blurr’s objective is to “externalize the profits that market makers and exchanges receive” back to users.
Transparent 1000x exposure is possible because Paper Trade reads Hyperliquid’s BBO directly on-chain through HyperEVM precompiles. At 1000x, a move beyond 0.1% can liquidate one side before the winning side offsets it, so volatile conditions initially restrict positions to BTC and ETH; blurr suggested SOL might work closer to 300x. The selling point is inspectability: “I specifically designed it so that you can see how it works.”
The hardest product problem is ordering transactions fairly when tens of thousands of users arrive together. Paper Trade routes activity through relayers and an internal mempool-like scheduler, prioritizing liquidations over closes and closes over opens; direct contract access would otherwise become a gas war in which interface users are “lambs to the slaughter.” If HyperEVM pauses while prices move, Paper Trade may freeze the market at its last price rather than let the first returning block become a race for free money.
The trading economics combine zero slippage and no funding with hard limits on scalpability and system exposure. Positions execute around the Hyperliquid BBO, but size restrictions prevent infinite positions; Paper Trade charges 1% on profits and losses, while its Rollbit-inspired payoff curve imposes a heavier charge on tiny moves. A $6 million winner against a $5 million bankroll would receive $5 million and enter a queue for the remaining $1 million.
Token emissions are strongest when the bankroll most needs recapitalization, creating a potentially unbounded but economically decaying supply. blurr could not recall the initial figure precisely, saying it might be $100 or $1,000 per paper before tentatively describing $1,000 for every $1,000 of paper losses; above roughly $2 million of bankroll, emissions ratchet downward toward zero. The token owns house commissions after a gas allocation, must be staked for emissions, and will not be tradable on day one: timing depends on how the live system behaves.
Paper Trade evolved from a playful mechanism into infrastructure designed around the possibility of 20,000 users on day one. blurr repeatedly attacked his own design—“How would I break it?”—including removing administrator-level attack paths, and sees failure as either a technical breakdown or nobody finding the product useful. His broader ambition is for others to inspect and iterate on the queue, because “it is a very new and unique mechanism.”
blurr’s larger crypto thesis is that “the purpose of a system is what it does,” and what crypto demonstrably does is move value safely online. Thread Guy’s examples included Bangladeshi click-farm workers using cheap USDT on Tron amid local-currency erosion, and the difficulty of splitting a restaurant bill in the United States before Venmo and Cash App. blurr’s conclusion is that financial adoption should precede attempts to use crypto for broader political or social goals.
His career—and his warning to younger builders—centers on taking asymmetric risks before the edge becomes institutionalized. RuneScape led to CryptoKitties genetics, Compound’s ticking 18-decimal balance, Uniswap sniping, three hours as effectively Curve’s only farmer, and a $13 million on-chain incident; today, venture capital absorbs many would-be “pirate developers.” With conventional labor buying less freedom, he thinks young people may still need to “flip the coin,” even as interconnected technology makes civilization feel like repeated heads until one terminal tail.
1. Paper Trade began as a mechanism that finally became buildable
blurr and Jez had worked together for roughly five years, but neither had independently built a complete product from scratch. New tools collapsed work that once required “a team of seven or eight people,” particularly interface and infrastructure tasks outside a founder’s core expertise.
The seed was an article, apparently written in summer 2024, about a queue-based casino bankroll. They first modeled deterministic games such as blackjack and coin flips, where outcomes were easier to express, but concluded those products were not interesting enough to launch.
Asked why accept the reputational downside when success might barely change their lives, blurr offered no convincing grand mission. He emphasized that the risks could outweigh the benefits: if everything went well, they might be no better off, while failure would hurt both of them. He then used the maxim, “A ship in the harbor is safe, but ships are not made for that.”
Jez’s tweet announcing the project drew more interest than expected. blurr realized that 20,000 users on the first day would require a fundamentally different product, so they redesigned around that possibility and around every way traders, administrators and infrastructure could break it.
2. The queue removes the house bankroll without removing the house edge
A conventional casino takes the other side of every bet and therefore needs enough capital to survive winners; its bankroll determines maximum bet size and creates a severe cold-start problem. Paper Trade instead pairs the system with an emission curve that rewards users more aggressively when the house is least solvent.
The incentive is intentionally counterintuitive: users who believe the house will be profitable over time may accept losses because those losses effectively buy into the system. “You encourage losing money to effectively buy into the system.”
blurr’s industry map is blunt: market makers and exchanges are the two crypto businesses that earn substantial money, and most of that money ultimately comes from users. HLP- or LLP-style liquidity pools gain specialized advantages, particularly liquidation flow and related fees, that ordinary market makers cannot easily reproduce.
Paper Trade keeps an exchange-like edge but tries to return the advantage it creates to users, including users who lost in the system, rather than leaving hidden economics with outside market makers. blurr’s framing is to synthetically put everything on the exchange while making the resulting house advantage flow back to users.
3. Hyperliquid’s on-chain BBO makes inspectable 1000x trading possible
Paper Trade reads the Hyperliquid BBO as its oracle through HyperEVM precompiles, rather than importing an off-chain price feed. blurr described one HyperEVM block and three HyperCore blocks per HyperEVM block and emphasized that the oracle can be read directly from the chain.
That access was decisive: other oracle systems become a product’s “stumbling block,” introduce authorized operators and theoretically expose keys or update processes. Hyperliquid can itself be manipulated, but blurr said its risk parameters are intended to ensure that manipulation is not profitable for Paper Trade.
Thread Guy pressed on why this architecture is not standard. blurr’s answer was technical difficulty: Hyperliquid’s precompiles made the approach possible, whereas many existing 1000x venues are effectively internal systems whose mechanics are not publicly visible.
The leverage remains dangerous despite the oracle. At 1000x, a move beyond 0.1% can liquidate one leg before the winning leg offsets it, leaving the house exposed. In volatile conditions, Paper Trade offers only Ether and Bitcoin; blurr said SOL might eventually support something nearer 300x, subject to much more complicated risk controls.
4. Fair execution is primarily a scheduling problem
blurr deliberately preserved the ability to close positions even if markets must be frozen. He described oracle restrictions that could set the value to zero and stop new positions while still allowing existing positions to close.
A HyperEVM maintenance pause creates a specific exploit: open a large long-short straddle, close the losing leg at the stale price, then close the winning leg after the chain resumes. If the deviation is extreme, Paper Trade may freeze the market at the last price instead of turning the first post-pause block into a race.
Direct contract access during a large launch would produce a gas war. Everyone using the interface would be “sent like lambs to the slaughter,” so early activity passes through relayers and an internal mempool-like ordering system.
Priorities reflect trader harm: liquidations outrank closes, and closes outrank opens. Deposits and other operations are expensive because Paper Trade must deploy proxy contracts, and the system also considers size and other factors. The goal is to make 50,000 simultaneous users workable rather than merely first-come, first-served.
5. No funding and zero slippage come with hard exposure limits
Paper Trade charges no funding. blurr argued that Bitcoin and Ether perps do not often enter significant backwardation or contango because spot is globally liquid and transferable; funding would matter far more for an RWA whose market closes overnight or over a weekend.
A low-leverage long could consequently be cheaper to hold on Paper Trade than elsewhere. That design would not safely generalize to oil or similar assets, where a known 10% opening move could make financing “a significant part of the deal.”
Execution uses the average BBO with zero slippage, but size is not infinite. Size restrictions are meant to stay invisible to ordinary users while blocking an attacker from opening a position large enough to overwhelm the bankroll.
Paper Trade takes a 1% commission on profits and losses. Its Rollbit-derived payoff curve charges proportionally more when a position barely moves and improves the payout farther from the mark, making easy scalping difficult.
6. Insolvency creates queued winnings, not an immediate claim on the bankroll
The bankroll is initially limited to about $5 million, restricting both the house’s exposure and the economic reward for manipulating Hyperliquid’s BBO. blurr estimated that moving the relevant Bitcoin order book by 0.01% could require at least low- to medium-eight-figure amounts in some cases.
If a closed position has $6 million of unrealized profit against a $5 million bankroll, the trader receives $5 million and holds $1 million of debt. That balance sits in the queue and is paid as later traders lose.
Thread Guy tested the catastrophic case: Bitcoin rises about 7% in an hour, everyone is highly leveraged long, and debt reaches perhaps $150 million. blurr conceded that this is theoretically possible, though he does not expect it; such a situation could cause participants to stop or leave.
Each position is isolated and temporary. A trader’s opening balance remains theirs rather than disappearing into a pooled margin account. blurr acknowledged that a scenario in which everyone wins could technically leave the system with its original balances but no ability to pay all paper gains immediately.
7. Token emissions act as an automatic recapitalization bid
The token owns the commissions generated by the house, with Paper Trade retaining a small amount for gas. Gas deserves its own carve-out because blurr expects Paper Trade could become one of the few very large users of HyperEVM and could otherwise spend more on gas than commissions generate.
blurr could not recall the maximum starting emission precisely. He said it might be $100 or $1,000 per paper, then tentatively described the mechanism as beginning at $1,000 for every $1,000 of paper losses. The transcript does not establish a clean token-per-dollar figure. Emissions are paid immediately after a loss and can be staked without a waiting period.
Once the bankroll reaches roughly $2 million, token output per dollar of loss begins ratcheting downward toward zero. Supply could technically expand indefinitely if the bankroll repeatedly falls below that threshold, but blurr expects the ratchet to create a practical ceiling.
Above the adjustable $5 million bankroll limit, anyone can press a button to distribute the surplus pro rata to stakers: at $6 million, roughly $1 million becomes distributable. The token will not trade on day one; release depends on live results that blurr compared to an uncertain Monte Carlo simulation.
8. The mechanism matters even if the first product does not
blurr’s most immediate failure case was “many people getting angry and tweeting at Jesse,” followed by the more substantive possibility that Paper Trade simply receives no useful adoption. He does not describe the project as a child, but the pair have invested enough effort that irrelevance would still matter.
He recalled Lighter publishing part of its prover stack and offering rewards for improvements. The expected gain was perhaps 20–30%, but the result was about 10 times faster. Thread Guy referred to roughly 3,400 people; blurr responded that this was about how many tokens were burned trying to do it, leaving the exact participant count unclear.
blurr said the Paper Trade design had been worked on by only two or three people in addition to the model. He hopes public release produces successive iterations, just as CryptoKitties and later NFT systems built on earlier experiments: “It’s worth seeing the light of day because it’s a very new and unique mechanism.”
