Great article my father passed away this year at the age of 90. He was the first masters recipient in probability and statistics at Oregon State University. Our favorite phrase was “what are the odds” 🙂. I am also a professional poker player which is one of the hardest disciplines to learn. I am constantly failing but I am also constantly learning due to those failures and one day of lack of discipline ruins 100 days of following disciplines since human tendencies seem to want to break the rules we set for ourselves. It’s a great recap that we are not betting to necessarily win but to follow our rules and once we do that, we are at peace with the outcome. Jesus Ferguson was a great poker pro and they asked him one day how much of poker is luck and how much is skill. He said over the course of one tournament it’s 90% luck but over the course of a year it’s 90% skill but we must be disciplined in order for these odds to play out. You had a great father teaching you these things and so did I🙂
Nice work. Very well written. I agree with your findings and was taught by my father who taught me how to bet on horses from a very early age. I teach very similar thinking at TradingView where I help people learn how to invest. I’m 60 now and have been investing since I was a teenager and worked on Wall Street too after college.
Same here with my DAD and brother, We grew up 9 blocks from Santa Anita and I starting trading at Age 26 in SF and 53 now with more system and complex system analysis that allows me to have an edge in the markets , plus these new predictions markets and most recently , building AI agents to execute my models. SML over LLM with ML pipeline.
I usually distill a few key points from your pieces. This time it is: "constantly reassess the odds with new information". This is Bayesian statistics, formally. When I was in college for engineering, statistics was my fave elective. After getting a terminal degree, I found myself holding statistics dear, and publishing works that relied on Bayesian inference.
But take everything I say with a grain of salt... because I worked for decades for my 120+ HRV and good health, and dove into Bitcoin as quickly as I found out about it. Cheers.
Check Energycostmodel.com and add on opportunity cost from large mining companies who could pivot many assets to instead use GPUs to mine high performance compute. Non bitcoin miners also can use GPU mining nd most in the crypto ecosystem converts to stablecoin or bitcoin.
Kardashev scale is relevant but that underappreciates Buckminster Fuller’s Empheralizarion.
In addition, if bitcoin is a perpetuity and claim on the per Joule value of energy, three things govern its value. One is energy costs increasing. The other is the value per unit of energy increasing. The other is the total energy consumption/production as multiplying the joules consumed multiply the value per joule since no one will spend joules without it adding more value in some way than it is worth.
There is a massive flywheel.
“Circular financing” is also not functionally much different than the banking function and federal reserve. NVDA is trying to become like the new federal reserve. GPUs are their bond. That bond produces yield in either intelligence or non bitcoin crypto mining (GPU mining) which will get converted to stablecoin or bitcoin. Those mining bitcoin also own valuable real estate, technical knowledge, relationship, efficient access to energy markets, access to cooling infrastructure, etc and those training new AI models will pay for their high performance compute via cloud services. This creates a tight correlation.
Fundamentally, bitcoin is a claim on the per joule value society provides and AI is driving the same increase in conversion of energy to something more productive.
The banking function is bonds plus dollars. The amount grows over time. And that function largely prices in the economic backbone or intrinsic earning capacity. Loans are made based on collateral plus earnings. The industrial power drives the growth and historically drives the military output which in turn secures the banking function and in turn the government / public money secures the banking function. It’s circular financing worldwide
AI is an extrinsic bet on the disruption of that model.
The competition between bitcoin and AI for energy accelerates energy innovation and development.
Which in turn increases the per joule value of society, which increases bitcoins value which increases the demand to mine and hash rate, which pulls money away from AI which increases the cost of high performance compute which in turn pulls it away from bitcoin increasing the rewards to mine and increasing the investment in mining it.
Bitcoin and AI and SMR/OKLO/NNE/TSLA and New wave energy create a powerful feedback loop.
In more simple terms bitcoin expands the Kardashev scale and new wave energy increases “ephemeralization” which through Javon’s paradox increases the demand for energy accelerating investment in new wave energy.
Bitcoin expands the demand side of energy, AI competes to increase the value provided at that cost per joule.
The upward pressure for AI to deliver enough value to be worth the increased cost of energy, or subsequently to pair with enough energy output to exceed the needed demand to keep price the same is perpetual and enormous. But they are making a massive bet that it will be worth the investment.
As long as you are talking Horse racing and poker, why not add blackjack. Ed Thorpe of “beat the dealer” and of Fortune’s Formula and Kelly criterion also invested himself using correlation pairs playing convergence and divergence of assets for statistical arbitrage.
And that introduces a play that would be more aligned with the Buffett/Munger reference.
You can play bitcoin and AI in so many different ways.
1)Data Center Real estate.
2)AI data center companies
3)GPU infrastructure
4) new wave energy
5)Cryptocurrency miners converting to High performance compute farming hybrids
6)Specific energy plays and curtailment rights
7)selling 30-40% OTM leap puts to benefit from the massive volatility premium and allowing for directional value add.
8)the Ed Thorpe play between all these assets or rebalancing an asset allocation to capture capital flows one way or another less betting on the theme and more betting on the movement of capital and normalization of price.
