365: Constellation Software, Fusion, Copilot for Scientists, Cloudflare & Datadog, and Bizarro Star Wars
"We must always read the fine print"
It isn't where you came from, it's where you're going that counts. –Ella Fitzgerald
🤔 📙If you could only read one book for the rest of your life, what would it be?
📺 👹 🏆 This a follow-up to what I wrote about Mr. Beast’s Empire in edition #364, and based on a conversation I had with friend-of-the-show Trung Phan (💚 🥃 🇨🇦🇻🇳) about it:
The longevity aspect has opportunities and challenges.
Mr Beast began making videos when he was 13!
Imagine if he sticks with it for 20-30 more years, the kind of compounding he could get in terms of audience and skills!
But it remains to be seen if he can maintain his popularity and ride successive cultural waves (entertainment media is rather fast-changing), or if he will have his apex moment and then lose the crown.
It’s rare for entertainers to stay on top for super long, and back when they did (John Carson, David Letterman, Jay Leno, etc), the media world was a lot less fragmented. There were many gatekeepers, and few open spots, so competition was limited. Today, there’s a million people gunning for him on YouTube and Tiktok.
I’m also curious to see if he’ll succeed in building an enduring organization around him and leverage it to expand production in ways that don’t require him to be at the center of everything, or if he’ll keep it mostly as a vehicle for himself.
There are pros and cons to both approaches, and both can be successful.
💚 🥃 🐇 This is the free edition of Liberty’s Highlights with 13,700+ subscribers.
You can get 2 extra editions/week full of juicy stuff + access to the private Discord 🗣🗣🗣 community by becoming a paid supporter (it’s quick & easy).
Paid posts since last week 🐇:
🔒 363: Google vs ChatGPT, DoD Cloud, Constellation, Microsoft vs Discord, Rolex, SpaceX, and Tarantino
I really enjoyed thinking about the potential impacts of ChatGPT and similar LLM AI models on Google’s Search business and wrote a relatively in-depth segment on this. Tthe most interesting aspects are the 2nd and 3rd order impacts.
🔒 364: ChatGPT, Mr. Beast, Microsoft, Palo Alto Networks, Semiconductors, Wind Turbines, and Base Editing
Mariah Carey, cast iron, shipping container prices, and silicon challenges & opportunities for Intel, AMD, Mediatek, Samsung, Qualcomm, Apple, etc. Lots of goodies in this one.
If you click the links above, you can see the intros for free and there’s a link to get a 7-day free trial of the paid version. What have you got to lose?
🧩 A Word From Our Sponsor: Heyday 🧠
Do you have 100+ browser tabs open right now? 😬
Give your memory a boost with Heyday, the research tool that automatically saves the content you view, and resurfaces it within your existing workflows. 👩💻
It’s like cheat codes for your memory. 😲💡
🧩 Give your memory a boost today 🧠
🏦 💰 Liberty Capital 💳 💴
New Spin-Out at ✨ Constellation Software ✨ + Large Acquisition (WideOrbit)
More fancy footwork from CSI!
Its recently created Lumine Group, which is itself a part of Volaris, which is an operating group of the larger Constellation Software, is acquiring WideOrbit (I told you it was a fractal company!):
WideOrbit helps media companies do more business by making it easier to buy and sell advertising. WideOrbit is the system of record for more than $37 billion in advertising spend annually…. including NBCUniversal, Fox, Paramount, AMC Networks, Univision, Gray, and Sinclair
Annual gross revenues for 2021 was ˜$167M USD at about 25% EBITDA margins and they have approximately 450 employees.
The price is $490m as a combination of cash and 13.8% of the new Lumine. That’s about 3x WideOrbit’s estimated 2022 EV/sales and 12x EV/EBITDA. I think it’s a good sign they can buy larger assets that *aren’t* distressed.
But that’s not all! Lumine is being partially spun out to Constellation shareholders:
Under the proposed transaction structure, Constellation will remain the majority shareholder of Lumine Group and will distribute to the current Constellation shareholders, pro rata and by way of a dividend-in-kind, nominal value subordinate voting shares of Lumine Group. [...]
