SITALWeek #210
Stuff I thought about last week 9-15-19
Welcome to Stuff I Thought About Last Week, a collection of topics on tech, innovation, science, the digital economic transition, the finance industry, giant flying reptiles, and whatever else made me think last week. Please grab me on Twitter with any thoughts or feedback.
In today’s post: an essay on the rapidly evolving food delivery, restaurant and grocery industries; legislating rideshare labor could paradoxically cement the incumbents position due to regulatory capture; influencing social change with “terms of service”; the complexity of commerce with the Santa Fe Institute; understanding quantum mechanics; the latest neuroscience on the lack of free will; and, lots more.
Click HERE to SIGN UP for SITALWeek’s Sunday EMAIL (please note some ad blocking software may disrupt the signup form; if you have any issues or questions please email sitalweek@nzscapital.com)
Stuff about Innovation and Technology
Congrats to NZS Capital co-founder, and now legendary runner, Brinton Johns for his 27th-place finish in the 100-mile race through the mountains of Colorado!
I had the pleasure of attending the Santa Fe Institute’s working discussion on The Complexity of Commerce this week hosted at Shopify. To give some sense of the actual complexity of commerce, this event was originally slated to be hosted by a retailer that subsequently declared bankruptcy ahead of the event. As the WSJ reported, there were more store closures in the first half of 2019 than all of 2018. The event covered a wide range of topics with presentations from some of my complexity science idols: Brian Arthur, Bill Gurley and Michael Mauboussin. This link for the event has some suggested reading (some of which I’ve highlighted in the past). Here are a few things I found interesting:
Increasing-returns guru Brian Arthur explained that he expects network effects will continue as they have for decades in tech – creating/sustaining natural monopolies, which should continue with AI. Proprietary data sets and who has access to them are increasingly the most valuable things in tech platforms.
Arthur reminds us that regulation always comes to the Wild West, but that is can take decades. The “Dickensian” period of the Industrial Age in Britain took 30 years to be addressed and even longer to fully get under regulatory control (100 years?).
Arthur suggests we are in the “distributive age” where it’s not about GDP (which is a weak measure of progress) and jobs, but maximizing access to the economy for the most people, i.e., intelligently distributing the wealth of the economy.
Stitch Fix CTO Cathy Polinsky and SurveyMonkey COO Tom Hale both discussed the importance of humans and AI working together to improve algorithmic outcomes. You can read more about Stitch Fix’s AI efforts in this article.
John Hagel from Deloitte suggested we could see infrastructure layers of the economy continue to concentrate, while products and services could continue to fragment into more and more niches. I’d take this one step further to say the massive infrastructure layers like AWS seem to be enabling the fragmentation itself.
There was a fair amount of discussion about the increasing role of search as a gatekeeper in commerce. I think this applies not just to search engines, but e-commerce platforms as well, which increasingly serve ads to buyers. At some level, if the infrastructure layer is consolidating more, then it will be possible to put a higher and higher tax on access to it (see article below on clothing brand Everlane's impossible cost of acquiring new customers).
There was some good discussion on regulatory capture cementing current platforms, but also potentially paralyzing or slowing them down (See more on regulatory capture below in the gig worker paragraph).
At the event, I chatted with Alex Komoroske about his views on complex systems, which you can find in this essay: algorithms that have emerged from technological progress are increasingly optimizing for the wrong outcomes because they maximize for attributes that mattered during the scarce resource period that drove human evolution, but can be dangerous in times of abundance.
Nearly 150 CEOs sent a letter requesting action from government on gun control. In SITALWeek #195, I discussed Salesforce.com’s decision to stop selling their e-commerce platform software to gun sellers in the US:
Salesforce is telling its retail customers to stop selling automatic and semi-automatic rifles or they will be blocked from using their software. Cynically, I could argue this is motivated by limiting their own liability (e.g., if Salesforce’s Demandware e-commerce software was used to sell a weapon subsequently used in a crime). However, maybe it’s more than that, or maybe it could be more. With simple “fair use” policy changes, the handful of cloud software platforms underpinning the entire economy could globally influence social change. Marc Benioff, Satya Nadella and a few other tech leaders could set policy and dialog for the world based on their end user agreements. Once you think this through, it's not a question of "how could they justify this?", but instead “how could they not?”
