438: Constellation Software Keeps Winning, Celebrating Declining Businesses, Microsoft Office Copilot, IGTV, SpaceX, and Rick & Morty
"increase your luck/optionality surface area"
Inspect every piece of pseudoscience and you will find a security blanket, a thumb to suck, a skirt to hold. What does the scientist have to offer in exchange? Uncertainty! Insecurity!
🚢🎂⚓️ Tomorrow, July 20th, will mark the 3-year anniversary of this steamboat.
I created it on a whim, as a hobby.
And here we are.
You never know how things will go. The more open you are to trying new things, the more you increase your luck/optionality surface area.
It doesn’t mean that you should randomly flail around — you need to filter and curate ideas — but when something seems interesting or like it may make you grow, give it a try, especially if it’s a low-cost reversible experiment. Don’t pre-judge that “it won’t work” before you’ve even started, and even after you get going, give it some time before you judge the results.
Most worthwhile things take a while to really get going, whether relationships, exercise, learning about something new, acquiring a new skill, etc.
btw, if you want a good example of how far generative AI has come, compare the image above with my first attempt at making a steamboat in Edition #327, a short 11 months ago.
🕹️👾 Life is a lot like a video game.
It’s not a coincidence, it’s the other way around…
We took elements from life to make our games. We just made them more fun, with shorter feedback loops, higher odds of payoffs, and simplified simulacra without a lot of the messiness of reality.
In real life, you may grind on a problem for 6 months without any idea if you’re going in the right direction. In a game, if within 15 minutes you aren’t sure what to do, you can be pretty sure the game designers will have included a bunch of fairly obvious clues to steer you in the right direction (or you can Google how to get unstuck — which you can also do in life a lot of the time!).
Advice on how to play the game of life: play the game as it is, not the game you wish it was.
If you think the game isn’t how it should be and complain a lot about it, you may feel morally superior, but you’re going to get run over by those who play the actual game.
I’m not saying these people are doing anything unethical, just that they’re better calibrated and so their effort is more effective — they have less power leakage.
It doesn’t mean you can’t think about how the game could be better, or try to change it through raising awareness and targeted actions. But you should still aim to be effective.
Often this can mean finding a different game to play.
If you believe that X should be meritocratic but it isn’t, you can rail against that until you’re blue in the face while being miserable playing X, or you can instead go play Y instead because it’s a better game for you where you can be more effective and happy.
🎥 🎬 A follow-up to Edition #436 where I wrote: “How long until we have an action film where the bad guys are selling black-market GPUs to China? 🤔”
Long-time supporter and friend-of-the-show G.R. (💚 🥃) emailed me:
Frank Costello (Jack Nicholson) in “The Departed” sold stolen chips to a Chinese buyer!
“If you wanna nuke Taiwan anytime in this century, you better pay up!”
💚 🥃 🐇 Ideas are the most valuable thing there is.
Anything you do starts with an idea.
If you want to inject more ideas and exploration into your life, please consider becoming a paid subscriber to get 1-2 additional editions full of goodies each week:
🏦 💰 Liberty Capital 💳 💴
✨ Constellation Software ✨ Gets Gift from Antitrust Regulators 🏘️ 💼 ⚖️ 💰
Intercontinental Exchange is trying to buy Black Knight, but due to regulatory issues, they have to divest some assets. Big juicy assets. 🧃
Guess who’s perfectly positioned to quickly wire money over to this forced seller?
What’s funny (as friend-of-the-show Tobi points out) is that Optimal Blue is one of the assets that the ICE CEO pointed to last year when asked why he was enthusiastic about the deal. So much for that…
As for the terms of the deal, they seem very good for Constellation.
They’re only putting up $200m in cash, the rest is a $500 promissory note with 7% interest and a 40-year term! And they don’t have to start payments until 5 years after closing! (h/t friend-of-the-show Pearnick)
It’s estimated that Constellation is paying 7x Adj. EBITDA or 3.3 EV/S for an asset with 47% adj. EBITDA margins. Optimal Blue may be in a tougher cycle right now, but it’s not in secular decline or run-off or anything like that (as far as I know — more below on declining businesses). I’m a little surprised that Roper or Thoma Bravo didn’t swoop in and grab that one…
Basically, Constellation gets what seems to be a very high-quality asset at a fire sale price. It’s levered up with very long-term debt that allows them to meet IRR hurdles, and the debt will help provide some tax benefits as the cherry on top.
