SITALWeek #341
Welcome to Stuff I Thought About Last Week, a personal collection of topics on tech, innovation, science, the digital economic transition, the finance industry, transformers, and whatever else made me think last week.
Click HERE to SIGN UP for SITALWeek’s Sunday Email.
In today’s post: transformer neural nets are supercharging the arms race in AI; drones are soon to be commonplace; using far-UVC light broadly for killing airborne viruses; private market correction; the rise of unions in the context of failed immigration policies; the tech sector enabled Russia to wage war in Ukraine, and it continues to enable China to enable Russia, but much more can be done to prevent future wars and oppression; and, much more below...
Stuff about Innovation and Technology
A Transformer Walks Into a Bar...
Google’s Pathway Language Model (PaLM) scales to 540B parameters. The model was trained on 6,144 of Google’s custom TPU v4 AI chips, far exceeding prior pods of 2,240 Nvidia A100s and 4,096 v3 TPUs. PaLM is reported by Google to have reached breakthroughs in understanding language, reasoning, and coding. While PaLM barely edges out the 530B parameters of Microsoft’s Megatron model, PaLM “can distinguish cause and effect, understand conceptual combinations in appropriate contexts, and even guess the movie from an emoji...generate explicit explanations for scenarios that require a complex combination of multi-step logical inference, world knowledge, and deep language understanding. For example, it can provide high quality explanations for novel jokes not found on the web.” PaLM, Megatron, and GPT-3 walk into a bar in the metaverse. The bartender, Watson, says: hey, is this some sort of joke? PaLM is the only one that laughs. There is a massive arms race among tech giants for human-like companion bots. Today’s search engine will evolve into the next contextually aware, seemingly sentient AI assistants. The pace at which progress is being made is quite impressive, and it could mean we are closer to this realization than we think. And yet, early applications are still conspicuously lacking, as the WSJ’s Joanna Stern found out when she lived with Amazon’s Astro robot for a week. One of the underpinnings of most of the large language models is a type of neural network called a transformer. Transformer models can look across a block of text and process the sentences (and words within the sentences) in any order. That allows a model to infer context that might be coming later in a sample. Nvidia is building transformer models into its latest chips, and this type of neural net is being increasingly applied to other types of AI problems, according to IEEE. Transformer models are growing 275-fold every two years and fast approaching trillions of parameters.
Fortnite for Ukraine
Players of Fortnite raised an astonishing $144M in aid for Ukraine as the game maker donated all proceeds from the first two weeks of the most recent season of its game to various causes.
Texas-Sized Drone Delivery
Google’s drone delivery company, Wing, began deliveries to tens of thousands of suburban Dallas residents from Walgreens, Blue Bell Creameries, and Easyvet pet prescriptions. Walgreens employees load drones with orders for liftoff from store parking lots with Wing employees providing remote monitoring. In other drone news, Sweden’s Everdrone delivers defibrillators for medical emergencies. The drones, dispatched in response to an emergency call, can arrive before emergency medical personnel. They have already saved one patient suffering cardiac arrest by providing the lifesaving equipment within the critical 10-minute treatment window. Perhaps other emergency medical technology, such as the robotic CPR machine or anaphylaxis-reversing EpiPens®, could also be sent ahead by drones. After years of anticipation, it’s strange to think neighborhood skies could soon be filled with drone traffic.
Miscellaneous Stuff
Germicidal 222
Far-UVC light with a wavelength of 222 nm is highly effective at killing 98% of airborne viruses, like influenza and COVID, but does not impact human tissue. Unlike harmful lower frequency UV light (UVA, UVB, and 254 nm UVC) that can penetrate tissue and cause damage, 222-UVC is strongly absorbed by proteins, so it doesn’t get past our outermost layer of dead skin cells or the tear layer that protects our eyes, making it safe for human exposure. Convincing the public (which can’t grasp the difference between ionizing X-ray and non-ionizing 5G radiation, which is entirely safe for humans) could be the hardest part of broad 222-UVC adoption. Prior barriers to implementation included cost and material used to make far-UVC light sources, but new nitride-based semiconductor emitters could be a gamechanger. Although I am not familiar with the manufacturing, if it follows a typical semiconductor process, I imagine this wavelength could be added to any household bulbs cost effectively. In the meantime, there are commercially available room air purifiers that incorporate a germicidal 254-UVC bulb.
Jimmy Swap
On April Fool’s Day, late night talk show hosts Jimmy Kimmel and Jimmy Fallon pranked audiences and staff by hosting each other’s shows for the night. In some fantastic, pre-recorded footage, the two hosts kept swapping places with each other while fans were getting their photos taken. Some people didn’t notice the switch despite the Jimmys’ distinct looks (beard vs. no beard) and voices. It’s yet another delightful lesson that you should not immediately trust your brain’s interpretation of your surroundings.
DST Sunrise Map
In the unlikely event the US does move to permanent daylight savings time, Bloomberg created a cool visual of how many days per year the sun would rise before 7am depending on your latitude and proximity to a time zone line.