Given the number of tech companies that signed the letter, I would provocatively ask: if a tech company has a specific view or goals concerning social outcomes (whatever those may be), would it be more effective to drive change by changing your terms of service to explicitly promote that outcome? This should be considered legal behavior under refusal to deal rules. For example, if you want to end ocean pollution (I’m looking at you Benioff!), then incentivize your customers to eliminate single-use plastics from their products. Or, perhaps a little more extreme: if your company wants action on guns, guns are made of steel, so don’t sell software or hardware to the steel industry.
If you missed our recent appearance on The Stock Podcast, you can listen here (or iTunes link) or read a PDF the transcript here. I want to briefly expand on one topic Brinton covered in the interview: he mentioned that there are a handful of companies in the semiconductor supply chain that, if they were gone tomorrow, would send the world back in time and grind the economy to a halt. One company mentioned in that vein was ASM Lithography, which enables every single advanced processor, 5G, and AI chip made today. The world now runs entirely on these chips. Without them we’d go back in time 10-20 years. More broadly, this test is very useful for estimating the value of a company: if a company disappeared, what would happen? To give a provocative counter example to ASM Lithography, what if Facebook were gone tomorrow? Besides the obvious loss of $500B invested in its shares, it’s not clear the world would be significantly different by the end of the week. This is of course true of many consumer-habit companies like Facebook or Coca-Cola, two companies that exist by exploiting evolutionary mechanisms for gossip and calories. This doesn't mean Coke and Facebook won't be bigger 20 years from now, it's just an interesting exercise on the utility of a business. (Eventually, if the negative health consequences are big enough, habits can change with education.)
Online retailer Everlane said that the increasing cost of paying online ad platforms to acquire new users has made it so no online business can be profitable. Is the increasing toll taken by search, social and other online ad networks stifling innovation? Government regulation appears to be catching up to this risk.
As food consumption habits radically transform in the coming decades, the biggest loser might be grocery stores. As food-delivery platforms platforms vertically integrate with cloud kitchens and white-label brands, offering subscription packages with routed and scheduled meal deliveries, it’s possible delivery might become cost competitive with home cooking for some households. Grocery stores operate on razor-thin margins that are already under attack from e-commerce. The dramatic transformation of the trillion dollar US food industry would also change the type and mix of restaurants that survives and could even re-write the food supply chain. Checkout my long read on the death of food consumption as we know it.
Speaking of food, here is a great example of the creative destruction happening in the restaurant space: diet brand Whole30 is launching a delivery-only restaurant in Chicago in combination with Grubhub to cater to customers looking to improve their health. It’s not a stretch to think such a business could launch a subscription service with scheduled and routed meal deliveries, perhaps even subsidized by health insurance or employers.
Walmart looks to expand their 26% share of grocery sales in the US by expanding their unlimited delivery service to 50% of the country by year’s end. The service costs $98, which is vastly cheaper than the time/cost dedicated to driving to the store and pushing a cart around.
For gig economy workers, is job security more important than flexibility? As expected, California passed legislation this week that would require companies like Uber and Lyft to reclassify contract workers as full-time employees. Gig companies have plans to fight the legislation on multiple fronts, from saying that drivers aren’t core to the business to advancing ballot measures to reverse the legislation. It seems to me that more and more people are choosing to enter the gig economy (36% of US workers according this Gallup report), and there is a competitive market amongst dozens of gig economy companies for these workers. The legislation is based on the assumption that gig workers are currently underpaid and being taken advantage of. There is room for improvement of benefits and pay for gig workers, but I am not sure this legislation would achieve a net improvement. Uber said recently that this type of legislation would force drivers to work full time for only one rideshare company, which (contrary to the legislators’ apparent intent), would be a boon for Uber and Lyft. Think about how hard it would be for new entrants in rideshare/delivery to hire a large network of drivers if they are all locked into one existing platform? It could also consolidate market share as drivers choose to work for the largest platform. This lack of competition would be bad for both consumers and workers. Drivers would have fewer on-demand companies competing for their labor. Perversely, legislation could make everyone worse off except the largest gig companies themselves due to regulatory capture; we touched on this concept in our two tech regulation papers, and, as other smart folks like Bill Gurley have recently pointed out:
“Perhaps the key reason Silicon Valley should be wary of U.S. regulation of big tech is that it will make it that much harder for the new startup to disrupt said incumbent.”
