About Me

The following is a collection of ideas that I like. I came up with some (but not all) of them.

Probabilistic Chess and Go

Instead of 2 players taking turns making moves in these games, we might want to determine who makes the next move by flipping a coin. So, the probability that white makes the next move is 50%, for example. This probability can vary, for example, the game could be set up so that if player X made the last move, there’s a Y% chance that X moves again. So for example, we could have a game where it’s like a normal game, but occasionally the same player makes more than 1 more in a row (based on randomness). My intuition is that the strategies could be very unique and interesting. By introducing the luck factor, it could be much more fun to play this game for money.

New version of Poker (with an Auction)

Players are dealt 2 cards like in Holdem. There is an auction as follows. Each player at the table places a secret amount of chips in front of them. The player who placed the most chips is the winner and gets extra 2 cards. (If there is a tie, the winner is the most out of position player.) Next, ALL of the money that people put in front of them goes into the pot. (So the auction losers have to put the money in the pot as well). Then we play holdem as normal. The player with 4 cards can use AT MOST 2 cards in his hand (They could use 1 unlike in Omaha).

The possible strategies are interesting. Let’s say you bid $99, but someone else bids $100. Your $99 go into the pot, but you get no benefit, but the guy who beat you by $1, gets the extra cards.

What types of hands should you be bidding a lot with? Is there a risk that you will reveal the strength of your hand by bidding a large amount?

There are possible variations to this game. The auction round could happen on the flop for example (or on the river).


If you want outsized returns (like 100x), you need to buy investments that nobody else is buying, that are not cool. It must be something that nobody cares or talks about. Ideally it must be something that’s hard to explain. Even when you explain it to your friends, they should ideally still remain skeptical. It’s a good sign if there is a small group of smart people that is extremely obsessed with it.

If lots of people are already invested and everyone is talking about it on Twitter, it’s usually already too late. Maybe you could still get a 2x return, but not 100x.

Signs of a good investment

  • Very smart founder. Listening to him speak teaches me something new about the world, or new ways of thinking. (Examples: Elon, Vitalik, Peter Thiel, Eliezer Yudkowsky, Robin Hanson)
  • Founder is multidisciplinary. They understand a very broad range of topics.

Signs of a bad investment

  • Founder tries to imitate someone (For example Elizabeth Holmes and Charles Hoskinson tried to imitate Steve Jobs). Vitalik Buterin, on the other hand, does not try to imitate anybody and is very original.
  • Lots of secrecy. Usually good ideas are so hard to grasp for people, that there is no need to hide anything. People have an irrational fear that others will copy their idea. In general people are slow to adopt new ideas. For example, look at Tesla, nobody even thought about copying their idea for like a decade. Elon openly talked about his master plan for years and nobody cared. Same thing with SpaceX. Nobody thought about copying Ethereum when Vitalik published the white paper in 2013. Ethereum was being developed completely in the open and nobody cared. When it was finally released, nobody cared. People started copying it only in 2017/2018 when it became extremely obvious that the market is interested. People start copying you only when it’s EXTREMELY obvious that the market likes your product. The insight here is that people start thining about copying irrationally too late. It has to be ridiculously obvious.

Types of truths

There are two types of truths. One is a truth about the world. It is true regardless of what people think. For example the speed of light. Another type is a social truth. It is true if most people think it's true. For example, what the borders of a certain country are or who is the president of the US. Another exmple is how much much a certain piece of paper is worth (Ie, a University degree or a dollar bill.


There is a better system than Democracy. It’s called Futarchy. The problem with Democracy is the participants (voters) are not rewarded for making good decisions and are not punished for making bad ones. There is no feedback loop. This is not the case for stock markets or prediction markets. If you make bad decisions there, you will lose money. Futarchy applies this market principle to governance.

Investing and gambling

Investing is making smart bets. Gambling is making dumb bets. When investing, you must make bets no larger than what is suggested by Kelly Criterion. (This implies that every bet must have a positive expected value).

Ethereum is cash flow positive

Ethereum can be thought of as a business and can be analyzed like one. Revenue: customers pay transaction fees to Ethereum in order to interact with it. Expenses: Ethereum pays the validators (employees) to secure the network. Revenue is expected to be higher than expenses, so the business is profitable! What happens to the profits? They are distributed to the investors (ETH holders).

Are there more real numbers than integers?

If we assume that real numbers that can’t be generated with a Turing machine do not exist, then the cardinality of integers is the same as reals.

Time lock cryptography is underutilized

I heard someone say that they would reveal their secret on their death bed. It turns out there is a much better solution. There is an encryption scheme that allows you to encrypt a piece of data in such a way that it would take for example 20 years to decrypt (no matter how much money gets thrown at the problem. It is an under applied cryptography construction.

