476: Boring Success, Microsoft CapEx, LVMH, Nvidia, Data Center Electricity, Spotify & Joe Rogan, Tesla Model 3 Refresh, and Dragon Prince

"That’s the prize you want."

Study hard what interests you the most in the most undisciplined, irreverent and original manner possible.

—Richard P. Feynman

Prompt: The gold medal winner on the podium is such a nice person, everybody likes them, a good guy

🥇🏆🤼 If you want to compete with your peers on something —

Compete with them on kindness, compassion, helpfulness, curiosity, enthusiasm, energy, creativity, resilience, grit, integrity, honesty, adaptability, and passion. Compete on giving a damn, deep understanding, elevating the standards, on creating opportunities and win-win situations.

That’s the prize you want.

That’s the medal worth something!

I couldn’t care less if it sounds like a cheesy motivational poster. It’s real, it’s the way to go. Take some steps in that direction *today* then do the same tomorrow.

🛀💭 🤲 🖐️ Thought experiment:

What if humans — through some different evolutionary path — didn’t have the manual dexterity and/or coordination to type on keyboards above 10 wpm or play piano or guitar? Or shoot a bow or rifle accurately? Or paint or sculpt like the great masters? Or do very precise lab experiments with scalpels and pipettes?

I’m not necessarily thinking of something quite this bad:

I suspect even a 15% or 25% lower ceiling on manual dexterity would have *huge* ripples across our whole civilization. The whole world would be different.

💚 🥃 🚢⚓️ If you feel like you’re getting value from this newsletter, it means the world to me that you become a supporter to help me to keep writing it.

If you think that you’re not making a difference, that’s incorrect. Only 2% of readers are supporters (so far — you can help change that), so each one of you makes a big difference.

You’ll get access to paid editions and the private Discord, and if you only get one good idea here per year, it’ll more than pay for itself!

---

A Word from our Sponsor: 💰Watchlist Investing 💰

Investors face two main problems:

1. Identifying good businesses 🕵️‍♂️

2. Keeping track of them over time 👀

Watchlist Investing is a monthly newsletter devoted to studying great businesses and helping readers be ready to pounce when Mr. Market gets irrational (🤪).

Independent Research: Benefit from 40-50+ hours of primary research on each Deep Dive

Regular updates on Watchlist companies: Don’t miss opportunities by staying on top of important business and industry changes!

Other Benefits: Private Google Meetups, Detailed sum-of-the-parts Berkshire Hathaway valuation, Private Discord server, Subscriber introductions

Watchlist Investing founder Adam Mead spent over a decade in commercial credit, has skin in the game as a value investor, and is the author of ‘The Complete Financial History of Berkshire Hathaway’. 📕

⭐️ Sign up here and get 20% off your first year with the coupon code “Liberty20” ⭐️

Here’s a free taste from the back catalog:

• Deep Dive on Berkshire Hathaway - 2023 Update (BRKB)

• Deep Dive on Boston Beer Company (SAM) 🍺🍺 + Q3-2023 Update

• Deep Dive on the Electric Utility Industry

---

🏦 💰 Liberty Capital 💳 💴

My best attempt at illustrating succeeding in a boring way. It’s ok if it still looks exciting to you! 💚 🥃

It’s Okay to Succeed in a Boring Way 🥱💰🏆🎖️👑

I invest in individual companies because I enjoy the process. I have beaten the market so far, but at the cost of a lot of my time and brain cycles (on top of a lot of luck 🍀). If we adjust for that, my outperformance would be small compared to low-effort alternatives.

BUT

I still feel it’s worth it for me because of the direct and indirect benefits: the people I’ve met, the things I’ve learned about that have been useful in other ways, the good ideas at the intersection of fields, etc.

That said, it’s a conscious choice I’m making. I think that for the vast majority of people, the aim should be to succeed financially in a pretty boring way.

Low-cost index funds, dollar cost averaging, paying yourself first and aiming for a high saving rate from a young age, learning skills to increase your earning power, controlling your expenses (pay up for things that truly make your life better, cut what doesn’t), etc.

🥦 Broccoli vs the Lottery 🎟️

It’s boring, but *it works*.

The hardest thing about this may be that it’s unexciting, and so it’s not attractive to many people. Or they get it, but eventually fall out of it and switch to crypto of WSB calls or whatever.

We need to get back to basics. Boring basics.