Lyft’s co-founder John Zimmer discussed the issue at the Deutsche Bank Tech conference this week:
“...independent contractors, the way labor law is written, if you want to provide benefits to a small portion of your drivers, you then turn all your drivers into employees, which isn't the best thing for society because if you flip all the way to employees, you only get a certain type of worker, and in our case, 91% of our drivers drive less than 20 hours, 76% of our drivers drive less than 10 hours. And so going full-fledged employee mode, while there would be some pros and cons for us, the con being additional cost for a certain subset of workers, the pro being you'd have a different amounts of control where you could ask drivers to work specific shifts at specific times. So in some ways, it will be easier to manage the marketplace. But then you would hurt the majority of the drivers that are doing this on a more supplemental income basis and don't want to work shifts.”
I’d like to see all gig economy companies focus on building long-term, sustainable businesses with prices high enough to support fair wages and benefits while continuing to provide more value to all of their constituents. Regulation, as it’s currently looking, could make everyone worse off.
NVIDIA’s AI chips are often criticized by competitors for using too much power compared to custom alternatives. The company’s chief scientist countered in this interview that NVIDIA’s Tensor Core for deep learning is just as efficient as custom chips, and the 10-20% difference in GPU power overhead is offset by not having to fetch data back and forth. NVIDIA also isn’t yet seeing the need to jump on the chiplets bandwagon (the new trend in chip design, which allow you to go vertical and stack different chips in a single die). It seems that NVIDIA, with the millions of CUDA programmers, remains poised to own the network effect for AI. Related: chip giant TSMC is pushing hard to keep Moore’s Law on track for decades to come with 2.5 and 3D structures.
The long tail of streaming video, in aggregate, is now consuming more bandwidth than Netflix at 12.8% of bits vs. 12.6%. As I’ve said in the past, the media industry is in a “have your cake and eat it too” period where the rising value proposition of content and direct-to-consumer subscriptions can more than offset cord cutting. And, the market isn’t 100M+ US households, it’s billions of smartphone users globally. I discussed some of these points in my recent essay on ViacomCBS.
Further highlighting the lack of non zero sum in the music streaming business, Spotify, Google, Amazon, Sirius, and others protested a move to pay songwriters and composers more money. Apple, the sole proponent of the move, had this to say:
“Apple doesn’t think it makes sense for [songwriters] to be dependent on the business success of the services that use their music. They should get a consistent, predictable, and transparent per-play rate, and then it is up to the services to run the most innovative and efficient businesses possible for which they can recoup the upside.”
Companies like Spotify are effectively trying to arbitrage between what music is worth, i.e., what singers and songwriters should receive for their art, and what people are willing to pay for it. Right now, that proposition seems to be upside down: people aren’t willing to pay enough to support the art and provide margin for the distribution. Here is Spotify’s argument, which I will leave without comment:
“With lower royalty rates, we can develop a better product, and if we can develop a better product, more people will be willing to pay, the market will expand, and the total value to the entire ecosystem will grow.”
Thanks to tech advances and declining costs, every city in China could be solar powered at grid parity today without subsidies.
Apple’s new lineup of 4G iPhones announced this week could cause the company to lose share in China.
Miscellaneous Stuff
Even physicists don’t totally understand quantum mechanics, but if you want to learn why that is the case (and get a shot at understanding it yourself), checkout Sean Carroll’s latest book Something Deeply Hidden. Sean’s last book The Big Picture is one of my top-recommended books to people that ask me for suggestions. I am about one third of the way through Something Deeply Hidden, and if you want to try to understand the implications of Quantum Mechanics, it’s the most approachable book I’ve come across. Spoiler: Sean is a big proponent of the “many worlds” interpretation of quantum mechanics (and so am I, or should I say so are we...me and my infinite copies...well at least I hope I mostly agree with myselves!). Sean, a professor at Caltech, an external staff member at the Santa Fe Institute, and a physics advisor on Marvel movies, explains in this Wired interview: “We see tables and chairs and people and planets moving through spacetime. Quantum mechanics says that there are no such things as tables and chairs—there’s just something we call a wave function.”