Following the Spin-Out and the Acquisition, CSI will indirectly own a 61.05% interest in the Company and the CSI Shareholders will own a 25.12% interest in the Company, in each case on a fully-diluted basis. The remaining 13.83% interest in the Company will be owned by the Rollover Shareholder Group.
Here’s the growth by segment at the pre-acquisition Lumine (inside of Volaris), which should do around $250m revenue in 2022 at ˜32% EBITDA margin:
The new spin listing will be on the TSX Venture.
This is similar to what they did for the acquisition of Topicus, with a bunch of complexity to be as tax-efficient as possible.
The release mentions that WideOrbit will “operate as an autonomous business unit within the Lumine Group’s portfolio of companies”. So unlike Topicus, it doesn’t sound like this was about WideOrbit doing a kind of reverse-takeover of the group when it comes to management.
In the release, Mark Leonard writes:
I hope my grandkids are still holding Lumine shares fifty years from now.
Now *that* is long-term thinking!
🧫🔬👩🔬🧪 Copilot for Scientists 🤖🥼
I hope there’s an AI company working on this somewhere because this would have tremendous leverage for improving the trajectory of humanity!
I’m imagining something similar to how GitHub Copilot assists coders, but for research scientists.
This tool would help them on both the front end when generating hypotheses and deciding what to study and designing the methodology of how to do it, and on the back end, when interpreting results and writing the paper.
It could provide help with complex and nuanced things like statistics and methodology. With how to clearly explain and visualize results, generate graphs and charts, help clarify and summarize complex sections, help make the language more precise, correct grammar and structure, make sure all replication steps are included, etc.
This wouldn’t be just about polish and form. It would be about making sure the content is as robust as possible and avoiding as many common mistakes as possible.
How to run a good experimental study and then write a good paper about it inside a piece of software, basically.
Wouldn’t that be cool? 🤔
🎙 Analyzing Cloudflare and Datadog with MBI 🚨
New podcast, still warm from the oven 🥐:
Trivia: 50% of the reason why I do podcasts is because years later I get to relisten and relive fun conversations I had with friends.
How much is that worth? 💎
Employment recovery in the US by industry
🇺🇸 ‘Coal was the largest source of electricity generation for 15 US states in 2021’ 🏭🪨🪨🪨
In 15 U.S. states last year, coal was used to generate electricity more than any other energy source. Twenty years earlier, in 2001, coal was the largest source of electricity generation in 32 states.
In 2001, natural gas was the largest source of in-state electricity generation in seven states. By 2021, that number had grown to 23 states
🇬🇧 ‘U.K. Government Approves First Coal Mine In 30 Years’ 🏭🪨🪨🪨
Speaking of coal:
The Woodhouse Colliery, owned by West Cumbria Mining, will extract coking coal for steelmaking, which ministers say would otherwise be imported.
Most of the coal produced is expected to be exported to Europe. Planning documents show that more than 80% of the coal the mine will produce annually is forecast to, after five years, go to an export terminal on the east coast.
🧪🔬 Liberty Labs 🧬 🔭
⚛️ Fusion Breakthrough! ⚛️
Lawrence Livermore National Laboratory:
a major scientific breakthrough decades in the making that will pave the way for advancements in national defense and the future of clean power.
On Dec. 5, a team at LLNL’s National Ignition Facility (NIF) conducted the first controlled fusion experiment in history to reach this milestone, also known as scientific energy breakeven, meaning it produced more energy from fusion than the laser energy used to drive it. [...]
LLNL’s experiment surpassed the fusion threshold by delivering 2.05 megajoules (MJ) of energy to the target, resulting in 3.15 MJ of fusion energy output, demonstrating for the first time a most fundamental science basis for inertial fusion energy (IFE).
This is a great scientific and technological breakthrough. This is very complex, very difficult stuff. It’s nice to see progress! Fusion nerds can read a more technical piece here.
But I don’t think we should quite celebrate the end of humanity’s energy woes just yet, as the path to commercial fusion is still quite long and arduous and we must always read the fine print:
NIF’s UV lasers use inefficient infrared amplification because they were low-cost options at the time of design and construction in the 1990s.
