Discover more from Liberty’s Highlights
197: Amazon's 20% Rivian Stake, Cloudflare, Microsoooooft, Working from Home, 3D Vision with Analog Computing, Zuckerberg's MANGA, and a Dune Scene Deconstructed
"impressive for a bunch of wet meat inside your skull"
The main fuel to speed the world's progress is our stock of knowledge, and the brake is our lack of imagination.
🪶 This edition will be a bit lighter than usual.
Both kids stayed home sick with a cold on Friday (COVID test results on Saturday: Negative), which is when I usually write a lot of Monday’s edition, and over the Halloween weekend I ended up doing a lot of family stuff, my parents came over, etc, so little time to write…
I almost decided to skip today’s edition, which may have been the savvy “marketing” thing to do since there’s a lot of new subs because of the Infinite Loops episode I was on — 👋 everyone! — and I don’t want a first impression to be a sub-par edition.
But oh well, this steamboat is also about the ups and downs of real-life, the cycles and seasons to everything. Why pretend otherwise?
Sometimes there’s lots to write about and lots of time to write it, and sometimes there isn’t ¯\_(ツ)_/¯
Here’s the full tweet I was going for:
🎃 Since nobody asked, here’s my theory about candy distribution at Halloween:
It’s better to make a small number of kids really happy than be forgotten and thought of as stingy by a large number, so I give huge handfuls of a large variety of candy, even if it means closing shop early because I rapidly run out.
☠️ A hill I'll die on:
Companies don't "miss estimates".
Analyst forecasts miss reality...
“Missing” isn’t a value judgement here, btw.
Of course the future is impossible to predict, especially for complex, multivariate events, where dozens of internal *and* external factors on which management has no control can change the outcome. If you decide to play that forecasting game, you’re going to be wrong a lot, it comes with the territory.
But it’s still not the company’s actual operational performance that “missed” those estimates, wherever they come from, including management guidance (forecasting is hard even on the inside, lots of factors you don’t control).
I’m just saying the terminology is backwards, I’m not saying it’s a huge deal). If I try to predict the outcome of some event, like a hockey game or an election, and it doesn’t turn out the way I predicted, it’s not the event that “missed” my prediction.
Nate Silver shouldn’t say that the election missed his prediction…
(yes, I get there should be movement when there’s a mismatch between reality and the forecast because large parts of the market are basing their estimates of value on these forecasts — I’m not saying there shouldn’t be movement. I also know about all the weird games with guidance and how management uses it as a target so it has an impact on actual results, incentives to under-promise and over-deliver, etc. Even how management often “misses” on its own guidance — it’s still the management forecast that was wrong, not the actual real-world performance that is “incorrect”… This is all besides the point I’m making about terminology. I’d be happy on this if the media said: “Consensus estimates missed by 3c this quarter” instead of “Company X missed estimates by 3c this quarter”.)
💚 🥃 The price of a couple coffees or one alcoholic drink isn't a bad trade for 12 emails per month (plus 𝕤𝕡𝕖𝕔𝕚𝕒𝕝 𝕖𝕕𝕚𝕥𝕚𝕠𝕟𝕤) full of eclectic ideas and investing/tech analysis. That’s 77¢ per edition.
If you make just one good investment decision per year because of something you learn here (or avoid one bad decision — don’t forget preventing negatives!), it'll pay for multiple years of subscriptions (or multiple lifetimes).
As Bezos would say of Prime, you’d be downright irresponsible not to be a member, it takes 19 seconds (3 secs on mobile with Apple/Google Pay — if you don’t see paid options, it’s because you’re not logged into your Substack account):
iNVESTING & bUSINESS
Amazon reveals 20% stake in Rivian
I wrote a few times about EV truck maker Rivian, most recently in edition #171.
I knew Amazon had a big stake in it, and had probably gotten a lot of it as part of a huge order for electric delivery vans… But I didn’t quite realize that Amazon had a 20%-big 🥩…
As of Sept. 30, 2021, Amazon held equity investments, “including preferred stock of Rivian Automotive, Inc. representing an approximately 20% ownership interest,” which had a carrying value of $3.8 billion, up from $2.7 billion in Dec. 31, 2020, the company disclosed in an SEC filing. Amazon has invested more than $1.3 billion in Rivian to date, Rivian said in a separate securities filing earlier this month. [...]
Rivian said Amazon has some exclusive rights to Rivian’s electric delivery vehicles for at least four years, and the right of first refusal after that. Amazon has ordered 100,000 vehicles by 2030. The company plans to have 10,000 new electric delivery vehicles from Rivian on the road as early as next year. (Source)
Not only have they made a big order, they’ve basically bought the whole supply of ‘em for the foreseeable future, which reminds me a bit of how Apple locks down the early supply of every new node at TSMC with huge orders…
(20% 🥩, get it?)
‘Off the Charts’
🏡 Working From Home’s Higher Variability 🏚
I love working from home, I’ve been doing it for about 15 years and I’ve set up my life so I can do it. I know it’s for me, it works with how my brain is wired.
But I’ve also known plenty of people who it drove crazy, and they were either going nuts at home, or seeking out cafés and public places to work just to be surrounded by people (extroverts, amirite?).
But there’s another axis of variability with WFH that is more compressed in an office (for all the other flaws of that system): Most people tend to have fairly similar working environments at the office.
Some have bigger offices, some have doors, some don’t.. Some have the nice cubicle near a window, some are stuck near the bathrooms… But generally, it’s kind of all the same environment.
It lops off the best, but also the very worst.
