NZS Capital, LLC

SITALWeek

Stuff I Thought About Last Week Newsletter

SITALWeek #292

Welcome to Stuff I Thought About Last Week, a collection of topics on tech, innovation, science, the digital economic transition, the finance industry, Sparklemuffins, and whatever else made me think last week. Please grab me on Twitter with any thoughts or feedback.

Click HERE to SIGN UP for SITALWeek’s Sunday EMAIL

In today’s post: Fitness takes off in VR; labor force shifts may be the unlikely cause of post-pandemic economic changes for low paying jobs; indoor air quality challenges; Nvidia’s Grace; creating more than you consume; Sparklemuffins and T. rexes; too much cash in one place and not enough in another; difficult NFTs; and lots more below...

Stuff about Innovation and Technology
DillBot
Dill is a robot from Pickle Robots that can unload around 1600 packages per hour of different shapes, sizes, and weights from the back of a truck. Designed to work alongside a human, this co-bot is an interesting choice for fulfillment centers that intake a lot of packages. The company projects that one person could oversee five Dills at a time, stepping in when necessary to help with any awkward loads. Dill is about twice as fast as the Boston Dynamics Stretch bot, which is designed to work without any human oversight. Pickle’s strategy is to solve for most, but not all, use cases as a way to speed up co-bot adoption. Although Dill is much faster, more accurate, and less prone to injury compared to a human, it runs around the same cost as one humanoid ($50-100k for both upfront and annual operation costs).

Dance-Fighting VR Fitness
Exercise is the killer app for virtual reality, according to this Verge interview with the maker of the “dance fighting” game Supernatural. An unexpected benefit of exercising in VR is that things hurt less because your perception of your physical body becomes somewhat disconnected, which allows you to potentially work harder without noticing it. “We’re building a lot of different sort of music-driven, cohabitated social experiences...working with Pharrell, and OKGo, and Justice, and building these really cool, immersive, music-driven experiences where you’re moving your body and you’re inhabiting crazy, magical experiences.” I am reminded of the omnidirectional treadmill from Ready Player One.

Labor Logic Inversion
As I was reading about the extreme labor shortages my local restaurants are grappling with, this line really stood out: Is the classic restaurant model fundamentally doomed and/or outdated—in a way the pandemic made clear for those providing its life force? In that sentence, “those” refers to wait and kitchen staff. While investors fret about labor inflation, I am wondering if something different is happening, a sort of inversion of logic: you can’t have labor inflation if there is no labor. If people have moved on to other types of jobs or exited the labor force, it could cause a shift in the underlying economy more profound than the pandemic-induced consumer behavioral shifts. Perhaps the main issue isn’t that people no longer want to eat out at casual dining restaurants, it’s that casual dining restaurants have lost their labor engine. Even Uber and Lyft are struggling to find drivers as the economy reopens. It's as if there is a nationwide strike or walkout against low paying jobs that lack benefits and safety nets. Most of us have been thinking about the post-pandemic world in terms of consumer behavioral shift, but what if a labor behavioral shift is the more important angle? Logic inversions are rare beasts, and when you come across one, you want to examine it closely even if it ultimately proves wrong. The economic downturn disproportionately slammed people under 30 and those making under $30,000 per year (as I read recently in a client-only report from Empirical Research). Pandemic closures also pulled forward a lot of retirements, according to Empirical, with retired folks climbing from 19% to 20% of the US population.

The entire situation seems to again support the idea of a hollowing out of the middle of the restaurant industry, leaving a long tail of niche/high-end food providers and a large head of national chains that can invest in technology (like Chick-fil-A’s delivery robots or Domino’s new driverless delivery pods from Nuro). What’s left for the large middle segment of the industry that relies on seated dining and drinking? Rather than consumer behavior causing a shift toward delivery, will all of these factors instead force a shift to delivery for casual dining? The labor cost to draw people back into the kitchen and dining area might be even higher than the cost to deliver a meal, especially for a more efficient ghost kitchen with cheaper overhead costs. If this continues to play out, will the currently unsustainable and zero- to negative-sum value proposition of delivery morph into something more viable? The problem with delivery, which was recently very well explained by Jason Hirschhorn here, is that it appears to be an unnecessary abstraction of the relationship between the consumer and the business (similar to the credit card situation that I wrote about last week, which Square is trying to fix). There is also a big push to make the temporary alcohol-to-go rules more permanent, which would boost delivery margins as well. A Deloitte survey – which probably merits a degree of skepticism since who knows how we’ll all behave when the world reopens – suggests eating at home is a behavior pattern people are keen to stick with: only 7% of people surveyed said they would cook less at home post-pandemic. I believe most extended unemployment benefits taper off this September, and I think we should hold all predictions until we are a few months past that point in time. Recent government data on unemployment is encouraging, suggesting that there may not be a tidal wave rush back into sub-optimal services jobs.