In the announced NIF experiment, roughly 300 MJ of electrical energy was consumed to generate only 2.05 MJ of laser emission energy to trigger fusion. That 2.05 MJ of laser energy produced 3.15 MJ of fusion thermal energy output. The result is only a ~0.0105 (or 1.05 percent) engineering gain, not including the conversion of fusion thermal energy to electrical energy.
The laser system itself only operates at a ~0.68 percent efficiency. State-of-the-art solid-state lasers could operate at 10 to 20 times that efficiency, but that is nowhere near efficient enough to reach breakeven engineering gain.
So to get the ~2 megajoules of energy from the laser which made the reaction generate ~3 megajoules, it took ~300 megajoules to charge the lasers.
These aren’t the most efficient lasers, but still…
Why is commercial fusion still far off? This piece explains it well (I recommend the whole thing)
Conversion of thermal fusion energy to electrical energy for consumption would undergo further losses in the range of perhaps 20 to 40 percent. For a NIF-like design to be practical for energy applications, and ignoring economic considerations, fusion target gains would likely need to reach the range of 20 to 80 (or 2,000 to 8,000 percent) instead of the announced 1.54 (or 154 percent).
Further, NIF can run experiments like the one it announced infrequently (no more than a few per day in the extreme), but for a NIF-like fusion design to be a practical power plant, such fusion reaction events would be required to run at roughly the equivalent of ten times per second.
That’s for inertial confinement fusion (ICF). Magnetic confinement fusion (MCF), like at ITER with Tokamak reactors, has a whole different set of very complex, difficult, and expensive challenges.
I used to think that fusion was the answer to all our problems. That it would solve everything that was wrong with fission.
But after learning more about both, it seems like what’s wrong with fission isn’t very wrong at all in the first place. In other words: What people say they would like about fusion, they can mostly get with fission today, but they’ve been convinced by decades of anti-nuclear campaigning to dislike fission.
Reality is: Waste is not that big a deal, uranium isn’t that expensive and represents a small fraction of the cost of operating a power plant and we have billions of years of supply (even today’s ‘waste’ spent fuel can become fuel in future breeder reactors and keep powering humanity), nuclear power plants are extremely safe and even the worst-case scenario of a meltdown mostly just means a very expensive piece of equipment is destroyed (few people know that Chernobyl’s other reactors kept operating after the disaster, the last one was shut down in 2000).
As for the benefits of fusion, they also come with downsides.
Fusion reactions are inherently a lot less stable and harder to keep going than fission. If you’re running your power grid on something, you don’t want to have it accidentally shut down too easily…
It’s also not clear to me that fusion would be all that cheaper than fission, at least not without many more breakthroughs.
Even if you replace uranium fuel with isotopes of hydrogen or deuterium or helium-3, you still have all the other costs of building a large, complex building with giant steam turbines, fancy high-grade concrete and steel, and huge lasers and magnets. There’s the ongoing maintenance of vessel walls that get damaged by neutrons, etc.
I also doubt regulation would be that much easier/quicker/cheaper to get fusion plants built than fission — and if there’s some much-needed reform there, the streamlining would likely help both.
I’m very much in favor of fusion research and development, I just hope it’s not a shiny thing on the horizon that distracts us from using what we have today, because there’s a lot of inertia in the world’s energy system, and we can’t wait decades to make big chances if we want energy security and abundant clean power.
💪 Chris Hemsworth and Peter Attia (Podcast)
I enjoyed this conversation between Peter Attia and Chris Hemsworth. Both participated in a documentary series on how to live better and longer called ‘Limitless’.
This is about their experience and some of the science behind it, but it’s also just a fun conversation between friends:
It made me want to watch the series. I’ll give my highlights when I’ve seen it.
🎨 🎭 Liberty Studio 👩🎨 🎥
🤖 Using AI to create a strange alternate reality version of Star Wars: A New Hope
This is trippy. Feels like the TV stations from parallel universes in Rick & Morty.