9)warrants or leap calls on crypto treasury to create asymmetric optionality particularly against your short OTM put to more or less create a synthetic futures trade on crypto with the ability to create a net credit and potential spread arbitrage.
10) those benefiting from AI the most directly or indirectly.
11) in against the economic disruption. E.g. shorting bonds (banks, financials, insurance), or long growth vs short industrials/etc.
12) regulatory reduction and overhead cost reduction plays like biotech. The cost to recover from failed drug trial by pivoting or the ability to reduce failure rate through gaming out and simulating while also reducing the research costs creates a lot of benefits to biotech.
Great framing, Jordi. Though with Blackrock and other tradfi institutions clearly "at the window," I'm a bit perplexed by your assertion that the vast majority of "real capital" is still seeing Bitcoin as a 100-1 longshot. Wasn't the thesis of your IPO Moment piece that the rotation from OG believers to people with real capital is exactly what's been happening for the past year? And that that rotation has been happening in trades of real size? And if that is the case, doesn't that mean that a lot of very wealthy people and institutions are already in your camp and not seeing this as a 100-1 longshot anymore? As a 63-year-old man who has probably oversized his bitcoin bet given his age, I would be very interested in how you would square the conclusions of these two essays. Thanks, as always for your great work.
It seems to me that it is defeatist to say I will own bitcoin because technology is changing so fast I can't handicap the actual business winners. I mean technology has always been changing fast, and there have always been winners, and they in theory should do better than a proxy for universal basic income which is what bitcoin is to me. Yes I see the argument. Perhaps UBI will even come from distributing government owned bitcoin, but it seems premature and if you have any skill or investment talent you can do better and have a less volatile tech portfolio.
If AI consumes the world (big IF), I'm not sure what that has to do with bitcoin? If we distill down your argument:
- AI would put pressure on earnings (maybe?)
- They wouldn't provide the earnings growth they have and money would flow out
- it would flow into Bitcoin because we'll need " digital money that moves at the speed of AI agents" - that's kind of a leap but assume that's true - btc is not digital money, and it most definitely does not move with speed! Your framework works better for a digital dollar which actually is "digital money that can move fast".
That being said I'm not bearish btc! But the reason to be long btc is I don't think AI based. I lean more on the debasement thesis.
Great article my father passed away this year at the age of 90. He was the first masters recipient in probability and statistics at Oregon State University. Our favorite phrase was “what are the odds” 🙂. I am also a professional poker player which is one of the hardest disciplines to learn. I am constantly failing but I am also constantly learning due to those failures and one day of lack of discipline ruins 100 days of following disciplines since human tendencies seem to want to break the rules we set for ourselves. It’s a great recap that we are not betting to necessarily win but to follow our rules and once we do that, we are at peace with the outcome. Jesus Ferguson was a great poker pro and they asked him one day how much of poker is luck and how much is skill. He said over the course of one tournament it’s 90% luck but over the course of a year it’s 90% skill but we must be disciplined in order for these odds to play out. You had a great father teaching you these things and so did I🙂
Nice work. Very well written. I agree with your findings and was taught by my father who taught me how to bet on horses from a very early age. I teach very similar thinking at TradingView where I help people learn how to invest. I’m 60 now and have been investing since I was a teenager and worked on Wall Street too after college.
Same here with my DAD and brother, We grew up 9 blocks from Santa Anita and I starting trading at Age 26 in SF and 53 now with more system and complex system analysis that allows me to have an edge in the markets , plus these new predictions markets and most recently , building AI agents to execute my models. SML over LLM with ML pipeline.
I remember the Derby piece; I Loved it.
I usually distill a few key points from your pieces. This time it is: "constantly reassess the odds with new information". This is Bayesian statistics, formally. When I was in college for engineering, statistics was my fave elective. After getting a terminal degree, I found myself holding statistics dear, and publishing works that relied on Bayesian inference.
But take everything I say with a grain of salt... because I worked for decades for my 120+ HRV and good health, and dove into Bitcoin as quickly as I found out about it. Cheers.
Spent a lot of good afternoons at Emerald Downs
Fix the money fix the world
Check Energycostmodel.com and add on opportunity cost from large mining companies who could pivot many assets to instead use GPUs to mine high performance compute. Non bitcoin miners also can use GPU mining nd most in the crypto ecosystem converts to stablecoin or bitcoin.
Kardashev scale is relevant but that underappreciates Buckminster Fuller’s Empheralizarion.
In addition, if bitcoin is a perpetuity and claim on the per Joule value of energy, three things govern its value. One is energy costs increasing. The other is the value per unit of energy increasing. The other is the total energy consumption/production as multiplying the joules consumed multiply the value per joule since no one will spend joules without it adding more value in some way than it is worth.
There is a massive flywheel.
“Circular financing” is also not functionally much different than the banking function and federal reserve. NVDA is trying to become like the new federal reserve. GPUs are their bond. That bond produces yield in either intelligence or non bitcoin crypto mining (GPU mining) which will get converted to stablecoin or bitcoin. Those mining bitcoin also own valuable real estate, technical knowledge, relationship, efficient access to energy markets, access to cooling infrastructure, etc and those training new AI models will pay for their high performance compute via cloud services. This creates a tight correlation.