9. Jez supplied the rare collaboration blurr could not replace alone
blurr had historically found collaboration costly because explaining and delegating work took longer than completing it himself. Jez is the exception: he can hand Jesse a small logical problem, receive a solution and explanation, and finish faster than he could have alone.
That complementarity is personal as well as technical. They met in New York in summer 2021, apparently through NFTs, quickly understood one another, and Jez became “probably one of my best friends.”
blurr is not especially worried that Hyperliquid will simply absorb Paper Trade. Its core goals differ, and Hyperliquid has little reason to optimize gas capacity for this exact workload, though sustained Paper Trade demand could eventually force a discussion.
His pseudonymity was accidental rather than ideological. He entered crypto around 2017, before crypto Twitter personalities were an obvious career strategy, and simply never developed an interest in personal social media. He worked alone behind a laptop and saw little upside in visibility; Thread Guy described Paper Trade as more “cult classic” than mass-market acquisition funnel.
10. Crypto’s real purpose appears where legacy money fails
blurr’s governing heuristic is that “the purpose of a system is what it does,” not what its founders said it should do. Crypto’s demonstrated function is secure value transfer online; financial applications are therefore the clearest evidence of what the system is for.
Thread Guy supplied the example of paying a Bangladeshi click farm that accepted only USDT on Tron. Cheap transfers mattered while local cash allegedly lost around 40% of its value annually; blurr agreed that crypto solves problems that people in wealthy countries often do not encounter.
Thread Guy also described his personal conversion. Britain already offered instant, free bank-to-bank transfers, but splitting a restaurant bill in the United States around 2018 resembled an international wire—roughly $30 and potentially weeks—making crypto payments suddenly legible before Venmo and Cash App filled part of the gap.
blurr remains skeptical of forcing blockchains into voting, charity or broad social coordination merely because early Ethereum culture imagined those uses. “This thing is for finance. It trades. It’s value on the internet.” Skilled financial actors will otherwise overwhelm idealists using the wrong tool.
11. RuneScape trained the instincts that Compound and CryptoKitties monetized
Before crypto, blurr played RuneScape for roughly 12 hours a day and had never held a conventional job. His father joked that he would end up “super dead, super rich, or in supermax”; a privileged upbringing gave him enough time to find an internet-native economic outlet.
The decisive moment came toward the end of 2018 after depositing a couple hundred dollars into Compound. Watching interest accumulate through all 18 decimal places—even fractions of a cent—felt like a collectible game with real money. He thought that if it affected him so strongly, it could become important.
At 19, influenced partly by the Ethereum white paper’s tokenization thesis, he left university shortly afterward. The downside was recoverable; if the thesis was right, working on it had “extremely high EV.” His description: “I threw myself off a cliff.”
CryptoKitties provided the first edge. Each cat encoded visible and recessive genes in a byte string; blurr bought cheap cats carrying valuable recessive traits, bred them until the traits surfaced, and sold the offspring at multiples. Three weeks of that replaced RuneScape permanently.
12. Early on-chain fortunes came from reading mechanics before the crowd
Uniswap sniping rewarded being first because an initial buyer had little immediate downside beyond liquidity removal or being frozen. Contracts could enforce “first or do not buy” with strict slippage checks, turning transaction ordering into a game before developers learned to counter it.
Ethereum’s roughly 15-second blocks meant implementation quality barely mattered. Even “the worst piece of JavaScript” could finish within the block; the competitive advantage was game logic and knowing how to make the code act on-chain, not shaving hundredths of a second in low-level code.
Curve’s launch is blurr’s cleanest specimen. An account called 0xChad deployed public contracts from GitHub, written in then-obscure Vyper; blurr figured out the farming mechanics and was effectively the only farmer for about three hours, created the initial Uniswap pool and sold individual CRV tokens at “insane” prices.
At Fei’s launch, depositing ETH returned stablecoins plus roughly a 20% token bonus, while a penalty escalated as the stablecoin deviated from its peg. blurr sold early and pushed it far enough off peg that later holders could barely exit, capturing what he regarded as literal free money.
13. Crypto must absorb finance before it can redirect political power
After roughly four years of feeling that the government had “a knee on our neck,” Trump’s November 2024 win briefly felt like permission to build and a declaration that crypto had won. blurr then watched “old demons” reappear, tempering that conclusion without changing his belief in the rails.
Political transformation requires capital first. He cited BitMEX co-founder Ben Delo and Christopher Harborne, whom he described as a Tether founder and effective inventor of the stablecoin, giving roughly $100 million to UK political parties—by his account, more than all parties together had received in any year—as the model: build useful financial products, accumulate wealth, then demand change.
Meme coins are loud but, in his view, do not contain especially serious money relative to crypto’s larger financial system. Their real cost is reputational: many outsiders’ only exposure is fraud-like behavior, so the entire technology becomes “tarred” by the most visible casino.
Adoption is generational and often invisible. Borrowing the line that science advances “one funeral at a time,” blurr expects younger users to treat digital value rails as normal; the strongest victory would be people using crypto under the hood without knowing or caring that they are doing so.
14. Risk avoidance is shrinking opportunity while technology raises the stakes
blurr attributes part of Europe’s technology deficit to education that systematically suppresses risk. He contrasted a 16-year-old who wants to start a bank and receives encouragement with the response he associates with Great Britain: “What are you doing? Stop this.” The resulting company map is the United States’ mass of bubbles versus Europe’s rare Spotify.
Conventional labor no longer buys the freedom it bought prior generations: its value is “tunneling to the center of the Earth.” For young people, school and college followed by compliant employment work only through narrow channels; creating something uncertain may be mandatory even when some coin flips end badly.
Thread Guy framed civilization as repeated coin tosses whose frequency rises with technological interconnection: food stops reaching New York, one dependency fails, or an AI risk becomes real. blurr agreed that greater interconnectedness creates more risks. He separately described himself as a “doomer” after years of travel and exposure to the world’s worst conditions, while cautioning that most people should not adopt that burden.
Thread Guy’s counterargument was that catastrophe has not happened and “the game still lives.” blurr said he did not see an obvious holistic rebuttal. He argued that markets will build data centers despite local political resistance because “the market does what the market wants,” leaving few effective levers to halt the process.
15. A $13 million on-chain incident closed one pirate era without ending the game
blurr described a $13 million incident involving an Aave position in aTokens, with MKR against dollars. He had borrowed against the deposit, and the position sat almost exactly at the solvency boundary, so any adverse oracle move could trigger liquidation while he investigated an unfamiliar transaction and signature.
He could inspect MetaMask but did not know whether canceling an approval would prevent the action from being approved again. His instinct was to stabilize the position, take what he needed from the address and investigate afterward. “In the end, they got me.”
Thread Guy called blurr an “on-chain pirate,” but blurr thinks the lone pirate-developer archetype has faded. Young developers who once shipped experiments directly are now intercepted by a large venture-capital class—“Welcome to YC, buddy”—reducing the supply of strange public infrastructure that pirates once exploited.
Thread Guy described the remaining opportunity as a dark forest in which participants learn obscure contracts and hide their edge before others notice it. blurr agreed that a whole game remains, though the particular style he practiced is smaller than it once was. The game evolves, even if the old players eventually accept that “the Lost Boys in Peter Pan should return home.”
Verification Notes
- The transcript is unclear on whether the Lighter experiment involved approximately 3,400 participants or approximately 3,400 tokens burned; the digest preserves both attributions without resolving them.
- The transcript does not establish a precise initial token-emission rate or a token-per-dollar formula.
Full transcript
Yo, yo, yo. What's interesting? This is Thread Guy, and I'm here with a rare and special previously recorded in-person episode of the podcast. Joining me is a mythical guest: none other than blurr, in real life and on the go. My friend, you're not someone who often does these. As far as I know, you've done at least one.
Yes, that's true. Thank you for having me.
No, not at all. I'm so glad you're here. Obviously, we have Paper Trade's launch on the horizon, although the date is not yet defined. I tried to get this for you today. I don't think we'll get this today, but it is on the horizon because you work with the media.
Yes, we have something to talk about.
Also, as I found out, you have an incredible crypto history. To start, can you tell us what prompted you to start working on this idea of Paper Trade, and where did it come from?
This is a difficult question in many ways. We worked together over the last 5 years, but neither of us had ever created anything proper from scratch independently. With the set of tools that are now available, you can advance very far across many different verticals without necessarily having deep knowledge in that specific industry at the beginning.
Many people are very well versed in one specific area, but they lack 2 or 3 other things. Historically, that meant you needed a team of 7 or 8 people to do something because you simply didn't know how to create a user interface or something like that.
Now, many of those things, whether more complex or lighter, can be done. There isn't any annoying nonsense that takes hundreds of hours to study but isn't necessarily hard to do. It's just possible. That's what we were able to do.
As for the question of why, there isn't a convincing answer here. I guess it's quite a complicated topic for discussion. It's a pretty cruel question. I just wanted to know: why bother at all?
This probably won't significantly change matters for me or my partner, and in many respects there are more disadvantages than advantages for us, right? If everything goes well, we're the same as we were. If something goes very badly, it affects both of us negatively. The risk related to this is quite extreme.
But a ship in a harbor is safe, but ships are not made for that.
I like it. I like it. We talked a little off camera, and you mentioned in passing—although I didn't really ask the question—that there's another foundational aspect of Paper Trade's design: a mechanism that was in the background for a long time, but that now makes sense to implement with Hyperliquid and perps.
Yes. The basic system allows you to launch effectively anything that has coefficients tied to it from scratch, without the necessity of starting with a bankroll for the game. That's something new and interesting, especially for people like us who have historically often worked independently and haven't necessarily had to interact or deal with a bunch of other people in a semi-professional capacity.
That definitely spoke to us. We wrote an article, again, it seems in the summer of 2024, about this mechanism. Then we developed it for simple casino games such as blackjack and coin tosses, for things that have deterministic odds, so that you can model them more clearly and display the results.
But something like that isn't worth launching. It's simple. It's not interesting enough.
Can you explain what the mechanism is in the simplest sense?
Essentially, there is no bankroll. Or I have a bankroll, but that's all. In crypto, when you take a long position, someone on the other side takes a short one. They're working against each other.
In a casino, when you place bets, you bet against the institution that owns the casino. This means you need someone on the other side to take on that risk, right? If you go to a casino, you'll often be limited to a maximum bet. You can't break it because the casino has to continue working; it depends on its bankroll.