Eat good food, get a good night of sleep, drink plenty of water, move your body, lift heavy things, get some sun, and consistently save money into long-term low-cost diversified vehicles over your life.

This should be table stakes, this should be the default for everyone. Then if someone knows what you’re doing, they can deviate from this to try to do better.

Sadly, the default is much worse, more ad hoc and random… It’s an uphill fight to even get to safe and boring.

It’s a low-hanging fruit in this world that you can do much better than most people simply by doing the obvious — no secrets to discover! — and avoiding doing the dumbest things.

But here we are ¯\_(ツ)_/¯

🏃‍♂️🔑 Success in one area can free you to go wild in other areas 🔓🤪

Remember, this is very important: It’s not because you’re boring in one aspect of your life that you have to be boring elsewhere.

In fact, building a strong financial foundation and taking care of your health should allow you to take more risks and do more interesting things elsewhere!

🔌 Electricity consumption from data centers, AI, and crypto could double in two years 🤖⚡️

The International Energy Agency:

Electricity consumption from data centres, artificial intelligence and the cryptocurrency sector could double by 2026.

After globally consuming an estimated 460 terawatt-hours (TWh) in 2022, data centres’ total electricity consumption could reach more than 1 000 TWh in 2026.

This demand is roughly equivalent to the electricity consumption of Japan.

A thousand terawatt-hours!

That’s a lot of juice!

It’s not too surprising when you look at the amount of capex that the cloud players are forecasting (see below👇).

🤖 Microsoft will spend $45bn in CapEx in one year 😲💰💰💰

Friend-of-the-show and supporter Alex Morris (⚾️🧢💚 🥃 ) writes:

"[Microsoft's] FY24e CapEx will likely be around $45 billion, comparable to Microsoft’s cumulative CapEx spend over the 15-year period from FY02 to FY16."

That’s bonkers, but not unusual these days.

Google and Amazon are spending similar amounts, and the smaller clouds are also all joining the frenzy. The best proxy is to look at Nvidia and AMD’s data-center revenues on the GPU side.

Generative AI and LLMs are driving a lot of this. Unsurprisingly, there’s a lot of R&D effort going into making these models less expensive to run:

[Microsoft] has formed a new team to develop conversational AI that requires less computing power compared to software it’s using from OpenAI [...]

The GenAI team develops small language models. They mimic the quality of LLMs such as OpenAI’s GPT-4—which powers ChatGPT and Microsoft’s AI Copilots—while using far less computing power. GenAI is separate from another Microsoft team, Turing, which has developed large AI models to improve Bing and other Microsoft products, said the person with knowledge of the situation.

Many approaches can be mixed to reduce costs. The simplest queries can be sent to smaller models and more complex ones can be kicked upstairs to full-size models like GPT-4.

The models themselves, big and small, can be optimized in various ways to use less RAM and require fewer GPU cycles per token generated. This can be done with techniques like pruning, quantization, knowledge distillation, sparse training, Low-Rank Adaptation (LoRA), etc.

A good demonstration of the rapid progress there is how much faster and cheaper GPT-3 has gotten over time. What was a slow cutting-edge model not that long ago is now almost free and very fast.

Expect GPT-4 to reach that point once even larger models and faster hardware are deployed.

And so it goes. 🔜

👜 LVMH Follow Up 🍷💎

This is a follow-up to the LVMH Sankey chart from Edition #474. Reader Elliot W. wrote me:

Regarding the LVMH chart, perfume etc doesn’t really need to be a profit driver because it sort of serves as customer acquisition cost.

Lower ASP means it’s accessible to more people and therefore introduces more people into the brand universe without degrading desirability of the real cash cows (handbags and other leather goods). 

Source: been studying the fashion industry recently and that’s what I’ve picked up from books, other thoughtful investors in the space. 

That makes a lot of sense to me. Thanks, Elliot 👍

🗣️🎙️ Joe Rogan renews with Spotify with NON-EXCLUSIVE $250m deal 🎧🤝💰💰💰

Rogan’s fresh deal—estimated to be worth as much as $250 million over its multiyear term—involves an upfront minimum guarantee, plus a revenue sharing agreement based on ad sales.

Under the new licensing agreement, Spotify will sell ads for and distribute “The Joe Rogan Experience” across several podcast platforms, including in a video format on YouTube, the company said Friday. Under his previous deal, the show was exclusive to Spotify.

