Welcome to Stuff I Thought About Last Week, a personal collection of topics on tech, innovation, science, the digital economic transition, the finance industry, Lagrange points, and whatever else made me think last week.

Click HERE to SIGN UP for SITALWeek’s Sunday Email.

In today’s post: the power law in advertising share; location tracking and the rising use of ad-hoc wireless networks; the possibility of capturing data center heat; DeepMind's new language model is 25x more efficient; fake vs. real 🎄; the James Webb Space Telescope; the dearth of successful companies transforming complex, analog industries; and much more below.

Stuff about Innovation and Technology
RollerBot
Wheel-legged robots have canid/humanoid form factors but with wheels instead of hands or feet. The wheels can be locked and unlocked on demand, allowing the robot to walk on two or four limbs, or drive like a car. An example is ANYmal from Swiss-Mile, which has a Jetsons-like feel to it, and is sort of adorable as it flails from quadruped to biped – a maneuver (video) that apparently takes a lot of physical space to accomplish.

Advertising Trifecta
Google, Facebook, and Amazon have doubled their share of the advertising industry over the last five years, according to GroupM data reported in the FT. The three now garner half of the ads outside of China. The global ad industry (including China) is set to hit around $763B in 2021, up 22% from 2020 and 19% from 2019. Digital is around two-thirds of the $763B market, with Google, Facebook, and Amazon controlling around 90% of it outside of China. It’s perhaps even more remarkable to add up the market cap associated with that 90% share. If we attribute the vast majority of Google and Facebook’s market value to their ad business and part of Amazon’s retail business (I’ll use a round number of $500B for Amazon, which is proportional to their advertising market share relative to Facebook and Google), that’s around $3.3T in market value derived from several hundred billion in annual advertising revenue at the three companies.

Family Safety App Serves Up Tracking Data
Family safety app Life360 has been tracking and selling the precise location data for its 33M users, including kids. In defense of the policy, the CEO said: “We see data as an important part of our business model that allows us to keep the core Life360 services free for the majority of our users, including features that have improved driver safety and saved numerous lives.” Apparently, the Life360 CEO learned everything he knows about business ethics from Madison Hotels’ Eric Gordon.

AirTags Enabling Car Theft
To add to your location-tracking phobia, Canadian police are warning that car thieves are surreptitiously affixing Apple AirTags to high-end vehicles to track them back to the owner’s garage and then steal them. I mentioned the proliferation of these ad-hoc wireless networks back in #296. With the current issues of compromised privacy, it’s not clear to me that the costs of tracking devices outweigh the benefits. Data hungry Life360 is acquiring Tile, one such ad-hoc tracking device network who assures users they do not sell location information.

Capturing Data Center Heat
Bitcoin mining company MintGreen runs their servers immersed in a non-conductive coolant that allows them to siphon off heat, reportedly recouping 96% of mining energy. Each server generates enough energy to heat 350,000 square feet of space. MintGreen will be partnering with Lonsdale Energy Corporation (LEC), a city-owned utility in North Vancouver that uses geothermal heat pumps and natural gas boilers to heat water that provides radiant heat to 7,000 apartments. By using MintGreen’s harvested thermal energy instead of natural gas boilers, LEC expects to save 20,000 tons of carbon over 12 years. Facebook is recovering 100,000 MWh of energy at a data center in Denmark and Amazon is partially heating their Seattle headquarters from a nearby data center. Examples seem to be few and far between, and it's not clear to me why we don't see more of this effort to recycle heat generated in data centers globally.

Look-Up AI
Google’s DeepMind has devised a way to use a look-up system for a large language model. The results are a staggering 25x improvement in efficiency over the current method. RETRO, as it’s called, uses only 7B parameters compared to OpenAI’s GPT-3 (with 175B) or Microsoft’s Megatron (with 530B). It’s this parameter reduction (by two orders of magnitude!) that affords the improved training/performance. To compensate for the smaller neural network, RETRO has a database with 2T passages of text it can access when it comes across something similar. It’s also easier to track how and what RETRO learns based on its database access history, making the look-up system less of a black box than other language models and potentially allowing researchers to monitor and edit out bias and offensive phrasing.