That's the problem: these things need a cold start, which is much higher than usual. But with tokens, you can effectively sell access to the bankroll, or increase the bankroll by involving people in the games.
The idea is that you have an emissions curve, and emissions are much higher when the casino is more insolvent or closer to red. Thus, if people assume or think it will be profitable in the long term, they're encouraged to lose money—to effectively buy into the system.
Generally speaking, if you take a step back, in cryptocurrency there are 2 businesses that earn a lot of money: market makers and exchanges. That's it, right? And where does all the money they earn generally come from? It comes from users, such as retail users.
In some exchange cases, they actually earn money through PvP promotions with market makers. Usually they work like that, but the money just flows down—or up, I think. Maybe it's better to say that all of this is extracted from users.
It's very difficult to develop a system where this doesn't happen because market making is a really complicated business. It's very difficult to do it well, and it will destroy you if you don't really know what you're doing.
Things like HLP or LLP, exchange liquidity pools in AMM styles, work pretty well, but they have huge asymmetric benefits to the exchange, which allow them to do this without requiring the same degree of mathematics or HFT skills.
What advantage do they have?
It's something like Hyperliquid, which allows you to buy something like liquidation flow. The same applies to LPs. A fee is charged for liquidation. Liquidation of similar assets is a great place to collect fees because, in general, no one includes it in the mathematics of trading those assets, right? Nobody thinks they're going to be liquidated, so this is a simple way to weigh against the user in the casino's favor.
For Hyperliquid, LPs serve a specific purpose. If they weren't there, some coins simply wouldn't have enough liquidity to make them tradable. They provide something like backstop liquidity in the desert, where it's hard to find liquidity.
So the thought was: how can I try to externalize the profits that market makers and exchanges receive to users? This is very difficult to do.
The system we created with queues, in the case of insolvency, somehow fits here. It allows you to synthetically put everything on the exchange. You give the exchange a significant advantage, right? The exchange has a very large advantage because I offer you many benefits that you otherwise wouldn't be able to get.
At the same time, all of this advantage simply goes back to the people who own the house—in this case, the users who lost in the system. The exchange has a certain advantage over the user, but everything it creates turns back to the users.
Unlike classic exchanges, where everything is simple: it is listed on the exchange, right? For example, in the case of Hyperliquid or Lighter tokens, they are buying tokens, but at the same time a lot remains pulled out from under the hood, especially by market makers. But that's not the case here. All returns to users.
As a user, this seems understandable enough. I want to ask later about the market-maker business. It seems a little shady.
I'm not a market maker.
No, I know. I'm just curious. But okay, I guess it's worth taking a step forward, even at a high level, and we can talk about some details.
How do you explain to someone who asks you what Paper Trade is? How do you explain what it is and what it does? Some of these things existed before. For example, very basic trade casinos with 1,000x leverage already existed. Rollbit proposed some of these things with leverage.
Maybe it's not worth looking at the mechanisms under the hood. At a high level, how do you explain what Paper Trade is and how it differs from what's on the market now?
I never used Rollbit. The mathematics of these casinos, which usually appear and then disappear every 2 or 3 years, reaches a point in their lifecycle when it makes sense for them to get out with the money. So you never feel secure investing your money there, right?
It's a very dark game, and online casinos themselves thrive on crime, right? That's just how they work.
In this case, we use Hyperliquid BBO as an oracle, directly. HyperEVM works like this: there's 1 block on HyperEVM and 3 blocks on HyperCore per HyperEVM block. We're in a chain, right? So our oracle is readable directly from the chain.
That's fairly unique, isn't it? People often tried to do this in the early stages of DeFi; we simply took the Uniswap price as the oracle. It's very dangerous because, obviously, it's very easy to manipulate.
Hyperliquid itself can be manipulated, but they have risk parameters to guarantee that this is never profitable, at least for Paper Trade.
But our oracle is incredibly effective and decentralized. There is no risk that I can deceive you at all. I specifically designed it so that you can see how it works. Everything is very public.
I haven’t seen a real stock exchange like 1000x where it wasn’t just internal under the hood, with some kind of small expense, essentially.
Of course. Where would it be public and invisible? So why isn’t this standard? How do you do it?
This is difficult to do. It’s really hard technically. Hyperliquid has access to precompiles, which is new and interesting, and that’s why we decided to create it ourselves there. It just made sense.
But is this also dangerous to do in general?
Yes. 1000x is a lot, right? If you open a position at 1000x with a long and a 1000x short position, the price only has to move more than 0.1% very quickly. If you’re cleanly selling a house, you’ve lost it.
When the situation is very volatile, we only offer positions on Ether and Bitcoin. If you tried to offer something like Solana, it would become much more dangerous, because someone could come in and trade against the price. Even stochastic price movement could simply overwhelm you.
For example, if it moves very quickly, the liquidation of one side of your position doesn’t outweigh the side that wins. You could win more than you lose on the liquidated side. So this is a dangerous game.
I could offer something like Solana at 300x, perhaps. That can work, and we’re considering adding more assets in the future. But there are risk-management restrictions that significantly complicate the product beyond what people are ready to try, I think.
What is your goal regarding how users and traders interact with the protocol?
We very quickly created a V1 that functioned essentially the same way as this one does now. Then Jez posted a tweet announcing it, and it turned out a little better than we expected. Looking back, it was a bit silly, considering who Jez is.
But we realized that if this thing had 20,000 users from the first day, it would be a completely different product that we needed to create. Cryptocurrency is cool because it handles a lot of unpleasant server-side problems that are an integral part of creating complex products, such as databases.
Our problem is that I handled cryptocurrency very well at the beginning, particularly launching things. People who created things did what they wanted, but maybe they were brilliant programmers and not very good economists. What they built wasn’t quite what they thought it was, and the game theory was a little wrong.
You can look at something and say, “This is how it will go if I do X, Y, and Z. For example, I can get free money.” I realized that I didn’t want to create something that would have that kind of life cycle, if that makes sense.
Creating something that is genuinely fair is very difficult, because you need to cut off every angle that would allow it to break down. I spend a lot of time looking at the system and thinking, “Okay, how would I break it? How would I break it? How would I break it?” Then I recursively work through it, inventing as many ways as possible.
There were many ways to break it. Some of them were funny. I even liked the ones where the administrator could break it, which I had to get rid of.
It seems there’s no way—even if you can’t break it, there are still things I can do. I can suspend the markets, which essentially destroys them. But you still have to be able to close your position, regardless of the price.
I have a system of oracle restrictions, so I can set the value to zero, but you can always close. You just won’t be able to open new positions.
Under what scenario would you need to close a market?
One of the strange things about Hyperliquid—and this is actually a strange problem with Hyperliquid EVM, I would say—is that they often suspend the chain. For example, they might do it on Sundays to improve something.
Of course. What takes place in that kind of scenario?
Usually it’s only for a minute, but if it were longer, you could be in a position where someone opened a large straddle—a long and a short position—and the price changed a lot.
They could close the losing side at the old price, and then, when the chain comes back, close the winning side. The price isn’t updated while the chain is suspended because it’s still flowing from the original price.
So if you get into the first block after that, and the price starts moving rapidly, you can close the short position and leave the long open?
Yes, exactly. Then you have a block, and it becomes a race. I don’t really want that.
There are scenarios like this where, if the price deviates too strongly, I’ll have to freeze the market. Otherwise it could become insolvent, and people would leave in droves with free money. So I would freeze it at the last price.
It would happen very rarely, and I don’t think it’s likely. But a lot of this is about thinking through all these strange scenarios—the cold paths that you otherwise wouldn’t have considered and that might never arise—and asking, “How could I break this?”
That raises the question: Hyperliquid EVM apparently isn’t the most performant or best choice for building something like this, at least if you’re trying to move a pile of perpetual transactions on-chain. If you could choose anything in a vacuum, it would probably be Hyperliquid EVM. Why?
Yes. Why? Precompiles. I get a free oracle, and no one else has that.
Can you explain the importance of that, and how you would solve it somewhere else? There are many oracles everywhere, and they’re usually a stumbling block for a lot of products.
It’s just not a nice thing to have to deal with. The fact that it seems to be solved for us is enough. It’s simple.
Our approach at the beginning was basically to have fun with the draft of the product, which actually got out of control. The starting point was, “Wouldn’t it be fun to create something like a mechanism-bankroll casino on top of perpetuals?”
When Hyperliquid’s precompiles launched, we could make it work that way. It was like, “Yes, of course.” Those events happened much later in the cycle, though. By that point, we were already deep into this.
It’s also simple. In the end, it’s not the same product, right? It’s just a different thing. This isn’t exactly a perpetual exchange. It’s not quite something else either. It’s something in the middle.
What is it? How do you classify it?
It’s paper trading.
Paper trading is wonderful. What’s special about this oracle?
It’s incredibly decentralized. It doesn’t need any supervision, because there’s no pause mechanism. It just is what it is.
Unlike everything else—Chainlink, Pyth, or someone else—they’re all essentially authorized oracles.
And therefore they’re more vulnerable to attack?
None of them has suffered very much. Sometimes they have been affected. For example, I think Pyth was the one that broke silver prices at one point.
In theory, you can hit a centralized oracle hard. It’s just a faster failure case if the keys are compromised or something like that. There’s an additional element there.
Here, that risk doesn’t exist at all. Even if the North Koreans seized everything here, you could still just take off. You have a 7-day period, and they can’t take anything from you.
There’s an entire stack of failure modes. If they take over DNS, there are problems. But generally speaking, with the chain itself, there isn’t much anyone can do. I don’t know—there are still things I can do.
Practically nowhere else works with this degree of theoretical autonomy. That’s pretty cool. For me, it’s clean. You understand?
So if this thing launches and 50,000 users are trying to block-trade at once, what happens? Hyperliquid EVM is pretty slow.
How does it work with gas?
I give priority to transactions. At the beginning, I have to run everything through our own relayers. If I let people interact with the contracts directly, the whole thing would turn into a nightmare gas war.
I’ve seen many launches historically grow into that, and I understand that it’s simply bad. The people who leave—the people who win—are usually a certain subset of users.
The people who win?
Yes. It’s just a subset. Anyone who goes through the user interface is effectively a leak. They’re being sent like lambs to the slaughter.
That isn’t the goal at all.