While we’re on the topic of unviable businesses of the past, let’s look at brick-and-mortar retail trends. McDonalds is shutting down hundreds of locations inside of Walmarts. That seems pretty pessimistic to me – do we really think people won’t go back to stores in any meaningful way? That would seem to spell serious trouble for just about every retailer that isn’t either a long-tail niche provider or Amazon. As I wrote in #287, even Walmart appears “stuck in the middle” and saddled with the wrong business model. Without impulse buying and a good way to recapture lost trade promotion dollars – as foot traffic shifts to click and collect or delivery – traditional retail is upside down on its business model, physical space, and fulfillment capabilities. To see the challenge retailers are facing, look no further than the convoluted ways in which Kroger is creating duplicate robotic warehouses and Target is putting items on shelves, only to take them back off to send to a different sortation center for direct consumer shipping.

Indoor Air Quality Conundrum
The WaPo has a story on the Sierra Mar restaurant at Big Sur’s Post Ranch Inn and its new air purification system implemented to mitigate viral spread and improve air quality during California’s smoky wildfire season. While bad indoor air is certainly at the root of a host of medical problems/ailments (e.g., poor cognition – see #227), Sierra Mar’s elaborate system of air purifiers (including mini tabletop units), vents, and sensors seems perhaps intrusive and a little bit of a damper on the otherwise magnificent ambiance of the location. The proven, more elegant solution for protecting indoor air quality is the vertical laminar flow we’re used to seeing in semiconductor plants or other clean rooms: a uniform, steady flow of air venting from the entire ceiling sucked down through a grated floor. Granted, this setup would be a challenging retrofit for most buildings. Even just adding heavier filters and/or an air exchanger to constantly recycle air and/or seasonally heat/cool introduced outdoor air is going to significantly increase the HVAC energy burden. Just opening the window helps, but it’s the movement of air down and out that I suspect makes the greatest difference. Wholesale retrofit of our current infrastructure without significantly increasing energy usage seems a daunting task.

Masking Transaction Data
Encrypted messaging app Signal discussed why they are starting to trial payment platforms inside of Signal in a blog post last week:
“As the world stands today, the future of transaction privacy does not look great. The existing landscape is dominated by traditional credit companies, who over the past decade have been steadily pushing their networks for increased access to user data. They (and their data customers) are on a track to getting SKU level data of every purchase everyone makes everywhere. There are other contenders, such as regional online payments networks (like Venmo in the US), but the data story there is similar.
This is not a future we are particularly excited about. At Signal, we want to help build a different kind of tech – where software is built for you rather than for your data – so these are trends that we watch warily.”

Further, Signal sees much of the crypto world focusing on speculation over user interface and usability (I agree! See the last paragraph below on NFTs). For now, the company is open to connecting to wallets that meet its requirements around privacy, starting with one called MobileCoin.

Semi Fab Semi Shortage; Nvidia’s System-Level Advances
The chip industry, already struggling to keep up with demand, now faces a circular reference problem as the equipment needed to make more chips is itself facing a chip shortage. However, that’s not stopping Nvidia from pushing the boundaries of compute, as Jon reports: at their GTC conference earlier this week, Nvidia added an important piece to their ambitions in solving the world's toughest computing problems with the announcement of Grace, their first CPU for the data center. Grace will be based on the Arm instruction set, and Nvidia's goal (at least for now) isn’t to compete head-to-head with Intel for the vast majority of workloads, but to optimize at a system level for the next decade of large AI models and high performance computing (HPC) problems. In classic Nvidia fashion, the Grace CPU is much less about the CPU itself than its contribution to the overall system performance – Grace appears to be a relatively off-the-shelf ARM-based CPU, but it will allow Nvidia to leverage its NVlink interconnect to create a very high speed pipe between the CPUs and GPUs to access system memory, removing what was previously the bottleneck in training large AI models. As Karl Freund from Cambrian-AI Research put it: “Since no CPU vendors are building such an interconnect to enable this level of outrageous performance, NVIDIA had no choice but to design it themselves”. Jensen's ambition with Grace is to enable AI models with trillions of parameters, which seems a ways off given the reigning champ for large AI models is GPT-3 at 175B parameters, but we also know these models are doubling in size every 3.4 months. The broader takeaway is that the pace of innovation at the system level is in some ways starting to outpace innovation on the silicon itself, with Nvidia now offering the CPU, GPU, and DPU, and, equally importantly, their own interconnect technology in NVlink. Grace is scheduled to be deployed alongside Nvidia GPUs in the Swiss National Computing Centre's “Alps” supercomputer in 2023; Alps will be used for heavy simulations in weather, climate, materials science, astrophysics, etc. and will offer 20 exaflops of performance, a 10x improvement over the US DOE's “El Capitan” system that is also scheduled to come online in 2023.

Amazon’s Welcome Refocus
Following a string of disappointing annual letters from Jeff Bezos that seemed to put politics ahead of innovation (see #241), I thoroughly enjoyed this year’s letter. Sure, it had a lot of not-so-subtle bragging meant to put Amazon on a pedestal, but it also had a lot of great insight. Create more than you consume” is so important to us at NZS Capital that we named our firm after it. It’s the win-win, non-zero-sum outcomes that we seek in companies we invest in and it’s how we run our own business.