AlphaFold’s Omicron Analysis
In other DeepMind news, researchers have used the company’s AlphaFold tool to model in silico the antigen-antibody interactions between Omicron and four antibody variants obtained from infected patients. The study suggests that, despite notable mutations and somewhat reduced binding affinity, Omicron’s conformation is still similar enough to prior SARS-CoV-2 variants that it can’t completely avoid detection and neutralization by the immune system.

Miscellaneous Stuff
10-Year Carbon Crossover for Fake Trees
Apparently, fake Christmas trees are only more environmentally friendly if you use them for at least a decade, according to the Guardian. That’s when the 40 kg of greenhouse gas emissions crosses over with the 4 kg from a chop-em-down type tree. The study did not take into account the costs for when you hastily cut down a tree in your front yard and end up with a wild squirrel running around the house.

JWST Launch: T-Minus 10
The James Webb Space Telescope is finally ready for liftoff on December 22nd. I wrote about the project, which should allow us to glimpse the Universe in its youth a few hundred million years after the Big Bang, back in #309 (reprinted below), and this recent article in Quanta Magazine has a lot of great info on the project.
After 25 years and $10B, the 7-ton James Webb Space Telescope (JWST) is nearing its launch into heliocentric orbit 1.5M km beyond Earth, from where it will take the deep space exploration helm from the 30-year-old Hubble. Once deployed, “eighteen hexagons of gold-coated beryllium mirror will open out, like an enormous, night-blooming flower. The mirrors will form a reflecting surface as tall and as wide as a house, and they will capture light that has been travelling for more than thirteen billion years.” The JWST will orbit in line with the Earth/Moon so the mirror array can be isolated from radiation generated (or reflected by) the Sun, Earth, and Moon via a tennis-court-sized parasol. To keep the telescope out of the shadow of the Earth/Moon, it will also orbit the 2nd Lagrange point (L2) on the far side of Earth perpendicular to its heliocentric orbit (see embedded animation). Ball Aerospace made the mirrors and cryogenic actuators that can move each mirror and curvature to within 1/10,000 the width of a human hair. The more light we can collect, the deeper and longer we can stare back into time – potentially 13.5 billion years – toward the beginning of the universe as we know it, some 14 billion years ago. Unlike the Earth-orbiting Hubble, which could be repaired and upgraded (Ball Aerospace fabricated instruments to fix its fuzzy focus, among other improvements), the JWST will be beyond our physical reach, so let’s hope for a flawless deployment.

Adaptability Key to SNL’s Longevity
Saturday Night Live debuted when producer Lorne Michaels was 30. Now, 46 years later, the show is the longest running on television next to The Tonight Show, which Lorne now also produces. In a feature on Michaels in the Washington Post, they describe the backdrop for the first season: “Both Nixon and Saigon had fallen in the preceding 14 months. The United States was stagflating; New York teetered on bankruptcy. ‘Saturday Night,’ as the show was first called, was designed to give rebellious voice to young baby boomers and revolutionize comedy the way the Beatles had blown up pop music.” We humans seem to perpetually exist in a state of crisis (whether major or minor, real or imagined), and comedy is always there to shine a light on the absurd, to point out the obvious, and to carry us forward on a wave of laughter. The only way for something to survive in the public eye as long as SNL has is by adapting and evolving, and Seinfeld describes it best: SNL is “amoeba-like in its ability to shift and react and survive in the ‘toxic and self-correcting ecosystem of comedy’ — all due to one man’s vision and work ethic.”