So a lot of the work involved restructuring things and announcing them after the fact, asking, “How do we create this for 50,000 users who all launch the thing simultaneously?”
The entire backend system was built around that. We essentially have our own mempool, something like software.
Of course. What we use to find out what goes where on the stack, right? Things like closing outweighing opening because they’re more important. Liquidations are much more important than closing.
It all depends on priority, as you said?
Closing outweighs opening. Closing for us is more important than opening. Of course, someone who closes a position is more important than someone who opens a position.
Yes, that’s true. I want to close. I want to close. To you right now, is the person who opens the position treated a little smaller because you don’t know whether you have the capacity?
Of course.
Okay, so there are a bunch of things. There’s something like, “Okay, if I were a trader, how badly do I see this?” So you’re manually arranging priorities for these various things?
Yes, there are a lot of these things. This is also something like deposits and other operations, right? They’re very expensive from a gas point of view because I have to deploy a whole set of proxy contracts for you.
There’s weighting for size and other things that are optimal for exchanges, but also optimal for trader psychology under the hood. There’s a lot of this where you just sit and use the system, and you’re trying things. You’re like, “Damn. Okay.”
You’re trying to invent scenarios that might arise. How do I wish this had unfolded? You’re methodical. You’re dealing with one person out of 10,000, and you’re like, “Aha.”
That was a really interesting cycle, because you can sit and think that something works in a certain way, and then you use it and think, “Oh, this is completely different. That’s not how it works.” Actually, that’s the problem, and I need to solve it.
As a person who had never done anything like this before, I had always been a user. I had a picture of how everything should look, and I was putting it together. Then you collect everything and it’s like, “Oh, okay. There are so many other things that I hadn’t even calculated. They were part of the delivery, and there are problems here.”
Can you explain how funding works on perps compared to traditional markets?
We don’t have any funding.
So how does it work?
Bitcoin and Ether don’t very often go into backwardation or contango. The price of a perp doesn’t often deviate from the underlying spot price, right?
Okay, what is this called?
Backwardation and contango. Those are the two technical terms for it. The spot is easily accessible, right? You can literally just transfer it. So they basically track one another very clearly, and they’re such liquid assets everywhere in the world that this isn’t a problem.
If you tried to do it with an RWA, it would be a huge problem because during weekends and, for example, overnight, you could get financing at 1,000%. That becomes a significant part of the deal, right?
If everyone knows that, for example, the price of oil will increase by 10% tomorrow, financing will be completely distorted. But I can’t explain this because there’s no previous contract for financing, right?
In general, financing is a function of the math of the payments. For Bitcoin and Ether, it isn’t a large enough component to be materially important. I don’t charge you anything. So if you want to hold a long position on Paper with low leverage, you can do that, and it will possibly be cheaper than doing it somewhere else, where you pay financing.
It’s up to you. But as a general part of the mathematics, it isn’t significant enough to really matter from an exchange-operations perspective.
Can you explain bankroll management and the concept that the more money you have in your Paper bankroll, the bigger the risk you take on yourself in exchange for the upside?
Okay. Do you remember from yesterday? It’s structured, right? Something like “more on the line”—the bigger the bankroll, the bigger the hole.
For example, if I have a bankroll of $200 million, the cost to move the Hyperliquid order book enough to manipulate the BBO and receive benefits from the agreement is some number. It’s quite high. It’s something like 8-digit numbers or higher in some cases—at least low- to medium-8-digit numbers.
Do you publish that, or is it an internal number?
You can simply check it out. If you run the exchange, for example, you can see what it would cost to move the Hyperliquid Bitcoin order book by 0.01% or something like that.
I meant your bankroll number. Paper’s bankroll.
Yes, that will be on-chain.
First, I limit the bankroll size to about $5 million. This simply means that the bankroll never exceeds that amount, and therefore the payoff curve for attempts to break through it isn’t very large.
What happens if my position has about $6 million of UPNL and the bankroll is equal to $5 million?
You take $5 million and get $1 million of debt.
And will you pay the same $1 million, or not?
That’s in the queue. When people lose, this will happen, but yes.
Could you have a scenario where there are, for example, $150 million of debt—a huge amount of debt?
Yes. Theoretically, that would break the system.
Would there be several people who simply stop because the debt is that large? If Bitcoin increased by approximately 7% in 1 hour and people tended to take long positions, everyone could have highly leveraged long positions.
Yes, there is a scenario where all of that happens. Technically, it’s bad, but also no.
Your balance is yours. When you make a trade and close it, your opening balance is always yours. It doesn’t go anywhere into a super-account. It stays with you. All positions on Paper are temporary.
They seem to be separated from everything else. Your balance isn’t like a total account or something like that. Each position is its own thing.
But imagine a scenario where everyone enters and everyone wins. The essence of that is that you just end up with what you started with. There’s no such thing as—
Yes, I understand. But it is technically possible. This is one of the reasons why you wouldn’t create something similar, because synthetic bookkeeping that is completely internal carries risk.
You can try to weigh it in favor of the exchange and determine how much makes sense before it starts to become meaningless. We did the math and understand how it works. In the end, it’s simple. You know how the system works.
Because of the way it’s designed, I don’t necessarily predict that this will happen. I don’t think it will happen. In an ideal scenario, we never use a queue, but it should be there to overcome the cold-start hump.
Most traders lose money. More than 50% of traders lose money, so in this specific case, the system is also weighted in favor of the exchange.
Do you have people who would like this to be considered a benefit of the exchange? You don’t charge a conditional commission, right?
We charge commission. You can open a position for $20 and close it again in a second, so there’s no reason for that without commissions. We charge commission on profits and losses, so there’s a percentage of the profit.
We also use an old mechanism—Rollbit invented it, or maybe it existed before, but we took most of the math from them. There are small changes, but there is a payoff curve.
One of the reasons for this is that Rollbit specifically allowed people to use stupid leverage on many shitcoins. I know many people who earned a lot of money essentially trading on Rollbit, floating with half-payoffs, which doesn’t make sense.
But one of the things they did was make your payout improve the farther your position moves from the mark.
Oh, I understand.
If the price increased by only a very tiny percentage, you pay a much larger commission. You pay a much larger commission, so you can’t scalp very easily. This is one of the protections that we have.
That’s good. That’s interesting.
How does slippage work?
There’s no slippage—zero slippage. It’s based on the average BBO.
With zero slippage, can you get anything like infinite size?
That won’t happen endlessly. If I allowed infinite size, you could open an infinitely large position on the perp, and that would lead to problems. There will be some kind of size restriction.
A regular user should never face this. It should be as if it isn’t there. It’s specifically there to block attacks like that.
Can you explain the token emission? Obviously, that’s what a lot of people want to know: what the token is and how to earn it.
The token owns the commissions that the house generates, basically.
Everyone?
Almost everyone. We have to pay for all the gas, and that can be expensive. There’s a small fee that we also take, but is that just for gas?
Gas is interesting because there’s math around how much it makes sense to pay.
And when it gets a little strange for you to manually decide when it happens, how does it work, essentially?
Yes. There are also a bunch of economic aspects where people could try to attack you on this side.
On the gas side?
Yes, because I’m trying to manage this. If I manage it incorrectly and burn too much gas compared with the commissions paid, eventually I’m just burning money.
Yes, that’s right. This isn’t sustainable. You just stop transactions?
Theoretically, I would probably just continue to trade and burn money, because, you know, whatever. But if you tried to manage it properly, you would design it according to that.
At the same time, I predict that I will probably be one of the few, if any, very large users of this chain in general, and that everyone who comes will do it for some other reason. That’s exactly how I imagine the development of events: I’ll actually be the only serious gas user on HyperEVM. So I’ll generally be able to control the price of the gas I pay for.
These are gas-adjustment algorithms used in chains like this, or at least in EVM-style chains. I don’t remember which EIP it was—maybe EIP-1559—but there was one that changed the adjustment mechanism. I don’t know. They weren’t really created for single-user chains.
It’s very similar to if there were this pool and many people were playing in the system, and then it adjusts, right? If it’s only you, it becomes a little strange because you can decide where everything goes.
What’s the caveat?
The caveat is that for large traders like me, obviously, it would be necessary to make sure that I can exit quickly. If you’re a small trader withdrawing funds or something like that, you don’t want to pay $30 in gas to get through this. That’s simply unreasonable, so I would wait until gas comes down to get through it.
There are scenarios when you still want to make the transaction, though. There’s a chance I still spend a lot of money on gas.
So there’s this one commission.
The commission is 1% of all profits and losses from positions.
1% of all profits and losses is the commission?
Yes, 1% of profits and losses is the commission. Then there’s whatever overflows.
I think the overflow is probably the most interesting aspect here. Everything exceeding $5 million in the bankroll is simply a button you can click, and it gets paid to stakers pro rata.
Is $5 million a hard limit, or would you correct that?
It’s adjustable. I think I’ll probably adjust it upward, but it’s a kind of game. We need to see how it gets played. If there are $6 million in paper losses, you press the button and $1 million gets paid out.
So if the bankroll gets stuck at $5 million?
Yes, that’s how it works.
The emission curve itself—I don’t remember if it’s $1,000 or $100 per paper—but it begins, I think, at $1,000 for every $1,000 of paper losses. That’s basically paid to you immediately after the loss.
They pay directly?
They pay directly. Nothing is waiting; it goes right into your account, and you can stake it instantly. There are no waiting delays or anything like that.
Obviously, you need to stake it to earn emissions. Then there’s the curve. As soon as the bankroll reaches a certain level, the curve changes, because it completely depends on how solvent the bank is. Once the bankroll reaches, I think, $2 million, anything above that gets onto this ratchet curve, where the payment—for example, the number of tokens you get per dollar of losses—decreases.
This continues toward infinity.
Why does it work like this? So the token supply isn’t completely unlimited, right?
The assumption is that we finish at $5 million, and then this begins to taper off, so we find ourselves on this curve and it seems to drop to zero. But if it ever returns below $2 million, it returns to $1,000 per dollar. Is it like infinite minting?
Technically, yes?
There are ways in which it can be minted endlessly, but I think it will be on this ratchet curve and therefore have a certain ceiling. That’s exactly it, though—I don’t know, because it depends on how much is happening on the bankroll side below $2 million.
But yes, it’s specially designed to promote losses when the bankroll most needs help. It was specifically developed for that.
Of course. Is it possible to trade it on the first day?
No, it can’t. If it were possible to trade it on the first day, most of that trading would probably take place on HyperEVM.