Miscellaneous Stuff
Flamboyant Arachnids
Sparklemuffin and Skeletorus are two types of Australian dancing peacock spiders. This type of adorable jumping spider, of which there are 92 known species, is vibrantly costumed and employs elaborate mating dances. National Geographic has the story and the video for anyone not arachnophobic.

Quantifying Fossil Finding Serendipity
A mere one out of every 80 million T. rexes that lived to be adults has been found as a fossil. This calculation is reported in a new Science paper that predicts around 20,000 of the adult giant dinosaurs were alive at any given time during their 2.5-million-year reign of terror. With an average lifespan of 19 years, that gives 127,000 generations for a cumulative total of 2.5 billion T. rexes in the history of Earth. Clearly, becoming – and being found as – a fossil is an extraordinarily rare event. There are lots of fun facts in the summary link – for example, their bite strength apparently increased by an order of magnitude as they entered adulthood (at around 15.5 years old), suggesting that the juveniles and adults existed as different populations with different prey species, which could explain why there is no fossil record of other medium-sized predators coexisting with T. rex, as the juveniles would have filled that niche role.

Time’s Arrow from Orthogonal Neural Coding?
As the brain takes in sensory input, it has to find a way to quickly distinguish new information from memories of the very recent past without getting the states confused. Scientists recently discovered the brain does so by rotating the neural signal to be encoded as a memory, thus preserving the ability for both states to utilize the same set of neurons without overwriting or muddling information. Physicist Sean Carroll suggests the feeling that time is flowing in one direction might come from this mechanism of constantly comparing past to present. This leaves me with a couple of questions: given thoughts and emotions are essentially a sixth and seventh sense, does the brain also flip these as your mind goes from one random idea or feeling to another? Regarding Lisa Feldman Barrett’s work on the brain as a prediction machine – where in this mechanism does the prediction lie, if it’s not encoded in the memory or the new information? I wrote more about Feldman Barrett’s seminal work in the final section of #272.

Quantum Computing Radiation Wrench
If the hypothetical idea of quantum computing turns out to be a real thing in some other universe (I suspect not this one), one of the big problems will be cosmic radiation. Scientists have been calculating just how much error correction would be needed to deal with cosmic rays hitting an array of qubits several times per minute. “An impinging particle ionizes the substrate, radiating high energy phonons that induce a burst of quasiparticles, destroying qubit coherence throughout the device.” The rays would cause a burst of errors that would need to be dealt with in order to get accurate results from the theoretical devices.

Stuff about Geopolitics, Economics, and the Finance Industry
From Way Too Much to Way Too Little Cash
VC funding reached a record smashing
$64B in Q1 2021. That tally is 43% of all the money raised throughout 2020. SPACs also raised $99B in Q1. There’s been a whole lotta money burning holes in select people’s pockets lately.

In contrast, outside of venture-backed startups, people are turning their empty pockets inside out to look for change. Pew reports that over 150M people fell out of the middle class in 2020, the first time the global number of middle-income individuals has shrunk since the 1990s. Most of this loss happened in emerging markets, with the middle class cohort’s total number somewhat buoyed by around 100M people dropping from upper-middle/high into middle class. The number of poor people, the lowest earners, rose from 691M in 2019 to 803M in 2020. If I circle back to the section above on restaurant labor, it’s quite possible the free lunch of underpaying the low-skilled service industry (not to mention the lack of benefits and savings programs for those jobs) might be forced to come to an end in the US with wages that need to be many multiples of where they are to reach stability. Or, maybe Dill will just take all the jobs.

NFTs Are No Picnic
Brinton and I each bought some Bitcoin back in 2014, not because we had some unique view, but because we were curious – it seemed interesting, but we didn’t understand it. What better way to learn than first-hand experience! In a similar spirit, I bought my first NFT this past week just to get a feel for what it's like to own something digital that anyone else can see. Buying Bitcoin even back in 2014 was quite easy. Buying an NFT now is anything but. Between the user interface of OpenSea (a leading NFT platform with backing from a16z), which makes a dumpster fire look like boat drinks, the kludgy interplay with Fortmatic (don’t even ask me what their role is), and connecting to a separate wallet, the entire process is for the brave at heart. It felt a bit like using eBay in 1996 when I had to go to the post office to get a money order to mail to the seller before they would ship my item. I assume it will get easier, and I look forward to continuing to experiment and learning more. I wrote more about NFTs in SITALWeek #284 and #287.

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 simply an informal gathering of topics I’ve recently read and thought about. It generally covers topics related to the digitization of the global economy, technology and innovation, macro and geopolitics, as well as scientific progress, especially in the fields of cosmology and the brain. I will frequently state things in the newsletter that contradict my own views in order to be provocative. 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.

jason slingerlend