Applying time lock cryptography

Suppose you want to be able to pass your crypto to your heir or friend in case you die. You encrypt the private key with time lock cryptography for 5 years and give it to your heir. Encrypt it for 10 years and give it to your friend. As long as you move your coins to a new address every ~4 years, you are fully in control of your coins. If you disappear, your heir or friend can claim your coins after solving the puzzle.

Be slow when it comes to selling investments

I feel like I learned my lesson about holding longer than what intuitively feels reasonable early on. Good thing I learned it. I’ve NEVER regretted holding too long. But I have regretted selling too early. Maybe this is because I invest in things that are generally good. Many venture capitalists say the same thing. Their biggest regret is selling too early.

Here’s a good heuristic. If you are unsure about buying some investment, then you must absolutely buy. If you are unsure about selling, do NOT sell. Most people do the opposite (which is one of the reasons it's a good heuristic).

If you do not have any decisions you regret, you are not taking enough risks

This is similar to the following, if you never get caught bluffing in poker, you are not bluffing enough.

Be loose where most people are tight and vice versa

You usually get disproportionately good deals on things where most people are not comfortable spending money. For example, unlocking extra features in some app that you use for a few bucks. People prefer to use the free version of the app for years. and not enjoy all of the features. Big mistake. People are comfortable spending money on weddings which is why they are extremely expensive and overpriced.

Things worth spending extra on:

  • Expensive mouse
  • YouTube Premium
  • Premium version of apps
  • Phone. (You use this device many hours per day for many purposes)

Where you should be tight

  • Expensive watch
  • Weddings
  • Expensive clothes
  • Cars like BMW

Ask yourself, would many people be willing to pay for this? If the answer is no, then you should definitely pay for it.

I bought an app for $6. But then I realized it’s not exactly what I was expecting or needed. Then I feel guilty for not using it. Then I realize that I spent more on Startbuck, but for some reason that doesn’t bother me.

Stay away from free things

Free things affect people on an emotional level. Usually there is a catch, stay away.

Do not try to optimize small purchases

Do not spend your brain power trying to save pennies, or even a few bucks. Paying 5 dollars in a situation where you don’t really have to hurts people emotionally. People drive around looking for a free parking spot instead of paying a few bucks. That’s a terrible behavior. People are so averse to small losses, that it’s irrational. For example, in some cases it would make sense to buy 10 different pairs of socks on Amazon to try them out. Donate all the ones that don’t fit and order a bunch that fit well. People have a fear of small losses so they avoid buying socks online because they might not fit. Following my suggested strategy saves so much time and let’s you focus on more important things. In my experience gambling is a great way to reduce the fear of small losses. After you lose win or lose a few thousand at a casino, your sensitivity to small loses goes down.

Gambling is a great way to experience how crazy randomness can get

Gambling has many benefits, including the one mentioned above. Your brain is very bad understanding randomness intuitively. When you gamble, you will realize how bad your brain is at it if you approach it correctly. Whenever it feels like “there’s no way this is random”, you must tell yourself this is what randomness is. It can help you think more rationally about other aspects of your life. You will begin to see that a lot more things in life were due to randomness than you would have originally thought.

Incorrectly thinking about net worth

I see young people with like a 50K portfolio manage it like a retired grandma living off dividends. What young people are not taking into account is their future earnings from work when calculating their net worth. You can think of yourself as a machine that gets paid $100k every year from your job. That machine would be worth about $2M. And you own the machine. If you’re young, and have no money, you can think about investing as if your net worth is like $2 million. So if some good, but risky investment opportunity comes, it’s not unreasonable to borrow $20k and go all in on it. (It’s like 1% or your net worth)

How to solve traffic

There is a limited about of space on the roads. Whenever you are driving you are using some of this limited resource. How do we solve this kind of problem in similar situations? Whenever you are on the road, you are creating an externality. You should be paying for this externality. There should be a real time auction where you are paying for being on the road. Basically every road should be a toll road. Prices should be set in such a way that throughput is maximized. Question I don’t have an answer to: if we reduced the maximum number of cars on the road such that cars were always moving at least 100km/h (60mph), would that increase overall throughput? Intuitively it seems like when when we have traffic jams, throughput is much lower than the theoretical maximum.


All laws should have an expiration date. There should be a limit on the total number of laws. After the limit is reached, a law needs to be removed in order to add a new one. More precisely, the total Kolmogorov complexity of all the laws must be below than some limit.

Sunk cost fallacy

Imagine you go to a movie theater, you buy a ticket for $20. Right before going in, you realize that you’ve lost your ticket. Would you to and buy another one? You would feel a certain amount of pain, right? You might not want to buy another ticket. What if you realize that you lost $20 before buying the movie ticket, you would feel less pain, right? Even though the situations are identical! What if you realize that the value of your stock market portfolio decreased by $20? You wouldn’t care at all.