I don’t know what the terms are on the debt regarding early repayment, but if they are fairly flexible, there’s even a scenario where interest rates fall in a few years and they can refinance at better rates and do even better on the acquisition.
On top of everything (because there’s more!), this divestiture increases the chances of the deal between ICE and BKI going through, which would mean that Constellation’s other deal with them for Empower — estimated at $190m — is also more likely to close! Reflexivity ftw!
h/t friend-of-the-show Jerry Capital (💎🐕) for the screenshot at the top
📉⚰️🪦 Should Declining Businesses and Run-offs be more common (and accepted, even celebrated?) 🤔
Friend-of-the-show and supporter Matt Reustle (💚 🥃) sent me an email to discuss this line from Edition #437:
"Not every experiment has to last forever."
Our discussion made me think of the broader implications of this line of thinking.
Most businesses have a strong urge to try to live forever even if it’s not really in the cards for them.
This can be a side effect of the agent-principal problem, where management is getting paid a lot and so wants to keep that going even if sometimes, the most rational and profitable thing to do would be to accept a decline and manage it as well as possible rather than to flail around, trying to recapture some past glory, incinerating capital in the process.
If a business runs out of high-return places to invest capital, returning it to shareholders allows them to make decisions about where else to put their capital to work. Shareholders may have better options than management since they can move from one industry to the next more easily than some old, structurally bad business can suddenly reinvent itself and become competitive in a totally different industry.
👀📎 Microsoft prices Office Copilot at $30/month 🤖
Microsoft is to charge $30 a month for generative artificial intelligence features in its widely used productivity software [...] Companies using enterprise-grade versions of Microsoft 365 currently pay $36 a month for each user of the E3 version and $57 a month for the E5 version. [...]
For customers who sign up, the new features are set to add a hefty 53-83 per cent increase to the average monthly cost of business-grade versions of the Microsoft 365 service, a suite of software formerly known as Office 365 that is used by hundreds of millions of workers.
This compares to $20/month for ChatGPT Plus and $19/month for GitHub Copilot.
Here’s how Satya Nadella thinks of it:
“I would think of this as the third leg” of Office, he said, after applications such as Word and Excel and cloud services like Teams… he claimed the new AI features “are the same class of value”, automating routine work and increasing productivity. [...]
“Every time you get a spreadsheet you get essentially a junior analyst with it you can ask questions,” Nadella said. “It’s like having an analyst on demand.” (Source)
I think for many people it will be too much and they’ll be fine cutting & pasting stuff from ChatGPT or Bard occasionally.
However, for some workers who are deep into Office software and paid a high hourly wage to make important decisions that affect a lot of company resources, paying $30 per month to be a little more productive and a little bit smarter can be a no-brainer.
For example, if you’re paying someone $50 or $100 per hour (or more) to do white-collar work in Office, if this Copilot can save a few hours per month and make the work noticeably better (more polished, fewer mistakes, better logic, whatever), then that seems like a pretty good ROI.
The benefits over cutting & pasting from ChatGPT are that by being integrated into the software, you both save time and have the ability to have the Copilot do more than just input text — it can also interact with the rest of the software, help you use advanced power-user features or even discover features you didn’t know about, etc.
I wouldn’t be surprised if the AI models used for Office were fine-tuned and used user signals for RLHF to get even better at those specific tasks than more general LLMs.
In related news, MS also announced “Bing Chat Enterprise” with some assurances of data privacy and protection for enterprise customers. It’s rolling out at no additional cost to Office customers and is no doubt another way for MS to try to drive usage to Bing. We’ll see if that moves the needle… 🪡
📺📲 Remember IGTV? 🪦
I recently read something that reminded me of IGTV (Instagram TV). I hadn't thought about this one in a while and had to look it up to see what happened.
It was launched in 2018, with the goal of being a standalone app to compete more directly with YouTube. At the time it was a big shock & awe launch, with tons of hype.
It was shut down in March 2022 to focus on Reels, which was probably a good idea — even YouTube has become more like TikTok since.