Damn it, all the gas is needed. I can pay. Is it possible to trade it at some point?
Yes.
And when? It’s undefined or uncertain?
Yes, it’s a bit uncertain. I can’t provide an exact date because it’s simply a Monte Carlo simulator. I don’t definitely know how it works, but when the time comes that I can let it out, we’ll do it.
Maybe you don’t want to say this on the podcast, but what are the attack vectors that worry you the most? What could the failure modes look like? How probable are they, theoretically? What are you trying to avoid?
Many people getting angry and tweeting at Jesse, probably. Is that the right answer to this question?
Everything that leads to that result. Obviously, we’ve spent quite a lot of time on this. A lot of founders treat their businesses like their children. I don’t think that’s quite the case for us, but we’ve put a lot of effort into this.
If it doesn’t receive any usefulness or use, it will just be something like sludge.
The failure is that it somehow experiences failure. That’s probably the better question: how big can this become? If the mechanism is simple, used for its purpose, and works very well, how important an element of the trading universe can it become?
I think the basic model is very interesting.
That’s funny. Lighter held this little experiment a few weeks ago where they published part of their prover stack. They said something like, “If you want to run an agent on this and try to improve it, because we can verify it deterministically, we’ll actually pay you for any improvements in speed that you achieve.”
The assumption was that maybe it could improve by 20% or 30%. In the end, it was about 10 times faster. The agents found a lot of things.
There were about 3,400 people who worked on this, right?
Yes. That’s about how many tokens were burned trying to do it.
The point is that you have some of the smartest ZK engineers who have ever existed working on this full-time. But there are only 4 or 5 of them, and they can only do a certain amount of things.
When you do something with open-source code that works well for everyone in the world, everyone who manages these models can improve the system pretty quickly. This was worked on by only 2 or 3 people, in addition to the model.
I think that as soon as we release this and people see how it works, I hope there will be a lot of iterations on the base design. I don’t know who knows how this will work, but I think it’s worth seeing the light of day because it’s a very new and unique mechanism.
A lot of things in cryptocurrency have historically happened because someone started something, then someone iterated on it, and someone else iterated on that. NFTs came from that kind of progressive cycle of iteration, right?
Yes.
If there hadn’t been that first person—Dapper Labs, with something similar to CryptoKitties—nothing like that would have happened, right?
Yes, definitely.
So, in my opinion, the result could be pretty cool.
I came here for NBA Top Shot.
Really?
Yes, that’s right. It’s funny how these things arise. I got into cryptocurrency because of CryptoKitties. That’s how I started. I read the BBC article on a Tuesday evening in my dormitory. It looked interesting, and then I went down the slope.
Did you get any advice? I probably don’t, but do you have any advice on how to shape the business or what people should do? How should people approach it, otherwise they’ll simply lose money?
My answer to that question is to get involved in this agency.
You continue to say this. I like it.
I mean that it’s designed to be a feasible thing, and if it works well—
What’s it like working with Jesse?
It’s great. I love Jesse.
We met in the summer of 2021, I want to say, here in New York.
That was a long time ago—5 years ago.
Well, yes. That was the beginning for you guys, I think. We very quickly understood each other. We also accidentally met through NFTs, of all things.
We’ve always worked alone, as if I never really had colleagues or anything like that. I just felt light. When someone else appears, I have to unload things, so I spend more time thinking about unloading than on any benefit I get from it.
I’m terrible at micromanagement. Historically, it just hasn’t worked very well for me. Jesse is virtually the only person I’ve ever met where, if I encounter some strange, small logical problem that I’m having difficulty deciding, I know exactly what will happen: I’ll show it to Jesse quickly, explain it to him, ask him to understand it, ask him to solve it, and then ask him to explain it to me first, before I would have reached the answer on my own.
That is absolutely unique to me. I’ve never met anyone like that in my life. It was a kind of discovery that this guy was operating on a slightly different level than me. We’ve done some work together for many years, that’s true, but he is probably one of my best friends. That’s exactly it.
If you want to create this on a hybrid virtual machine, and it becomes very successful, aren’t you worried that Jesse would just say, “Yes, give me this”?
Risky question. Risky response. Not “yes.” This has to be done carefully.
I don’t think it necessarily makes sense for them to do it. For example, the goal they’re working on differs very much from what he works on. So it isn’t necessarily wise for them to do so. I don’t think they will necessarily work to our benefit when it comes to things like capacity, gas, and so on. It doesn’t make sense for them to optimize that for us.
Maybe eventually we’ll create a lot of noise around gas, and then there will be some discussion. But I don’t think it will really happen like that, just as, say, the Hyena [?] encountered problems trying to build a kind of deployment of 3 followers [?]. That’s something completely different, a completely different flavor.
Of course. So the real question is: do you like that? Because I heard—at least to me, there wasn’t much of this—you’re something of a mythical figure. That’s the word I use to describe you in Chain Wizard. I’ve never seen your face before. You don’t write on Twitter. You have 2,000 subscribers.
I wrote your name on Twitter, and everyone writes back to me: “Greetings, Cyrus.” Everyone tells me you’re Wilt Chamberlain, that you invented Uniswap farming. I hear all these crazy things.
I’ve never seen you before. I don’t know anything about you. I saw the Taiki interview—greetings, Taiki, legend—and now I’m meeting you personally. We’re in New York, and you’re showing your face on camera.
I don’t know how rich you are, but I think you’ve done pretty well. What prompted you to sign something with your own name, show your face, and put your brand and career at risk? How was it, and why are you doing it?
I never planned to be anonymous. It wasn’t a decision that I consciously made. You remember, I got into it somewhere in 2017, a year before cryptocurrency Twitter really came along, right? You lived on Discord. Telegram didn’t even exist yet. It was very early.
There wasn’t really such an idea as pure anonymity. It simply didn’t exist in that sense. This was among the first large online personalities who became something like BitBoy. There was GCR and so on, but at that point it wasn’t really a thing. That just wasn’t something I was ever interested in.
I don’t like that side of life. I’ve never been interested in social networks or personal life being on social networks. It’s just not something I’ve ever cared about.
Never?
Never. It’s simple. It’s like, “What are we doing, guys? This is a little gay, don’t you understand? Interacting with other guys online and posting gay Instagram stories?” I don’t know. Damn earrings, dude.
But yes, it wasn’t a deliberate decision that I made to avoid doing it. It was simply criticized, and then it never made sense to do it. I never had a strong position on it.
People just think, “What are you doing? You don’t do this, but everyone else does. Are you inclined to do so, or are you wrong?” I was also lucky that my career—however much that word makes sense as a description of what I do—was never connected to other people.
I was completely alone in a room behind a laptop, and it worked, didn’t it? I never had to do this. There was no function forcing me to leave and do it, so I never did.
I don’t think doing something like that would necessarily have a lot of advantages and not many flaws for me, considering what I do. I don’t get any benefit from it, except Paper Trade sales.
Honestly, I think that from a penetration perspective, at least at the beginning, anyone who is going to use Paper Trade probably already knows that it exists. I don’t think showing my face expands its reach. It’s more of a cult-classic product.
Yes, exactly. It became something of a topic of conversation, and that’s because Jesse had already tied into it, right? We already have one perspective, and Jesse is the one who—
Who is Jesse?
That means that side of it was already resolved for us, which is good.
I saw one of your interviews and thought, “What’s special about you, considering your reputation?” Let’s say—
Thanks, dude.
That’s not a bad reputation.
Okay, we’re not there yet.
But you really are a good interviewer, and I thought, “This is cool.” I really like it when someone pushes beyond the limits of who they are, or beyond what they’re expected to be. I thought it would be fun to chat.
Thanks, dude. Glad to.
Did you grow up as a child in your room by a computer? When you were 16, did you think about what you would become?
You just want to know that I played RuneScape for 12 hours per day.
Got it. What would you be doing now if you hadn’t found cryptocurrency?
It would be some kind of internet commotion [?].
I think that’s always fun, right? If we rewind a little bit—and excuse me, I’m sorry, Shu—you were looking at the internet in general from around 2005, 2006, or 2007 onward. Every time you played poker, “Poker died in the U.S.” If you wanted to earn money online, you were a lonely guy and you played poker.
You understand what I’m talking about, right? There was a large cohort of men who very often used the internet and historically weren’t especially economically active. Then cryptocurrency became a means and a channel through which people could interact economically without ever leaving the room.
I don’t really know what I would have done. My dad always joked when I was little that, at some point in my life, I would either be super dead, super rich, or in supermax.
What’s the third?
Supermax, like in prison.
Oh, God. Thank you, Dad.
I had a fairly privileged upbringing, I would say. But considering who I am, if I hadn’t been given the time to sit and play RuneScape for 12 hours a day, there’s a very high probability that I would be living under a bridge or something like that.
I’m really not good at many other things that contribute to working in a workplace environment. I’ve never liked doing that. I’ve never had a job. So it’s not as though I would have been very happy if that kind of thing hadn’t existed.
Do you think cryptocurrency will continue to be this kind of game?
I think so. You should be surprised by how long cryptocurrency has continued as a game. But if you throw away your assumptions about what cryptocurrency is—everything that it does—and look at it from a basic point of view, it’s simply the transfer of value in a safe way online. That’s all it is.
The structures we place on top of that, by design, create these emergent properties. But there’s no reason why it would disappear. You can go back and read articles saying, “Bitcoin is dead,” all the way back to 2009, right? There were people saying that then, and they always appear.
They always say that people don’t understand what it is under the hood. It solves a specific problem, doesn’t it? That problem is allowing people to interact economically in a digital world, and that’s genuinely useful.
So no, I don’t think it’s going anywhere. It’s too useful to get rid of, regardless of how many crimes or other things are happening around it. That’s why I think it will stay here.
I like your opinion because I’ve probably talked to many people like you, but a lot of people who graduated in 2017 or around then, achieved very good results in cryptocurrency, and were still active in crypto by 2024 are quite cynical and angry about everything that’s bad for the world—memecoins, Pump.fun, you know what I’m talking about.
They’re like, “This is beautiful…”
Yes, it’s funny, right? There’s a kind of heuristic in systems thinking that says the purpose of a system is what it does, right? At first glance, that sounds a bit tautological. The purpose of a system is what it does.
But the point is that the purpose of the system is not what it is supposed to be. It is exclusively what it does.
