440: Visa and Mastercard, Jensen, Breweries, Seth Godin, Danaher, Shopify, Norway, Richard Rhodes, Synthetic Data, and Geothermal
"I doubt Al Capone was a member"
Bad things can happen fast, but almost all good things happen slowly. —Kevin Kelly
🛀💭🌐 How well do I know humanity?
Yesterday, as I sat in the hospital waiting room with my son, this thought occurred to me. (Don’t worry, nobody’s sick or hurt — it was for a pulmonary function test to determine if he has asthma)
I couldn’t help but overhear some conversations, and it was a good reminder of how different people can be.
I’ve never really “worked with the public” or been in a position where I am exposed to *a random selection* of people, which is very different than if you meet a lot of strangers, but they’re all from certain slices of the population, like bankers or engineers or XYZ.
I ingest a lot of information (online, books, from art & media), but how representative is it?
Most of the people I spend time with are family and friends. That’s a rather small group of people who either have a lot in common because of genetics/geography/culture or other selection effects (common interests, common networks, common goals, etc).
How much direct experience do I have with the rest of humanity other than brief exchanges with random people at stores and restaurants? You get a sense of how people are this way, but it’s not deep.
I’ve gotten to know *a lot* of great new people in the past 3 years through this steamboat and OSV, but they’re definitely not a random sample. They’re almost all outliers.
If I could dedicate years of my life to getting to know a representative sample of people from all walks of life and spending extended amounts of time with them; people from different regions around the world, from cities and rural areas, with all kinds of jobs and backgrounds, rich and poor and everything in between, young and old, religious or not, etc.
If I did all that, how would my mosaic model of ‘humanity’ get updated? 🤔
I’m not saying that I think my model is worse than most. I suspect most people are in a similar boat, but we rarely think about it, or we overestimate our sample size and how representative it is. This creates reality tunnels.
Let’s dig ourselves out!
🗣️ 🍽️☕️ If you’re connected with Ed Thorp and you’re reading this, I have a favor to ask (I’m looking in your direction, J.A.S. 👀):
Please make a lunch between Ed and David Senra (📚🎙️) happen 🙏
It makes all the sense in the world and I’d be really happy if David got to meet the person who he thinks has “come closest to figuring out life”.
I bet Ed would enjoy it a lot too!
💚 🥃 🐇 What has value in your life?
Is it the stuff? Is it the people? Is it the ideas? Is it understanding how things work? Is it discovering new pieces of art that make you feel something? Is it playing pickup basketball with your buddies?
Hey, I’m just asking, I can’t tell you!
Everyone has to figure this out for themselves, it’s crucial because it has an influence on so many other decisions that will then shape your life. You end up in different places based on these forks in the road.
But if you value what I’m sharing here, I hope you’ll become a paid supporter so that you can get even more of it and help me keep this steamboat in the water!
🏦 💰 Liberty Capital 💳 💴
🍺 The number of breweries in the U.S. has gone parabolic!
Note that the number of breweries most definitely wasn’t zero during prohibition. In fact, it probably spiked…
But this chart shows the *official* number from the Brewers Association, and I doubt Al Capone was a member 😬 (or maybe he was 🤔)
👟🏢 Seth Godin’s ‘Strong Brand’ Test 🐮
Friend-of-the-show Ben Gilbert tweeted this brand litmus test by Seth Godin:
If Nike announced that they were opening a hotel, you'd have a pretty good guess about what it would be like.
But if Hyatt announced that they were going to start making shoes, you would have NO IDEA WHATSOEVER what those shoes would be like.
That's because Nike owns a brand and Hyatt simply owns real estate.
It’s an interesting idea. I wouldn’t take it too literally or make it too binary, but it’s a useful exercise.
I think Hyatt has a brand, but it isn’t very strong and is based on more specific and concrete things than Nike’s focus on more abstract human emotions and values (which inherently carries more easily across product verticals or even industries).
A lot of companies with strong brands have taken Nike’s approach of trying to associate themselves with concepts.
Car companies, beer companies, sugar water companies (Coca Cola, Pepsi), and Apple have all tried to tap into that, while the local furniture store may have a decent brand when you think of it, but it’s probably a lot more constrained and specific to what they do.
Does that make it not “a strong brand”? I guess it depends on how you define those terms…
Another way to think about it may be that Chipotle has a strong brand for a restaurant while Apple has a strong brand for a company. 🍎
If Apple opened a Mexican restaurant, you can kind of know what to expect 🌮, but if Chipotle made a phone… ¯\_(ツ)_/¯
💳 Visa and Mastercard: Deep Dive into the payment rails & conviction by subtraction 🌐 💵 💶 💷 💴
Here are some highlights:
Like some other companies that have held their market values this year, Visa and Mastercard are competitively entrenched cash gushers with secular tailwinds that have so far been unscathed by the macro.