Again, I’m not saying Threads will have a similar fate. I’m just being reminded of these interesting things I hadn’t thought about in a few years. I don’t know yet what will happen with Threads, but so far it has had a very impressive launch. I just wish they’d improve the timeline because right now it sucks.
SpaceX 🚀 📡 🛰️🛰️🛰️🛰️🛰️🛰️🛰️🛰️
The Information reports:
Elon Musk’s Space Exploration Technologies, the most highly valued private tech company in the U.S., has told some investors it expects to bring in about $8 billion in revenue in 2023, roughly doubling its revenue from the previous year, according to people familiar with the discussions.
Ok, how about valuation:
The company, valued in a secondary share sale at about $150 billion this month
That makes it more valuable than “publicly traded Boeing, Raytheon and Lockheed Martin, all aerospace companies with more than $60 billion in annual revenue each.”
Of course, these are not growing like SpaceX is, but it still gives a sense of scale.
Here’s the head-scratching part:
assured investors it expects to pull in about $3 billion in operating profits this year, at least by a measure that excludes expenses tied to building rockets and satellites.
I wish I had more details on what they mean because excluding these expenses seems a bit… unjustified?
Something may have been lost in translation between the person who saw the presentation and the journalist.
SpaceX’s satellite internet business, Starlink, which opened up pre-orders to the public in 2021, has also started to comprise a significant portion of revenue, the people said. The company has said publicly it has surpassed 1.5 million Starlink customers, at least some of whom pay $120 a month for a standard subscription
🗣️ Interview: Jack Raines on ♾️ 🔁
📈 Wages vs Inflation in the US 📉
🧪🔬 Liberty Labs 🧬 🔭
💥 BOOM 💥
I don't remember where I first saw this photo years ago, but Noah Smith (🐇) just reminded me of it — what a cool shot (literally and figuratively)!
It gives an idea of the kind of shockwave created by these old-school naval guns. In fact, it makes me wonder if the sailors on these huge battleships sometimes got concussions from their own guns 🤔
I was also recently reminded of the excellent memoir ‘With the Old Breed’ by E.B. Sledge. He was a marine during WWII and participated in the island-hopping campaign against imperial Japan in the Pacific Theater. These were largely coral islands where you often couldn’t even dig a foxhole, being on the receiving end of both naval artillery and ground-based artillery. Yikes.
🔌🚘 Can the Grid Handle EVs? ⚡️
Some of the basic math and systems thinking behind what it would mean to transition transportation to EVs in the US.
⚓️ Roman shipwrecks are mined for lead to build neutrino detectors ☢️
Ethan Mollick explains:
Lead is a great radiation shield, but refined lead has small amounts of radioactive ²¹⁰Pb isotopes that mess up advanced neutrino detectors. Only lead from Roman shipwrecks is old enough for the impurities to decay
'Millions of sensitive emails have been missent to Mali's .ML domain due to people mistyping the US military's .MIL domain' 🤦♀️
Millions of US military emails have been misdirected to Mali through a “typo leak” that has exposed highly sensitive information, including diplomatic documents, tax returns, passwords and the travel details of top officers.
The problem was first identified almost a decade ago by Johannes Zuurbier, a Dutch internet entrepreneur who has a contract to manage Mali’s country domain. [...]
Control of the .ML domain will revert on Monday from Zuurbier to Mali’s government, which is closely allied with Russia. When Zuurbier’s 10-year management contract expires, Malian authorities will be able to gather the misdirected emails. (Source)
We can invest billions in cybersecurity, zero trust, pen testing, etc, but sometimes it’s the simplest stuff.
🎨 🎭 Liberty Studio 👩🎨 🎥
📺 Rick & Morty: Origin story and how the show is written
A fun peek behind the scenes at this absolutely insane show (in a good way).
🎥🎬 ‘Denis Villeneuve to Shoot HBO Mini-Series ‘The Son’ This Fall, Jake Gyllenhaal Cast as Lead’
The series is based on a book by Jo Nesbo, a very famous Norwegian author.
The project is positioned to start production in October/November, with Jake Gyllenhaal starring in the role of Lofthus, an escaped convict and opioid addict on the run from the law who can’t recall his past.