Another situation is imagine you buy $100 ticket to a show and a $200 ticket to a different show. However, you are looking forward to the $100 show because it’s better. It turns out these shows are happening at the same time and you can only go to one of them. It would irrationally hurt more to go the $100 show because it would mean you’ve lost $200.

Random walks

In a random walk, after n steps, you would visit the origin O(sqrt(n)) times in 1D and O(log(n)) in 2D. The probability of being at the origin after n steps in 1D random walk is choose(n, n/2)/2^n (the probability that you flipped the same number of heads and tails after n flips). This can be approximated to O(1/sqrt(n)). So after a 100 steps, the probability of being at the origin is roughly 1/10, or 10%. In 2D, you need to be at 0 in 2 dimensions, so we need to square up the probability, so the probability of being at the origin is O(1/n). So after 100 steps the probability of being at the origin is 1/100, or 1%. In 3D, you need to be at 0 in 3 dimensions, so we need to cube the original probability and we get O(1/n^1.5). So after 100 steps, the probability of being at the origin is 1/1000, or 0.1%. The integral of these probabilities is divergent in the 1D and 2D case and convergent in the 3D+ case, which is why the probability of returning to origin is less than 100%. This also means that if we consider some fractal dimension (like 2.1) we would return to the origin only a finite number of times.

The whole world should switch to UTC time zone

The number of problems caused by multiple time zones is unbelievable! Everything would be so simplified if everyone switched to UTC. Flights. Zoom meetings. Banking systems.

Duodecimal system

10 is such a bad base to use. Base 12 is much better (why do you think a day is 12 hours, not 10?). 12 is divisible by 2, 3, 4, 6. 10 is only divisible by 2 and 5. Not being divisible by 3 is so annoying. 1/3 = 0.3333… in decimal. 1/3 = 0.4 in duodecimal.

QWERTY keyboard layout sucks

It was designed specifically to slow you down. It’s very unfortunate that the whole world uses it.

How to improve chess

In chess, there is an armageddon tiebreaker. White gets 5 minutes, Black gets 4 minutes. However a draw conunts as a win for black. The fact that 1 the difference is exactly 1 minute is so arbitrary and inelegant. The solution is to do an auction. Both players bid for how much time they would need as black. The player that bids less gets black. For example Carlesen bids 3:22, Nakamura bids 3:35. Carlsen wins the auction, gets black with 3:22 on his clock.

How to improve chess tournaments

Do the tiebreak games before the classical games. The winner of the tiebreak games is awarded 0.25 points.

How to improve soccer

Do the tiebreaker penalties BEFORE the game starts. The team that wins the penalty shootout gets 0.5 points. At every point during the game one of the teams will be unsatisfied and will need to score. This will eliminate the situation where both teams play defensively trying to get to the penalty shootout.

Other things to write about

  • Schelling points
  • What is money?
  • Cryptocurrency presentation
  • It’s annoying that we use suboptimal standards like time zones, decimal instead of duodecimal, etc
  • Things that annoy me
  • Homelessness is caused by incentives
  • Why don’t restaurants auction are off spots?
  • Income tax is inefficient. We want land value tax.
  • I wish there was a way send micro transactions to people
  • Newspapers need to be prediction marketalized. All sources of info should have pm quotes.
  • Crypto native is like scanning a book vs text file. Electricity in factories.
  • What’s after crypto?
  • Layers of abstraction?
  • In the modern world it’s very profitable to be a generalist that keeps an eye on many things.
  • Making money in the stock market or crypto is much easier than in poker or prediction markets
  • Tesla had a 1 in 100k chance of hitting what it did according to option prices
  • Money creation is not zero sum
  • Cooperate game experiment
  • Silicon Valley is successful because people trust each other
  • To get people from an environment where there is no trust, you need to remove them one at a time

EIP 1559 Market on Polymarket

Useful sites

Which blockchain projects are actually used?

How decentralized are the validators in a given blockchain network.


Key takeaway: index investing beats most fund managers.

Key takeaway: betting half Kelly is almost as good as full Kelly in terms of expected value, but the variance is significantly reduced. So when it doubt, bet half Kelly.

Key takeaway: there are things that gain from disorder and volatility. A good example of this is crypto projects. The fact that they are public and get attacked all the time makes them stronger over time.

Key takeaway: when you are young, you need to accumulate money. But when you are old, you need to strategically spend money. You should be spending like crazy when you are in your sixties. Your goal should be to have zero dollars by the time you die. Do not leave inheritance to your kids, instead give them money while you are still alive. Your kids will likely be in their 60s by the time they would receive the inheritance which is way to late. Give it to them in their twenties.

Key takeaway:

Key takeaway: everything you do in life is a bet. When you decide to move to a new city (or when you decide to stay in your current city) you are betting on a future outcome. Thinking of it explicitly as a bet is helpful.