Cross-border volumes, which are far more profitable than domestic switching given currency translation fees and lower rebate burdens, have ricocheted back stronger than anyone expected, to about 40% above pre-COVID levels. With most of their revenue tied to payment flows, the card networks are also net beneficiaries of inflation.
Over the last 3 years, Visa and Mastercard have grown EBITDA by 9% and 11%, respectively, even with Russia revenue (4% for both) reduced to 0, and I think there’s plenty of reason to think profits will continue to grow at a similar rate for the foreseeable future.
Nice eh? But here’s why David is great:
Rather than expound upon Visa and Mastercard’s moats, which are widely understood and appreciated by now, I thought it’d be more useful to discuss potential challenges. This is conviction by subtraction, seeing how much is left after chipping away the excess stone of competitive threats.
I wish this approach was more frequent, especially with companies that have well-understood strengths.
😎 Jensen Huang has been on the ball for so long… ⚽️
Doug O’Laughlin: Jensen is one of the longest-serving public CEOs. He founded the company in 1993. What’s amazing is that after twenty years sometimes people miss the boat and they get less good. You cannot say that about Jensen. He has definitely been on the ball the entire time. He’s been talking about accelerated compute for a decade. He’s been talking about all these societal changes.
Look, it’s happening. He’s been right the entire time. We should have listened. He’s been right about a long-term vision. Being able to see five to ten years in the future is rare. Jensen did it. He was always steering to this long-term target on the horizon.
Jordan Schneider: We talked on another podcast about James Gleick. There should be a James Gleick chapter on Nvidia when he hopefully writes his AI book. How absurd and beautiful is it that a company built to allow teenagers to play video games and render Wolfenstein ends up being the thing that ushers in an entire industrial revolution? There’s something wonderful and human about how our just wanting to play more realistic games has ended up turning into this incredible thing that’s going to change the lives of not just gamers, but everyone.
The whole thing is good, I recommend it (ChinaTalk sub required).
📊 Companies as Ships of Theseus, Danaher Edition
Friend-of-the-show Leandro made a good point:
[context: Someone was comparing the multiples of Danaher over time]
Prime example of why comparing multiples across history is misleading
The $DHR of 2013 has little to do with the Danaher of 2023, rendering multiple and valuation comparisons quite useless
Also it’s quite misleading to assume the 10-year average multiple constitutes fair value when the company has far outpaced market returns during that time frame
This can only mean that the multiple given to the company did not constitute fair value
This is something I’ve written about in the past in a few contexts (ie. stock indices).
Something may have *the same name* as before, but like the ship of Theseus on which every plank had been replaced, it may not be quite the same thing anymore.
If you’re making comparisons across time, make sure that what you are comparing is actually the same thing, or that you take the changes into account.
Danaher is a particularly good example of this because they made transformative acquisitions and spun off their legacy industrial businesses, so they are very different from what they were 10 or 20 years ago. Roper is in a similar boat.
Even if the changes aren’t as drastic, make sure you are not misled into thinking that something is more expensive or cheaper than it really is.
🇳🇴 Norway Makes Biggest Natural Gas Discovery In 10 Years
DNO ASA, the Norwegian oil and gas operator, today announced a significant gas and condensate discovery on the Carmen prospect in the Norwegian North Sea [...]
Preliminary evaluation of comprehensive data, including cores and fluid samples, acquired from the discovery well and a follow-on extended sidetrack indicates gross recoverable resources in the range of 120-230 million barrels of oil equivalent (MMboe) on a P90-P10 basis. At 175 MMboe, the mid-point of this range, Carmen ranks as the largest discovery on the Norwegian Continental Shelf since 2013. [...]
Carmen is DNO’s sixth discovery in the Troll-Gjøa area since 2021 and is located close to existing infrastructure with clear routes towards commercialization. The other discoveries are Røver Nord, Kveikje, Ofelia, Røver Sør and Heisenberg.
Norway is now the largest supplier of gas to Europe now that Putin has made Russia a pariah.
Kind of incredible that this small Nordic country of 5.5 million people has such geopolitical importance. 💪
🛒 Shopify’s Early Success 👷♀️🏗️👷♂️
Friend-of-the-show Fahd Ananta (🪳) makes an interesting point:
I feel like a big reason of Shopify’s early success was hiring former founders and builders and then giving them an ambitious mission
Companies today are almost afraid to hire founders and default to hiring career NPCs
It sounds very familiar because it’s a story you hear surrounding the founding of a lot of the most successful companies.
A group of highly-talented, highly-motivated, highly entrepreneurial and creative people feeling like they’re working on a very clear mission.