With working from home, some have a nice quiet home office with sunlight, ergonomic furniture and a fast computer with plenty of RAM, bird are chirping on the bird-feeder by the window, Norman Rockwell is painting in the yard…
Others are sitting on a hard chair, with a laptop set up on the corner of the kitchen table in a tiny apartment, looking out the window at a brick wall, hearing the sounds of the city (sirens and diesel trucks), with a spouse also working from home in the other room, hearing muffled talk from the constant stupid meetings audible through the paper-thin-walls..
I feel like this doesn’t get enough attention, possibly because most people who have big platforms to talk about WFH probably have nice homes and the means to set things up to their liking.
Friend-of-the-show and supporter (💚 🥃) Alex Morris wrote a good post about Microsoft:
Here’s my highlights:
We’ve reached a point where Cloud, which has seen its growth rate accelerate to the mid-30’s (%) since the end of FY20, accounts for just under 50% of Microsoft’s revenues. Said differently, even if the company’s remaining businesses are unable to grow (they can and they will), the current contribution from Cloud alone is enough to drive double digit topline growth.
MS’ cloud revenue as a whole (which is Azure plus all their saas stuff like Office 365) grew 9% in 2017 and 31% in fiscal Q1 2022… Part of this is a transition from perpetual licenses to recurring subscriptions, but it’s still kind of bonkers to see these massive business lines grow faster as they get bigger.
In addition to expanding the paid seats TAM, Microsoft has improved the O365 Commercial value proposition (Teams is a notable example, with telephony / voice via Operator Connect becoming an important contributor as well).
That’s the thing, right. It’s very different to only try to sell basically the same thing to more and more people, or to sell expanding/improving product lines to more and more people and the same people too.
Being able to offer more to existing customers and benefit from that distribution power is one of the main way these huge software businesses defy the odds, because the CAC is so much lower on an existing customer.
If your saas offering is basically bundling a bunch of things that the customer used to have to pay for separately, the value proposition is higher and so you can charge more. And because of your economies of scale, you can likely do these ancillary things more cheaply than your thousands and millions of customers could separately, so that creates a nice pricing umbrella where everybody can feel like they’re getting a good deal.
Concrete example: When you move from a perpetual copy of MS Office that you install and run locally to the saas version of O365, you don’t just replace X with X.
You’re replacing X with what is hopefully a better product, but you’re also probably replacing some low-value-add IT costs like just having people install software, patch software, do customer support, security on your own servers, etc. The bundle that MS sells you with O365 isn’t just the office stuff you type documents into, but the hosting and security part too, freeing up resources for higher-value things (like maybe have IT people that focus on writing good software of your own…).
It’s really impressive that Azure, with its estimated $30bn run-rate, basically grows as fast as Cloudflare at 1/60th the size.
This is a good example of base-rate-defying durability of growth by the new global digital businesses that have few historical analogs..
Is Zuckerberg an Anime fan?
sCIENCE & tECHNOLOGY
🛣 👀 🧠
I was thinking about how we perceive the 3D world through our eyes and brains.
We all know the “answer”, which is that we have two eyes, and there’s a distance between them, and we can tell how far things are based on the angle that our eyes point at and how the lenses inside our eyeballs are focused and all that…
But that’s still a superficial answer that has a lot of holes.
Think about how you would do this digitally, with cameras and GPUs and CPUs and software… The amount of calculations required to determine the 3D shape of a scene based on the differences between those two 2D images.
But our brain does all this, meshing the two images and giving us that “feeling” of 3D space and distance entirely in an analog way.
It reminds me of the early analog computers of Vannevar Bush mentioned in the excellent biography of Claude Shannon by friend-of-the-show Jimmy Soni. They would basically be able to do complex math on their “differential analyzer” by having a bunch of mechanical processes take place (spinning discs and metal balls and levers and all that — I don’t remember all the details of the descriptions, and I don’t understand how it worked well enough to explain it here).
Our brains somehow take two very weird images (because what we “see” in our mind and the raw feed of what our eyes see is pretty different — for example, there’s a blind spot in each eye where the optical nerve loops through the retina that we don’t even consciously see), process them with some analog process to mesh them together into our conscious field of view, and this field of view has 3D information encoded within it so that when we look or think of something specific, we “feel” how big it is and how far away it is.
A big part of 3D vision also has to be memory. We have a good idea of how big a house is, how big a tree is, how big a car is, how big an ant is… So whenever we see something we recognize, we can calibrate based on that. That’s why it’s much harder to know how big things are on a view without any recognizable element.
Of course that process isn’t perfect. Optical illusions are the best proof of that.
But it’s still damn impressive for a bunch of wet meat inside your skull.
tHE aRTS & hISTORY
Denis Villeneuve Deconstructs a Scene from Dune
I wish I had this for the whole film.
Hopefully, there’s a good director’s commentary out at some point (though I don’t have anything to play physical media like DVD or Blu-ray in my house, I hope there’s a digital version of the commentary).
There’s also this recent podcast discussion between Villeneuve and Christopher Nolan that I thought was kind of fun, mostly about Dune, why Villeneuve wanted to do it, how the project took shape, etc:
I Saw it Again (in theater, of course)
Update: I saw “Dune: Part 1” (because remember, there’s half of it that we haven’t seen..) again in theater on Sunday.
It was my second time, and my wife’s first. I was looking forward to sharing the experience with her, but wasn’t quite sure how much she’d like it. Our tastes overlap a fair amount — she liked Deadwood too! — but not 100%, so there’s always a chance…
But she loved it! Mentioned the music being amazing a few times, the actors being great, the visuals, etc.
This viewing wasn’t on an IMAX screen (though it was AVX), and I did feel like it lost a bit of its power during the sandstorm scene, but overall, I’d say I enjoyed it pretty much as much as the first time, which is a lot to say for a film seen twice in a period of 8 days.