Stuff about Geopolitics, Economics, and the Finance Industry
NZS in Complexity Investing Spotlight
Simon Evan-Cook writes in Wealth Manager about complexity science as the next investment frontier, featuring NZS Capital as one example of a firm tapping complex systems to think about investing:
“One good example, from a team who have written the paper on complexity investing, is the £29m Jupiter Global Equity Growth Unconstrained fund. Beneath this unassuming name is a process run from the US by Brinton Johns and Brad Slingerlend of NZS Capital.
NZS stands for non-zero-sum, which is integral to complexity science and NZS’s process.
It’s part of a concept known as ‘emergence’ in complexity circles. Emergent means greater than the sum of its parts. The idea doesn’t lend itself well to spreadsheets, as it often hinges on intangible factors, such as culture, trust or relationships. These things are crucial, but live between and beyond the data – so they are missed by many fund managers.
NZS looks for companies that benefit not just themselves, but their wider ecosystems too. Drug companies that abuse their monopoly-supplier status to gouge unfortunate customers are out, as are firms whose environmental costs outweigh their benefit to society. But sustainable companies that are instrumental in growing their customers’ and suppliers’ businesses, and so create as much value for them as for themselves, are in.
The firm believes that holding such high-quality companies, but selling them when they start to get become win-lose greedy, is key to antifragile returns.”

Innovation ETF
Congratulations to Harbor on the launch of their new Disruptive Innovation ETF. NZS Capital is pleased to partner with Harbor as one of the five managers contributing to the portfolio.

Amazon and Tesla: Strangers in a Strange Land
A lot of the innovation we witness in the world around us is largely software- and data-driven. It frequently involves just replacing an analog interface with a new digital one without actually transforming the underlying industry. A good example of that would be travel booking, an industry that transitioned from human travel agents to online bookings, without much changing under the hood (many of the exact same booking engines in the background have been running for several decades). It’s nice that it’s more convenient to book a hotel, but it’s not revolutionary. Two businesses, Amazon and Tesla, stand out in the ways they have not only innovated in software but transformed complex, analog industries.

Amazon grew their logistics business steadily from the start of the century until 2019, and then they doubled it over the last two years. It’s as if the world is not moving fast enough for Amazon, so they are willing their own future growth into being by buying (or building) warehouses, outfitting data centers, building their own overseas shipping containers, buying all the vans they can, hiring everyone they can, and raising wages across the US while they are at it. A widespread outage in the ubiquitous Amazon Web Services ground to a halt many business dealings (including Amazon’s own) in the US, and had such far reaching consequences that some people’s WiFi auto cat feeders stopped dispensing food for feline companions. The head of Amazon’s retail recently said that, by 2022, they would become the largest shipper in the US. In contrast, FedEx and UPS, companies that once claimed Amazon as a large customer, appear rudderless, offering customers the same service for more money as they listlessly circle the drain. Amazon is effectively setting the pace for many parts of the economy (not just in retail and logistics – e.g., Graviton3 is their new leading-edge data center chip) and no one else is keeping up with them, at any level of scale. It’s not entirely clear anyone is really trying that hard either. Maybe it’s institutional inertia – the paralysis Buffett warns against that keeps companies static and risk averse. As Amazon becomes more vertically integrated, hires more people, and buys more warehouses and land, they are effectively driving up the cost of doing business in their industries for every would-be competitor. The WSJ article on wages notes how ecommerce pet supplies company Chewy has struggled to keep workers in locations that are in geographic proximity to new/expanding Amazon warehouses. Chewy further noted on their earnings call last week that their shipping partner is raising costs on them in January, a move that will impact margins negatively in 2022 (this is the exact problem I’ve been trying to point out regarding FedEx and UPS: they are harming their customers to a point where those customers could lose out to Amazon, leaving the legacy shippers eventually with no customers, which is short termism and negative-sum behavior).

Like Amazon, Tesla has dramatically pulled away from their competition. They continue to generate reams of data to feed their AI as they improve autonomous functionality and rapidly expand battery manufacturing capacity, leaving the rest of the auto industry looking rather embarrassed. And, Musk’s SpaceX business is setting all sorts of new benchmarks for rockets and space-based Internet, leaving traditional aerospace contracts stuck on Earth. It seems like legacy auto makers have held as many press events as SpaceX has launched satellites, always proclaiming they will dethrone Tesla but without a shred of tangible evidence to back up their bravado.