This could describe the early days of Apple and the team that built the Macintosh, or the early days of PayPal and the group that kept the company alive against all odds, or the early days of Google or Amazon.
They each had very different cultures from one another, but the talent density, ability to attract builders, and the lack of bureaucracy and BS to get in the way of a clear mission were common for all in their early days.
🧪🔬 Liberty Labs 🧬 🔭
🎓🤖 AI Models increasingly trained on ‘synthetic data’ 🤔
Anything that grows exponentially pretty soon hits constraints. 🧫
Large AI models are getting to the point where they may be running out of easily accessible *and* high-quality data for training (on top of potential legal constraints).
To dramatically improve their performance and be able to address challenges in science, medicine or business, AI models will require unique and sophisticated data sets. These will either have to be created by world experts such as scientists, doctors, authors, actors or engineers, or acquired as proprietary data from large corporations such as pharmaceuticals, banks and retailers. However, “human-created data . . . is extremely expensive”, Gomez said.
The new trend of using synthetic data sidesteps this costly requirement. Instead, companies can use AI models to produce text, code or more complex information related to healthcare or financial fraud. This synthetic data is then used to train advanced LLMs to become ever more capable. (Source)
This reminds me of how AlphaZero was able to bootstrap to super-human level of play at Go by playing against itself.
This is another way of saying that its training data set was synthetic (machine-generated), and not from humans (unlike previous models that had been trained on a corpus of existing games, at least as a seed).
For those worries about EV battery packs needing to be replaced frequently 🔌🚘🔋
EV batteries aren’t perfect, but they last a lot longer than the urban legends your Facebook uncle will tell you about…
🗣️🎙️ Interview: Richard Rhodes, author of Pulitzer-winning ‘The Making of the Atomic Bomb’ 💣
I’ve only heard the first hour of this almost 4-hour-long epic, but I can already recommend it:
Kudos to Joe Walker! 💚 🥃
Deep, substantive interviews like this need to be out in the world, and a lot more too.
Injecting high-quality information and context into the zeitgeist can only have beneficial second-order effects, even if they are unpredictable.
Fervo Energy claims Breakthrough for Deep Enhanced Geothermal Energy
I’m always skeptical of claims made by energy startups, I want to see them proven out and working commercially before I get too excited, but here’s why I’m generally bullish on geothermal energy (which I’ve covered many times, check out Edition #219 and Edition #300 for more details):
The heat is there, everywhere on Earth (at various depths)
We know how to make electricity with heat
We’ve gotten a lot better at drilling in deep and in difficult conditions in the past few decades (the whole fracking revolution, deep offshore drilling, etc)
So it’s basically an engineering problem.
We need to keep getting better at #3 to bring the cost down to a level where we can do #2 with #1.
What if we took geothermal seriously and invested in it even just a fraction of what is going into solar & wind? I’m sure progress would be a lot faster.
What makes geothermal special is that it would be a source of dispatchable, 24/7 clean energy, and we don’t have many of these.
Nuclear and hydro are pretty much it if you want clean and a high energy density. You can also overbuild a lot of wind & solar combined with mountains of batteries if you don’t mind going with the more complex Rube Golberg low-density approach that uses vast amounts of land and raw materials.
But personally, I’d prefer if the majority of the grid in most places came from small buildings (except for hydro) that aren’t dependent on the weather and don’t have production swings of tens of percents on a daily basis while we need to keep the grid perfectly balanced micro-second to micro-second.
I got side-tracked. I’m sorry, it’s just that every time I think about what we’re doing with the grid, it hurts…
Back to the geothermal startup:
The Houston-based startup uses horizontal drilling techniques and fiber-optic sensing tools to access geothermal resources that are otherwise too expensive or technically complex to reach using existing methods.
During a 30-day test period, Fervo showed that its Project Red site in northern Nevada is capable of generating 3.5 megawatts of electricity. While that’s only enough to power roughly 2,600 U.S. homes at once, it’s still more electricity than any of the world’s 40-some “enhanced geothermal systems” have previously achieved
More details here. I hope they, and many others, keep going up the learning curve and driving costs down, because we could sure use all that clean, baseload energy!
🎨 🎭 Liberty Studio 👩🎨 🎥
🐦 How the now-defunct Twitter logo was designed
A lot more thought goes into designing a good logo than most people would imagine.
This thread by one of the designers of the Twitter logo shows a lot of sketches and thought process and influences. I found it interesting, maybe you will too!
Related to this, Jimmy now has a Substack (go sign up!). He wrote a post about the origin of “X” and the story behind Musk’s acquisition and re-acquisition of the rare one-letter domain.
💣🎬 I saw ‘Oppenheimer’ ☢️
But it was late last night (on IMAX!) and I’m still digesting it and putting my thoughts together. I’ll share my review with you soon, probably in the next edition.