I remember that sometime in 2017 or 2018, I had to pay for something on Reddit. I ended up dealing with a click farm in Bangladesh, and at that time, the only payment they accepted was USDT on Tron. I thought, “What?” This was before Tron had its moment, sometime around the end of 2018 or early 2019.
It happened because it was very cheap. This is no longer the case, but at one point, money on Tron could actually move freely. Ethereum transactions were expensive, and one-time transfers cost quite a lot of money, so it was logical for them to use Tron. I remember thinking it was strange. Why was this happening?
But it makes sense, right? It really solves a problem for them. If they keep their money in Bangladeshi taka, its value falls by approximately 40% per year. If you have cash, the cost of storing it safely is extremely high. Tether on Tron was a real salvation for these people. It completely changed how the economies of third-world countries worked.
The purpose of a system consists of what it does. That is the purpose of cryptocurrency: it solves people’s problems that we don’t encounter because they are not problems we have.
I remember the first time I came to the States, around 2018, as an adult. Ever since I became functional and had a phone, in Great Britain you could instantly transfer money between bank accounts, free of charge, from one bank to another. It had always been that way. You pressed a button, and everything happened. It was free.
Then I came here and needed to split the bill with a girl at a restaurant. I said, “I’ll just send you the money.” Is that possible? Is it difficult? It was a big problem. An international bank transfer cost me $30, and then $30 came back approximately 30 days later. I thought, “What is this?”
That was the first time the whole thing with cryptocurrency payments became understandable to me. I thought, “If the regime in which you live makes owning and moving your money this complicated, then it makes sense to try to create an alternative.”
Of course, Venmo and Cash App appeared and solved—or filled—this gap in the United States. But at the time, this side of things had never made sense to me in Great Britain. It was, “Okay, I think it should make sense, but the alternatives are good enough.”
This is an interesting epithet, because I think your view of what it does for people is potentially short-sighted. You just don’t have the same problems as other people. You don’t necessarily need the solution, but other people can.
What was the flashpoint or light-bulb moment when you thought, “Damn, this is really cool, and I can make a lot of money”?
The light-bulb moment was when I made one of the first deposits in Compound, right when it launched.
Robert Leshner?
Yes, Leshner. Leshner was the guy.
I remember seeing it and thinking, “What year is it?” This was probably sometime toward the end of 2018. Compound had probably launched toward the end of 2018.
I remember that there were probably 18 decimal places. I bet I deposited a couple hundred bucks or something like that. You immediately saw the 18th decimal place start counting up, and I remember watching it and thinking, “This is it. Is something going to happen, or am I wrong?” This was long before DeFi had a name or anything like that.
What do you mean by 18? A comma? Tokens on Ethereum have 18 characters after the comma, right?
Right. You just had a balance that was approximately 6.02.
Of course.
Eighteen numbers. Watching all these numbers ticking upward, even though they represented fractions of a cent, was so tempting to me. I just knew that if I could feel it, then—well, I was a big RuneScape player. I like collectible games.
I don’t usually play something this intensely, but I felt that this was going to become something important. At that point, I actually quit university shortly afterward so I could continue with it. I thought it would become something important.
What were you doing when you dropped out of university?
This was still very low-level, essentially. It was enough to live on, but there wasn’t much money to play with.
Did you assume there would be money, or were you simply captivated?
I was a believer at that time, and, in some sense, I still am. It was a hard road, but we’ll get there. We’ll get there eventually.
The other thing was the Ethereum white paper, which is actually quite hard to find on the internet now. I think it preached Ethereum’s tokenization thesis: we were going to tokenize everything—real estate, whatever. That was probably in 2018.
I read it and thought, “This makes sense to me. This is what’s going to happen.” There was nothing else. I was 19, and I thought that if I worked on this, the expected value would be extremely high. It would be very good.
In the worst case, I would be doing something I liked. I could just go and do whatever I wanted. But if I was right—if this was right—then, you know. That’s why I threw myself off a cliff, as it were. It was a very unwise risk from my side.
What was the scene like at that time?
There really wasn’t much of a scene. There were contracts being deployed, small games, things like hot potatoes, pyramids, and so on. There were a bunch of developers reworking everything they could, and tiny teams were pulling commissions from them.
There was a class, or cohort, of people who played these games. That was it. It was a very small subset of people.
On the one hand, there were people who really admired CryptoKitties, which stayed around because it was the only one of its kind. It was something, but everything faded away soon afterward. I was one of the few who remained at the end of the 2018–2019 winter.
How did you cope with CryptoKitties V1? Was it like, “Oh, this is going to be a huge moment”?
I was one of the first people to understand how crossbreeding worked.
How did you understand that?
I just understood it. I don’t know. I understood that every cat was represented by a long string of bytes. That string was broken into small pieces, and each digit in the string of bytes was a gene.
There was an algorithmic switch, which meant that if you had a recessive gene, it wasn’t expressed, but it was still present. There was a chance that when you crossed the cats, it would be passed along and change places.
You could buy a lot of cats with really good recessive genes that looked like nothing. They were practically free, and nobody bought them. You would buy a bunch of them, crossbreed them together, and every time one of the good genes appeared, you could sell it for many times more.
I never invested a lot of money, but I figured this out fairly quickly. For three weeks, it was a lot of fun. Before that, I had been playing RuneScape for about 12 hours a day, and I never went back.
CryptoKitties running on Ethereum was relevant for a while. It was fun.
What about CryptoKitties? What is the origin of Uniswap sniping?
Uniswap sniping is funny, because that is exactly how it was conceptualized. I think people assumed it wouldn’t work, but for us, it was very much like a game.
You launch a token and put a bunch of it in an LP on Uniswap. If the initial float is small and you buy first, you have practically no downside risk. Your downside risk is that the liquidity gets pulled out or that you somehow get frozen.
Because the game went through iterations, developers eventually started doing this themselves. It became a whole game of trying to avoid it. At first, it wasn’t really clear, but if you just bought first, there was no risk.
You could do it with contracts so that if you weren’t first, you didn’t buy at all. You would have strict checks on slippage, and so on. The metagame started to develop around the question, “How do you make sure your transactions are ordered correctly?”
There were a lot of things that were supposedly thought out, but it wasn’t that complicated under the hood. You just needed a certain set of skills to understand this very opaque world—the token launches and all the software involved. It was very difficult to learn if you weren’t already knowledgeable about those things.
You learn this just by osmosis.
Why did you want to make this? Because you wanted to play these games?
Yes, but that wasn’t really true. It wasn’t as if I said, “Okay, I want to do this.” It was a very slow learning curve. I thought, “I wonder if this is possible,” and then I tried this, and this, and this. Eventually, you look back and think, “Oh, I guess I can do it now.” I never noticed the moment when it became as technically complicated as it is, but eventually you get there.
A lot of my early code was terrible. I was interested in blockchains, or at least Ethereum, because the block time is so long. It was 15 seconds. You could write the worst piece of code you had ever seen, and it would still be executed if it completed within 15 seconds. Everything was fine. Fifteen seconds is a very long time for a code snippet.
You didn’t have to shave off hundredths of a second. The worst fragment of JavaScript would take 200 milliseconds or something like that for any checks you were conducting. It wasn’t going to be too expensive in terms of time, and it was easy enough.
Optimization at the limit was never really about being a good developer. You didn’t have to write these things in low-level Rust or anything like that. You could just write whatever you wanted. If the logic was good, you could win. It was just game logic: implementing the idea in code. That was all you needed. The code just had to know where to go, what to do, and how to do it.
You could just squeeze it in. That’s basically why it worked. So many people who had done this before were not professional programmers. You just had to figure it out. Generally speaking, you can figure out the code. You didn’t need Stack Overflow or anything like that. Agents just do it now, but it was interesting to learn this way. That’s exactly how I learned.
Then these strange niche skills appear. There isn’t much money in them. Then summer comes: DeFi.
Which one did I miss? I missed the summer of DeFi.
Yes, it was a wonderful time. It was madness. It’s a terrible thing to miss, right?
It was a very strange epoch, a thing worth grieving. An expensive thing, after which it was worth saying, “That was a very strange time.” But it was also a lot of fun. I slept very little. I shaved all my hair off.
Oh, really? Looking back at the summer of DeFi, was there some specific economic result you were looking for, something that changed the trajectory of your life at one point?
Not really. It was the culmination of many things. I had lived a very itinerant lifestyle for the last half-decade, quite a wandering way of life. A nomadic lifestyle isn’t the same as traveling, of course.
I think I would have done it anyway. I’m not sure that anything changed significantly. Part of it was that COVID happened, and I immediately went to Sweden, which didn’t introduce any restrictions. I stayed there for 6 weeks. I left, supposedly for a week, and I just never went back. This continued through cycles of lockdowns over the next 18 months.
After that, I continued because I liked it so much that it didn’t make sense to stop. I don’t think there was any particular economic result. Obviously, there were restrictions on what I could do, and I needed to be able to afford to fly, but I don’t think I would have done it differently regardless of the results.
Can you tell me about a bright moment on the DeFi blockchain? If you have one, and if not, we can miss it. Maybe something cool that you did, or how you interacted with something you broke. A small story, for example, about a coin. The launch was quite funny.
There’s a lot of information, and I’m not sure how much of it is public, but I’ll speak about one side of it.
Basically, Curve Finance wanted to launch its token, and there were some reasons why they couldn’t, let’s say. Then a new account on Twitter called 0xChad basically deployed the contracts for them. They were all on GitHub, so you could just deploy them. He deployed them on behalf of Curve, and they were already live.
Curve’s contracts were specially written in Vyper, which is a smart-contract language similar to Python. At that time, almost no one used Vyper, and no one really understood it. But the Curve contracts were live, and I thought, “Okay.” I understood how to farm them.
For about 3 hours, I was the only one farming any CRV. You can find all of this on the blockchain. It’s still there. Where else would it go? It was very fun. I sold the individual CRV tokens for insane amounts of money because no one else had access to them. I created the initial Uniswap pool, and so on.
It was fun because if you knew how to do something, you could just do it, and everyone else was saying, “How does this work?”
How did you find out how to do it?
I didn’t know at that moment, but it was possible to figure it out. You can just work it out.
Were you in a group chat with 5 people who were also doing it, or were you alone in the room?
I was alone. It was just me. There seems to have been one other guy who tried it on-chain, as far as I remember. You can see it. But no, it was simple.
There was also the launch of Fei and Tribe, if I remember correctly. That was a stablecoin. The launch was very funny because of how it was developed, wasn’t it?