It’s as if the world is frozen, save Amazon and Tesla. What’s surprising about their blistering growth is how atypical it is for non-software companies to achieve. Amazon and Tesla leverage plenty of software, but, by and large, they deal with messy, complicated, real world things – objects, people, commodities, extremely complex manufacturing/physics, etc. We expect this type of accelerating progress more from a purely software business, such as Google’s search engine. To achieve this type of growth in sectors that require the interface of the analog and digital worlds is remarkable.

I talk a lot about the transition from analog to digital – from the Industrial, to the Information, to the AI age. Amazon and Tesla are at the leading edge of this transition, and we hypothesize that we will see digital leaders emerge in other sectors/industries. However, the rollover isn’t inevitable by any means, especially for non-data-centric businesses. And even some data-centric sectors – like finance and healthcare – remain mired in the analog world. As we wrote in Pace Layers, regulation can be an impediment to change. Indeed, regulators have been long-time party crashers for finance and healthcare, yet have been late to the party for media, retail, and autonomous driving, leaving Amazon and Tesla to grow unfettered. Maybe that is part of the explanation for their success. We are seeing some fairly large digital finance businesses built, but it’s a wide open question as to whether they will survive the regulatory gambit.

If the underlying process (not just the interface) of a business is rooted in the Industrial Age, then the transition is more complex – as Zillow can attest. Their recent strategy shift was due (at least in large part) to the messy nature of trying to herd contractors into fixing up a house while simultaneously dealing with a complex financial market, interest rate risks, macro factors, and more. Energy is another messy, analog industry that’s struggling with massive logistical hurdles in storage, grid improvements, etc. in its bid to go digital. And yet, Amazon and Tesla have successfully dealt with some equally impressive analog hurdles, so what sets them apart? Vertical integration? Lack of regulation? Was the role of luck, timing, or some other unknown factor much more important than we might think? Does it come down to the personality of the lead entrepreneur – does breaking out of analog shackles require a higher risk-seeking propensity as well as creative genius? Will we wait decades for the next digital Einstein to emerge in finance, healthcare, or energy? Are Amazon and Tesla anomalies, destined to take over (and possibly evacuate) the planet in Buy n Large fashion? Or, is it just a matter of time before we see more transformational companies and leaders pulling the analog world into the digital future?

✌️-Brad

Disclaimers:

The content of this newsletter is my personal opinion as of the date published and is subject to change without notice and may not reflect the opinion of NZS Capital, LLC.  This newsletter is an informal gathering of topics I’ve recently read and thought about. I will sometimes state things in the newsletter that contradict my own views in order to provoke debate. Often I try to make jokes, and they aren’t very funny – sorry. 

I may include links to third-party websites as a convenience, and the inclusion of such links does not imply any endorsement, approval, investigation, verification or monitoring by NZS Capital, LLC. If you choose to visit the linked sites, you do so at your own risk, and you will be subject to such sites' terms of use and privacy policies, over which NZS Capital, LLC has no control. In no event will NZS Capital, LLC be responsible for any information or content within the linked sites or your use of the linked sites.

Nothing in this newsletter should be construed as investment advice. The information contained herein is only as current as of the date indicated and may be superseded by subsequent market events or for other reasons. There is no guarantee that the information supplied is accurate, complete, or timely. Past performance is not a guarantee of future results. 

Investing involves risk, including the possible loss of principal and fluctuation of value. Nothing contained in this newsletter is an offer to sell or solicit any investment services or securities. Initial Public Offerings (IPOs) are highly speculative investments and may be subject to lower liquidity and greater volatility. Special risks associated with IPOs include limited operating history, unseasoned trading, high turnover and non-repeatable performance.