Fundamentally, bitcoin is a claim on the per joule value society provides and AI is driving the same increase in conversion of energy to something more productive.
The banking function is bonds plus dollars. The amount grows over time. And that function largely prices in the economic backbone or intrinsic earning capacity. Loans are made based on collateral plus earnings. The industrial power drives the growth and historically drives the military output which in turn secures the banking function and in turn the government / public money secures the banking function. It’s circular financing worldwide
AI is an extrinsic bet on the disruption of that model.
The competition between bitcoin and AI for energy accelerates energy innovation and development.
Which in turn increases the per joule value of society, which increases bitcoins value which increases the demand to mine and hash rate, which pulls money away from AI which increases the cost of high performance compute which in turn pulls it away from bitcoin increasing the rewards to mine and increasing the investment in mining it.
Bitcoin and AI and SMR/OKLO/NNE/TSLA and New wave energy create a powerful feedback loop.
In more simple terms bitcoin expands the Kardashev scale and new wave energy increases “ephemeralization” which through Javon’s paradox increases the demand for energy accelerating investment in new wave energy.
Bitcoin expands the demand side of energy, AI competes to increase the value provided at that cost per joule.
The upward pressure for AI to deliver enough value to be worth the increased cost of energy, or subsequently to pair with enough energy output to exceed the needed demand to keep price the same is perpetual and enormous. But they are making a massive bet that it will be worth the investment.
As long as you are talking Horse racing and poker, why not add blackjack. Ed Thorpe of “beat the dealer” and of Fortune’s Formula and Kelly criterion also invested himself using correlation pairs playing convergence and divergence of assets for statistical arbitrage.
And that introduces a play that would be more aligned with the Buffett/Munger reference.
You can play bitcoin and AI in so many different ways.
1)Data Center Real estate.
2)AI data center companies
3)GPU infrastructure
4) new wave energy
5)Cryptocurrency miners converting to High performance compute farming hybrids
6)Specific energy plays and curtailment rights
7)selling 30-40% OTM leap puts to benefit from the massive volatility premium and allowing for directional value add.
8)the Ed Thorpe play between all these assets or rebalancing an asset allocation to capture capital flows one way or another less betting on the theme and more betting on the movement of capital and normalization of price.
9)warrants or leap calls on crypto treasury to create asymmetric optionality particularly against your short OTM put to more or less create a synthetic futures trade on crypto with the ability to create a net credit and potential spread arbitrage.
10) those benefiting from AI the most directly or indirectly.
11) in against the economic disruption. E.g. shorting bonds (banks, financials, insurance), or long growth vs short industrials/etc.
12) regulatory reduction and overhead cost reduction plays like biotech. The cost to recover from failed drug trial by pivoting or the ability to reduce failure rate through gaming out and simulating while also reducing the research costs creates a lot of benefits to biotech.
Etc
just buy $IREN
Super
Mr Visser, Great article, as always. I'm curious what you think about ETH and the building of a new financial system based on those rails?
Thank you Jordi, I love your stories and models. you are showing me how to be smarter and stronger in this crazy world, that´s priceless. ❤️🙏
Fantastic read 👍🙌
Great framing, Jordi. Though with Blackrock and other tradfi institutions clearly "at the window," I'm a bit perplexed by your assertion that the vast majority of "real capital" is still seeing Bitcoin as a 100-1 longshot. Wasn't the thesis of your IPO Moment piece that the rotation from OG believers to people with real capital is exactly what's been happening for the past year? And that that rotation has been happening in trades of real size? And if that is the case, doesn't that mean that a lot of very wealthy people and institutions are already in your camp and not seeing this as a 100-1 longshot anymore? As a 63-year-old man who has probably oversized his bitcoin bet given his age, I would be very interested in how you would square the conclusions of these two essays. Thanks, as always for your great work.
It seems to me that it is defeatist to say I will own bitcoin because technology is changing so fast I can't handicap the actual business winners. I mean technology has always been changing fast, and there have always been winners, and they in theory should do better than a proxy for universal basic income which is what bitcoin is to me. Yes I see the argument. Perhaps UBI will even come from distributing government owned bitcoin, but it seems premature and if you have any skill or investment talent you can do better and have a less volatile tech portfolio.
Thinking in bets. Excellent 👍🏻
Fantastic article to which I can truly relate! Thank you
So much wisdom Jordi. Thank you!
If AI consumes the world (big IF), I'm not sure what that has to do with bitcoin? If we distill down your argument:
- AI would put pressure on earnings (maybe?)
- They wouldn't provide the earnings growth they have and money would flow out
- it would flow into Bitcoin because we'll need " digital money that moves at the speed of AI agents" - that's kind of a leap but assume that's true - btc is not digital money, and it most definitely does not move with speed! Your framework works better for a digital dollar which actually is "digital money that can move fast".
That being said I'm not bearish btc! But the reason to be long btc is I don't think AI based. I lean more on the debasement thesis.