You put a bunch of Ether in a pot, and you were given a stablecoin in exchange for it. You also received something like a 20% bonus in a token representing shares of the protocol.
If I put $1,000 worth of Ether in, I get $1,000 back in stablecoins plus 20% on top. I can invest a lot of money, and as soon as I get the dollars back, I sell them, and then I also sell the extra 20%. I just earn free money.
Yes, it really was free money. Literally free money. But the funniest thing was that for each percentage point the stablecoin deviated from its peg, there was a penalty for selling it.
It was similar to the first algorithmic stablecoin, right?
No, I mean, that first algorithmic stablecoin was like the U.S. dollar, although I think there was something earlier. It has a very ancient history. It wasn’t even really an algorithmic stablecoin. It didn’t have that kind of expansion mechanism, and maybe it was something else. It was very complicated.
They had one of the funniest pre-launch audit reports. It had something like 17 critical issues. They had a strong collapse.
But if you were the first one through it, like me, you just sold a bunch of the thing. You unpegged it so much that no one could sell anymore. I did this and then blocked everyone else, because it seemed like no one could sell at all. The penalty was extreme.
The penalty was extreme.
The whole thing just got stuck, and I was the only one who got any benefit. Good entertainment. Beautiful entertainment.
How do you feel about the trajectory of crypto disappearing? For example, cryptocurrency in 2026, where we are now.
I think a good starting point here would be to rewind to November 2024. Trump had just won, right?
I’m not even talking about Trump specifically. We can talk about that additional information later if you want. But we spent about 4 years with the government pressing its knee on our necks. Every day it was basically, “Fuck Biden.” Gensler, you know. You look at Signature Bank being closed. There were many things where practically everything you were trying to do was illegal.
In a strange way, we thrived in that environment. We coped very well. You had all these theoretical future things that could happen, and it was optimistic.
Then Trump comes along and says, “Okay, now we’re ready. We can do everything, and this will work.” It was like, “We won. We’ll win here, won’t we?” The prices rose a little parabola-like.
Then the old demons raised their heads. This is what I’ve been working on for a long time, at least internally. It felt whole. It made sense. What we created, in essence, is a small piece of global accounting software, and it works well in that role. It’s utilitarian.
Remember the heuristic: the purpose of a system is what it does. This thing is for finance. What is it for? Is it necessary or not? That’s where the value is in the chain. But if you go back even further, to the early days of Bitcoin and the early days of Ethereum...
If you look at the people who were interested in this, they were, as a rule, politically active, sovereign personalities. Many people in the early days of Ethereum were like that, right? Vitalik was incredibly well-known on the left, right? Years later, in Remix, the first contract that I ever saw was something like a contract for voting.
If you read any of Vitalik’s work, you’ll find that most of it is dedicated to politics and sociology. It’s not about finance at all. They very much wanted these systems to be integrated into the social fabric of society. But that’s not what they’re for, damn it. They strive, don’t they?
We burned hundreds of millions, billions of dollars, on these ventures, and there simply isn’t any meaning. This is never going to work. None of them really worked. There were many similar tokens from the 2017 era. There were many like this, tokens used for whatever, but also many social programs, charitable organizations, and so on.
The point is that I’m just trying to understand it. This is not what it’s for. None of these things make sense in general. That’s all. They’re now dead. People have generally stopped trying to understand it.
Does that make sense? If you rewind the systems back, the purpose of the system is to do what this thing does. This is finance. It trades. This is value on the internet. It doesn’t mean that you can’t connect it to some social movement. Of course you can. I’m following. It’s just not intended for that.
Even if you try to do this, people who are really good and knowledgeable in mathematics will simply come and sweep away all the others, as a rule. That happens, doesn’t it? So you must try it, and it makes no sense. It’s like trying to adjust the tool to the problem, when the tool was not created for problem-solving.
A very large part of the politically active class in the cryptosphere, including many major ETH holders, still want this to be true to a certain extent, right? This has to be some kind of means for the public good. Their biggest concern is that you must have the ability to run a node on the most obscure software provision in the Andes by means of smoke signals, right? This is what they’re very worried about.
But obviously, no one else does this, so there’s no point in running it, is there? I think a really good example here would be politics in Great Britain. It’s not funded very much. It’s crowdfunded, isn’t it? For example, an individual running for Parliament can raise an eight-digit amount for campaign expenses.
Oh, I understand. In the UK, it’s just not right, is it?
But 3 weeks ago, Ben Delo, who was one of the founders of BitMEX, and Christopher Harborne, who was one of the founders of Tether and effectively invented the stablecoin, donated about $100 million to political parties in Great Britain. That was more than had been donated, for example, by all the parties together in any year. It was incredible.
I look at these guys and think, “If you really want to predict social changes or accelerate social changes, then you do it: accumulate wealth and govern, and then insist on what you want.” That’s exactly what these boys did, right?
They created a pile of financial products that worked very well and changed people’s lives. I think crypto is a wonderful thing. I think stablecoins are wonderful, and they did a lot of good. Then you can try to influence the world so that it changes the way you want. But you can’t do that without first having capital that can be spent on it.
If you really want to transform cryptocurrency into something like a movement, it has to absorb finance in general, and then you’ll get a lot of power. But this is like the chicken and the egg—not the egg and the chicken, if that makes sense.
I wouldn’t say that this dream is completely dead. I think Hyperliquid manifests itself in the way they did it, and there’s great interest from the TradFi side of the world, including major exchanges outside cryptocurrency, in the potential integration of this technology under the hood.
But, yes, I think it’s a shame. I think there was a lot of potential that we actually squandered.
Your side of the game is something like memecoins, if you want to call it that. It’s completely another thing.
In general, I think that although it’s very loud—because that’s how it should be, exactly how it sells itself—it’s just not true. There’s not a lot of money, and there never really was.
No, yes, relatively. This is not serious money, isn’t that right?
Yes, I don’t think so. Importantly, the only time when it matters is that it actually overshadowed many of the things that could have been done with this. It tarred it.
“Tarred” is like tarnished with tar and feathers, right?
Yes. When many people remember cryptocurrency, they just assume that it’s fraud because their interaction with it was like that.
Do you want to hear this?
If you ask a 40-year-old person about cryptocurrency who has never used it, they just stare. But my real opinion on this matter is that it’s not—I don’t think it has value. Technology is technology, and the impact is influence.
The most interesting thing for me is when, for example, a 19-year-old guy who is very clever, has good internet access, and has never used cryptocurrency sees it for the first time. You show it to him for 5 minutes and he says, “Okay, cool. I get it. That makes sense.”
It’s almost like a meme to me, maybe a meme coin. Trump Coin is quite unfortunate. But political coins are all about finances. It should have happened anyway. Cryptocurrency, being the best means, will also make everything happen, and that’s it. It’s something like a path that it went through.
Maybe it’s strange, but this is not true. There’s a saying that science moves forward one funeral at a time, right? As soon as you have finally formulated your understanding of something, it changes, doesn’t it?
I’ve always said that for this to work, it’s simply necessary for younger people who have grown up with this to go through it thinking, “No, this is how it works now.” It works, doesn’t it? There’s no other way to reach this level of change without this.
You can’t force 40-year-olds to use a world that they didn’t create. That would be very difficult, unpleasant, and difficult. We still win when no one knows what they’re using under the hood, but they do it.
This is a game. Everything is happening slower and faster than you think. It’s just a slow process.
Your previous parallel with cryptocurrency sounded like this: at first, you accumulate it all, and then you’re the one in power, and you can influence changes—for example, to promote whatever. Where do you think it is on this trajectory, and can it still achieve the result you believed it could achieve in 2021?
Look at Great Britain, right? For example, $100 million to political parties. It changes things, right? This is essentially the beginning. Great Britain is a particularly bad example because it’s one of the most restrictive countries in the world regarding cryptocurrencies.
We no longer produce. Now it’s generally a poor country. If you don’t move with the times and change, people will come along and leave you behind.
Let me postpone that question for a moment. What happened with Europe from a productivity point of view? Why does it happen that Europe is simply left behind in the world of technology, artificial intelligence, and cryptocurrency? Thousands of articles and opinion pieces have been written about this specific problem, and everyone has their own vision.
I think it’s generally related to our education systems.
Really?
Yes, I think so. If you meet a 16-year-old boy here and he says, “I want to establish a bank or something like that,” the reaction will be, “Wow, great. How do you do that? Will you do it? You’ll go to Great Britain or somewhere else, and they’ll ask, ‘What are you doing? Stop this,’ right?”
At every stage, knowing the history of your education, you’re forced to avoid risk as much as possible. In general, this is a society with a low level of risk.
Maybe there are slightly fewer deaths among children, but when you look at the level of social change, we simply don’t educate the people who are needed to push the boundaries forward.
There’s this absurd bubble chart of companies founded over the last 50 years and their market capitalization. The United States is just a huge accumulation of bubbles, while Europe is such a tiny thing. So, for example, Spotify, and this is basically—
Oh, really? Yes. Spotify is a cool company.
But yes, we just don’t expand boundaries anymore. It’s a shame, but that’s how it worked out. There are many reasons for this, probably too difficult to consider here, but it’s sad and all that.
Maybe you have a certain survivorship bias, but what’s your train of thought regarding youth risk in 2026?
I think you now understand where the world is moving, and you just have to do it. There’s such a concept as the happy middle ground, right? You can just play it safe, and a happy life will be very calm, right?
For our parents’ generation, it worked, isn’t that right? You could just do what was expected of you, and life in general would have been fine. But the value of labor has already passed through the floor and through the basement; it’s tunneling to the center of the Earth while we’re talking.
There’s no money there. You will not be able to live a comfortable life doing what you’re supposed to do. Other than very specific channels like school or college, there just aren’t any viable paths. This has no meaning.
If you want to have the same freedom and opportunities that we probably had, it will require some risk. It will require creating something or, for example, exiting and doing something new that people will pay you for.
You will never again receive compensation for your time in such a way that you could enjoy these privileges. It’s simple: full compression all over the world. And that’s exactly how it is, but yes, you must.
And you know, this is somehow bad because you tell the guys, “Yes, flip a coin,” and some of them will throw it and won’t get anywhere, but you still want to throw it away? You have to throw it up, and you have no choice.
Would it be fair for me to describe your bank account as post-economic? Or can I use that term here?