Citing a seven-year-old study and unpublished work, a misleading Atlantic article this week muddied the waters regarding the point at which we are consciously aware of what the brain wants us to do/think/say. The article brings to mind this old joke: Do you believe in free will? Of course, I have no choice. The idea that we could be fully aware of the activity in our brain before we are consciously aware doesn’t pass any kind of logical or real-world feeling test. Importantly, the Atlantic article only cited studies concerning brain signals related to motor movement; a more recent study showed “the outcome of a free decision to either add or subtract numbers can already be decoded from neural activity in medial prefrontal and parietal cortex 4 s before the participant reports they are consciously making their choice.” In Sapolsky’s Behave, he goes through some of the things beyond our control that nonetheless unconsciously impact our behavior, including: “blood glucose levels; the socioeconomic status of your family of birth; a concussive head injury; sleep quality and quantity; prenatal environment; stress and gluticocorticoid levels; whether you’re in pain; if you have Parkinson’s disease and which medication you’ve been prescribed; perinatal hypoxia; your Dopamine D4 receptor gene variant; if you have had a stroke in your frontal cortex; if you suffered childhood abuse; how much cognitive load you’ve borne in the last few minutes; your MAO-A gene variant; if you’re infected with a particular parasite; if you have the gene for Huntington’s disease; lead levels in your tap water when you were a kid; if you live in an individualist or collectivist culture; if your a heterosexual male and there’s an attractive woman around; if you’ve been smelling the sweat of someone who is frightened. On and on. Of all the stances of mitigated free will, the one that assigns aptitude to biology and effort to free will, or impulse to biology and resisting to free will, is the most pernicious and destructive.” (p. 597-598).
75 million years ago, North America had 500-pound flying reptiles with 30-foot wingspans. Discovery and analysis of a partial skeleton of a new species, dubbed Cryodrakon boreas or “frozen dragon of the north”, sheds new light on the evolution and behavior of the pterosaurs (which are “among the most popular and charismatic of all fossil animals”).
As we learned in the Hitchhiker’s Guide to the Galaxy, the answer to the ultimate question of life is 42. Mathematicians have finally solved the sum-of-three-cubes problem for the number 42 by harnessing the capacity of 500,000 idle computers around the world. The answer for x^3+y^3+z^3=42 is x = -80538738812075974 y = 80435758145817515 z = 12602123297335631
Stuff about Geopolitics, Economics, and the Finance Industry
Morningstar reports that passive funds are officially ahead of active in the US at $4.27T vs. $4.25T (note: earlier, they incorrectly reported that this crossover happened back in May).
Chinese influence over US companies continues largely unchecked. JP Morgan, whose CEO Jamie Dimon is one of the posterchilds for The Business Roundtable effort to go beyond shareholder value, has instructed all staff to not recognize Taiwan’s sovereignty and democracy as the bank looks to do more business in China. I guess lying is one way to go beyond shareholder value, just in the wrong direction. Also this week, Secretary of State Pompeo spoke to the Motion Picture Association of America, urging them to resist the rapid rise of Chinese censorship and creative control of the US movie making process.
“In exchange, Pompeo asked MPAA executives to stop bowing to Chinese censors. China is eroding Hollywood's freedom by setting the terms of the content of movies, as studios compete to have their movies accepted among the limited number of movies China authorizes to be shown in the country each year.”
(Richard Gere would probably agree, as he believes his support of Tibet has blacklisted him in Hollywood; Apple also recently canceled a Gere series for Apple TV+ despite winning a heated bidding war for the project.)
The FT reports on “Fear and oppression in Xinjiang: China’s war on Uighur culture”.
The Pentagon worries about US reliance on China for prescription drugs. "Basically we've outsourced our entire [drug industry] to China," retired Brig. Gen. John Adams told NBC News. "That is a strategic vulnerability."
"We can't make penicillin anymore," said Gibson. "The last penicillin plant in the United States closed in 2004."
Disclaimers:
The content of this newsletter is my personal opinion as of the date published and are subject to change without notice and may not reflect the opinion of NZS Capital, LLC (“NZS”). 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. I 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 (“NZS”). 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 has no control. In no event will NZS 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.