The most noteworthy aspect of this is that the podcast will no longer be exclusive to Spotify. The full videos will come back to YouTube and you'll be able to listen on Apple or Amazon or wherever else in the open podcast ecosystem.

Is this the end of big exclusive deals in the podcasting world? 🤔

---

🧪🔬 Liberty Labs 🧬 🔭

🔌🚘🔋 Mini-review: Refreshed Tesla Model 3

Seems like a nice evolutionary, iterative improvement, which makes sense for Tesla.

I’m glad they didn’t try to push a yoke steering wheel as default the way they did when the Model S was redesigned in 2021.

I have no problem with them offering weird stuff *as an option*. But some choices are just too opinionated to be made the default. Thankfully, Tesla caved to the pressure and started offering the option of a round steering wheel on the S, at first as an expensive retrofit, but eventually a no-cost option.

🇩🇪 Germany to Subsidize 15-20 new Natural Gas Power Plants 🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🏭🔥

This is so stupid 🤦‍♀️

After shutting down one of the very best fleets of nuclear power plants in the world in the name of protecting the environment — and the plants are still right there, they could be restarted if they decided to — Germany is now going to subsidize 10,000 megawatts of new natural gas capacity:

The government agreed to go ahead with four short-term tenders for up to 10 gigawatts, translating to about 15 to 20 new plants.

It requires developers to upgrade stations to burn hydrogen sometime between 2035 and 2040, and also leaves open room for carbon capturing and storage technologies, according to a statement Monday. [...]

The measure will cost between €15 billion ($16.1 billion) and €20 billion over the next 20 years, according to people familiar with the matter, and includes subsidies for investment and operational costs. [...] Last month, the government agreed to set aside €7.55 billion for the gas power plan.

The hydrogen mandate is laughable and won’t happen — it’s just a way for politicians to save face and pretend that they aren’t making Germany’s grid more dependent on fossil fuels — or if they do implement it, it’ll be a *stupid* way to generate electricity.

Hydrogen is not a source of energy because you can’t find hydrogen that isn’t bound up with other atoms (on Earth).

It takes energy in the first place to get pure hydrogen, so it ends up acting more as a storage medium for whatever that original source of energy is.

Because there are losses with every step of production and storage (ie. it takes energy to get hydrogen from electrolysis but you never get back 100% of the energy you put in, and because hydrogen is not very dense, it must be compressed to high pressure, which requires a bunch more energy, and then because there’s no hydrogen infrastructure, you have to build a huge specialized system, which takes a lot more energy, and then because hydrogen is very “leaky”, you probably lose a bunch of it in transport…).

In other words, whatever energy source you’d use to make hydrogen, why not just use that directly to power the grid?

And take a wild guess what’s the best way to make a lot of hydrogen?

Yep, nuclear power plants.

But even if you have a large nuclear fleet and want to store the surplus energy generated off-peak, hydrogen may not be the best way. Pumped hydro storage is very elegant if you have the right type of terrain. Stationary batteries are also getting better and cheaper over time.

Building a Rube Goldberg hydrogen system makes little sense when there are better alternatives.

Meanwhile:

The expansion of gas-fired power plants is necessary because Germany is also facing a power shortage in the 2030s due to a simultaneous increase in demand for electricity, especially in the transport and heating sectors

Are you telling me it's a bad idea to, on one hand, want to electrify transportation and heating and, on the other, shut down 26% of your baseload clean electricity generation capacity?

To borrow a word from a great philosopher: Duh!

---

🎨 🎭 Liberty Studio 👩‍🎨 🎥

🐲 ‘The Dragon Prince’ — The Kid’s Show for Adults? 🐉

I feel a bit weird recommending this to you even if you don’t have kids, but now that I’m about 3/4 of the way through season three with my kids, I’m hooked and really impressed by the quality of the storytelling, inventiveness of the writing, the art direction, the world-building, voice-acting, music…

Everything!

It has more nuance than most TV shows targeted at adults! If my kids stopped watching, I’d keep going by myself.

Don’t get me wrong, it still feels youth-oriented, but I’m having more fun watching this fantasy world than anything else I’ve seen since ‘Fellowship of the Ring’ (my fave of the trilogy).

If you still have a bit of childlike wonder in you, this is the time to feed it ⚔️🛡️🔮

And it’s on Netflix, so it’s not hard to find.

(Like most shows, it takes a few episodes to fully get going, so if you’re going to try it, I recommend committing to at least 3-4 eps. It becomes more complex as it progresses)