No, I’ve never had a bank account with any significant sum. Perhaps by choice. The question is, does this make sense? The term probably doesn’t have much value. But I have enough to feel comfortable.
Cool. I thought my question was about 100 years or more—I don’t even know how long. My history isn’t very impressive. If you had succeeded at something like that, you could just hold fiat money. You could hold U.S. dollars and sit there.
Yes, sit there and do nothing actively for as long as—I don’t know—as long as the U.S. dollar exists. Maybe we have a unique perspective on this in cryptocurrency, and you’re familiar with oscillations, volatility, and rotations. Is it the next-largest store of value or something like that?
Do you think about this—for example, how to store value in this situation? I don’t know. What about AI? I mean, something that is like that, something old. I think you’re on [unclear] the second decision on Europe and Germany[?].
I think one of the other problems we have is that when people are well-off, they very quickly switch to defensive thinking. It all comes down to having finished and entering a game where you need to protect what’s left—providing food, right?
Generally speaking, it’s very bad. It’s as if you’re in some continuum and always stop at the first stop, as if you never pass the first damn stop, right? So I don’t waste a lot of time thinking about it, actually. No, I definitely think so.
There are many opportunities for people who are agentic, interesting, and ready to work on things that, compared with everything that came before, are almost of no importance in many respects, right? That helps, starting from the prospect that you have an advantage over the starting line.
But if you have a specific advantage, use it.
I mean, we discussed this a little yesterday, I think, but I’m a kind of thinker. I’m truly a doomer, dude. But that’s good, because if I’m wrong, then it’s okay to be wrong.
You must be very optimistic. If you’re wrong, who cares? And if you’re right—yes, but does it have value? Or are you not on that side? Well, either you live or you’re dying. If you’re alive, you can do it. Truly doomed.
I think I said this—we talked about Malcolm yesterday. I spent the last 8 years traveling. I traveled to a lot of places and saw the world in all its brilliance and greatness, but at the same time, in its debauched anger, as well as the worst of humanity.
That gives you a view of things that I think most people just don’t have.
You didn’t see it. You don’t know.
I also feel that, whether it’s here or New York, there’s something like the beating heart of the world. Many people live on the edge of a knife, which people conceptually don’t understand and don’t see.
Right? And this—wait, explain it. Explain this edge of a knife.
It was madness. Is there something about New York like—
Daily, is there a food supply here? If the food trucks stopped coming, it would all be over, right?
Imagine COVID itself, but COVID is really bad. It kills, you know, in the style of the Spanish flu. Young people ride trucks, and so on. If no one else wants to bring food, what happens? What is actually happening?
Yes, that’s right. And that’s true.
The answer is no. No one really thinks about this, and I’m sure they have a set of government decisions and all that. But eventually, someone has to manage these things. If no one does this, you have a problem.
The world is so badly interconnected in many ways that there are failure points everywhere. If something goes wrong, the analogy I use is that we’re sitting here flipping coins. I hate it. I hate this. I hate it.
The rate of coin-tossing grows year by year by year. Why is it growing? Because we’re more interconnected.
Yes, more interconnected. There are more risks.
Well, yes. Technology is advancing—technology trees, right? And we need to toss every time, right?
It’s like Rosencrantz and Guildenstern Are Dead. It’s a play where the guy at the beginning just flips coins. It’s always heads. He asks whether we live in a deterministic universe, or what is happening.
They think so because Hamlet ends, so their story is told. But we sit here trying to toss these coins, and if we are ever wrong, let’s flip a coin. You won’t be able to cancel this coin toss. That’s it, right? It’s all over. Tails—it’s all over.
Still, this has not happened. There’s a concept of quantum immortality, if you will, where we are simply in the universe where we never turn up tails, perhaps. We can’t die. But it seems that in the end, you probably turn up tails, and it’s a game.
Moreover, I feel that in many respects we’re already here. Do you remember the guy who left Anthropic a few weeks ago? There was a whole story about it—what the hell, right?
He says, “We think there’s a 10% chance that it will kill humanity.” But in my head I’m like, “How is it only a 10% chance? That’s practically guaranteed, right?”
The difference between approximately 0% and approximately 0.01%, and the difference between 0.01% and 100%, is tiny. This gap is huge here. It just feels like something that can happen.
And if it’s something that can happen, it probably will eventually happen, right? That’s how it works. At least, that’s how it works for me.
I think this is how we built society, right? Our structure of power, the government, generally redistributes what exists. As a result, society decides where it goes. But in reality, it doesn’t necessarily decide what the result will be.
The result is determined by the market. The market does what the market wants, right? The market operates on money.
So you see these massive political battles at the local level regarding data centers, right? Nobody wants a data center nearby, and so on. And still, many data centers are being built.
Why? Because politics doesn’t really dictate the overall result. If the market wants to build a pile of data centers, they will somehow build them, right?
That’s the thing. The government doesn’t really control this side of affairs. And this side of things naturally leads to a bad end. We simply have no levers of power to stop it.
In my opinion, the only levers that make sense are the really bad ones, aren’t they? Something like ultraviolent revolutions, which may already be over.
So it’s something like, “Either everything is over, or everything is over.” That’s probably difficult to perceive mentally, I think. But I don’t see an obvious counterargument that would have holistic meaning for me.
All these ideas where we slow down and stop—it’s as if that’s not what people do.
What were the counterarguments?
Being a suicide bomber is something like a battleship. You can—
Yes, you just—it’s free, right? You can’t—it’s something like free. It’s undeniable. I understand. Do you understand what I mean?
For me, it’s like, okay, the counterargument is this: it’s as if this never happened. You understand what I’m talking about?
It’s hard for me.
You are a beautiful suicide bomber. Are you really getting ready for this?
Well, I mean, you're in New York. What will happen if this damn thing—will the truck stop?
I mean, no, not completely.
Yes, but the game still lives. Another part of how I lived is that I did many things that were quite risky for many years. But I never really loved it. I was just like, “Is that it? Is it so important?”
Wait, I was going to ask you about this before, but I forgot. Of your previous question about how, in the world, you have seen things that, in your opinion, many people must see—did this change your point of view? Is there anything that you think I should do or see?
I don't think so. Spending 6 weeks in sub-Saharan Africa won't necessarily harm you. It's an interesting place.
Seriously?
This also depends on the situation. For example, what do you want to engage in? I definitely don't think most people must be doomers. If everyone were doomers—not just doomers, but doomers for the world—it would be a little bad, wouldn't it? In many ways.
That's not good either.
This is my cross, which I have to carry. It's somehow very strange and arrogant, but yes, it's just not something I think most people must do. I think if you are a person who wants to accept decisions regarding such things and influence other people's lives, then it is very important that you know these things and understand them.
But generally speaking, if this is not what you want—if you don't like sitting at night and thinking about it—then definitely don't do this. It doesn't make sense, does it?
Yes. You can bring this to people who should accept such decisions. That's normal.
Yes.
By the way, about the bad things—sorry for the complete distortion—but what we spoke about yesterday was interesting, and I want you to talk about it on the podcast.
Yes.
I just want to find out the full story. You told me about this infamous hack of $13 million yesterday.
This wasn't infamous or evil. I just pulled out $13 million, which is a little crazy.
But then you started to tell a story about how you returned part of this on Aave. Honestly, I didn't watch this very carefully. Tell me, please. Tell me quickly. Tell me the story.
This was some kind of position in aTokens.
What does this mean?
It's like a deposit into Aave. It's a deposit receipt in Aave, right? You invest money in one of these lending protocols and receive a token back. But I had to borrow against it.
So this happens, and any mathematical scenario calculations that take place under the hood simply fit exactly to the limits that can be achieved. You get a position that is perfectly solvent, right? Any change in the oracles and the position is liquidated, right? So it's as if you're on a line.
But then I sit there and don't know for sure what I just signed. I can sort through MetaMask, but it's kind of a weird, new, unusual type of transaction. I don't quite know how to cancel this, for example, because I didn't like it. There was an approval, but that's not what I signed. It was something that could be revoked, for example, with a signature.
Even if I cancel this, is it possible that this could be approved again? How do you know how it works under the hood?
I'm trying to figure it out.
At the same time, if the price falls at all, I get liquidated, and even more disappears. So there's a whole problem here.
What was the position?
MKR against dollars, I think, on Aave mainnet. It was just some crazy impulse, like, “I feel like I should just [expletive] with this position a little so that it was safe, take everything that I need from the address, and then I can sit down and leave.”
Okay, that just happened?
Yes, that's what it is. In the end, they got me. But yes, we live, we learn.
What do you think the future of this archetype is—something like an on-chain pirate? Is that a fair word?
Pirate on-chain.
You came up with this?
Yes, just now.
I like it. I like it.
What a future does this have?
I think this is a little different now. This appears to be a person's class or style, because you also had something like the pirate developer, I think. That no longer exists.
Why?
Because I think there is now a very large cohort of venture capital investors who didn't really exist before. If you find a young developer who creates something, he would just create it and lay it out.
“Welcome to YC, buddy.”
Yes, and now you get all these people who are ready to learn everything. Malicious venture investors are wandering around in search of their own thing.
So pirate developers, using this analogy, created many things that allowed on-chain pirates to achieve success in this way. That simply no longer exists. There isn't as much opportunity for a single player.
It's like evolution. Maybe they need to start educating the exploiters. I'm saying it again, and maybe some people are watching this, laughing and thinking, “Of course, it still exists, my friend.”
Even when I was there, very few people understood what was happening. Maybe that's changed now, but yes, the game has changed somewhat. At least what I did is not as large as it once was.
But maybe now there is something else that is more hidden and casual.
Because this is all a dark forest, right? It's as if the edge exists solely because no one else knows about its existence. As soon as someone else learns about what you're doing, the whole game hides as much as possible so that no one can see it.
That's not very allowed.
Everything you need is to go and find it precisely in order to do dark arts on-chain.
What do you mean by dark arts on-chain?
For example, learning how to work, what to pay attention to, exactly what types of contracts exist, how to use them, and so on.
There's a whole game being played.
Yes, I think this specific style probably died, but the game has evolved since time immemorial. It's not like this particular copy disappeared and nothing else will appear.
I'm a little sad that you no longer spend 18 hours a day on-chain.
Yes, me too. It was fun while it lasted, that's for sure.
Of course.
But we all end up growing up. The Lost Boys in Peter Pan should return home.
You've never seen this before?
What?
Peter Pan? You've never seen it?