Stuff about Geopolitics, Economics, and the Finance Industry
VC Chill
Global VC investments dropped 19% in Q122 as the industry began a significant and necessary valuation correction (see Bubble, Bubble, Toil, and Trouble). Investment activity likely declined due to global uncertainties and a correction in public market valuations despite ongoing fundraising for new VC funds. As central banks begin to suck excess liquidity out of the economy, correcting the over-stimulus that took place during the pandemic, a far greater negative impact will be felt in the VC market. While investors will surely suffer, the biggest victims will be the companies that raised too much money at too high of a valuation and their employees, who will see paper profits – and possibly their jobs – evaporate.
Transparency, Workers’ Rights, and Automation
In the Information Age, it’s easier than ever for employees to see how they will be treated by employers. The rise of unions among service workers in the US comes as a large demographic headwind collides with years of the free market and government policy failing to arrive at fair levels for wages and benefits. The trend is likely to continue unless we have a significant, unlikely reversal in immigration policies (see demographic discussion in #331). Companies that are already maximizing non-zero-sum outcomes for their employees – in a way that still allows them to deliver products and services to customers at reasonable prices and sufficient margins – will be well ahead of companies that squeezed employees and product quality in favor of higher profit margins. Howard Schulz’s largely symbolic cancellation of Starbucks' share buyback plan is a tidy example of the shift from favoring investors to employees as the balance of power tilts in the economy. Ultimately, this type of reprioritization will create long-term shareholder value as these companies become invaluable supporters of their ecosystems. I'd be remiss if I didn't also tie this rise in worker power and wages to the prevalent (but likely misinformed) market theory that we will see deglobalization in the coming years. As I've written before, globalization is a one-way street, and the lack of labor in the West is a powerful reason why we are unlikely to see deglobalization on any significant scale before we have major improvements in automation capability and capacity (see #336-339 for more). Automation with software and robotics will be increasingly existential to the economy over the next decade. For example, Walmart recently raised long-haul driver pay from $87,000 to $95,000-110,000. Such cost pressures will hasten the development of autonomous and semi-autonomous transportation technologies. The situation is not isolated to the US, as the UK is seeing the fastest increase in starting salaries since 1997 amid labor shortages, according to Bloomberg.
Tech’s Increasing Role in Global Conflict
From the wheel to the computer, technology has done more to improve quality of life over the course of human evolution than anything else. Our ability to innovate defines our species. But, that same technology can also supercharge the destruction wrought by our primal, tribalistic instincts. If we want to prevent/minimize such future tragedies, it is imperative that technology executives, governments, and investors do much more to understand how technology enables war and oppression. We don’t need to keep relearning the unbearable lessons of war firsthand – we have archive and real-time footage of wars, historical records, and survivors attesting to its loathsome nature. Yet, despite this knowledge, not enough has been done to keep technology from enriching and enabling countries that want to wage war against others.
Semiconductors and electronic components provide the foundation for the analog-to-digital transition of the global economy. As the building blocks of smartphones, cloud computing, social network apps, industrial equipment, health technology, robotics, and more, there are increasingly few things in the world that won’t depend on chips in some way or another. This applies to military and lethal weapons as well. A drone can be used recreationally – or strapped with explosives. Add image sensors, facial recognition, cellular connectivity, and cloud computing, and you have extremely dangerous weapons of mass destruction potentially available to any government or individual.
Technology underpins today’s military, from tanks to fighter jets to satellites to communication. While it’s important for technology leaders to work with democratic governments and their defense organizations to safeguard humanity’s future and prevent war and oppression, it’s equally important to keep technology out of the hands of those that would use it proactively to threaten or cause harm to people and the environment on a more pernicious level.
Due to the complex nature of the semiconductor supply chain, with its multiple layers of distribution, it may not be easy to determine where chips ultimately end up. A microchip for a garage door opener or a child’s toy might also go in a tank. Government sanctions might effectively control semi manufacturing equipment (and the sale of advanced chips direct from manufacturers) but might be difficult to enforce more broadly. In a perverse example of difficult sanctions, Ukrainians currently can't get the chips they need to build their anti-tank drones. Software providers also bear responsibility, as operating systems and cloud platforms run algorithms that control electronics and process data in a way that could enable and/or render weapons more effective.
As a long-time investor in technology, I am a believer in the power it wields to bring about increased progress and innovation, but I fear a future where that technology is in the wrong hands. I call upon chip makers, distributors, and governments to do more in tracking and controlling the shipments of electronic components. Increasingly, as electronics require connectivity to function, it may be possible to securely authenticate chips for usage in different applications. Software makers, cloud platforms, and cellular service providers all have a role to play as well to gain more control over the technology supply chain and how semiconductors and software are used.
All of the important technologies of the Digital Age, including chips, hardware, and software, have been largely developed by democratic countries that value freedom and human rights. However, there is a direct line from the iPhone in your hand to Apple's unwavering support of China to China's unwavering support of Russia to Russia's mass killing of innocent civilians. If we, as a free society, value how technology is used, we should do whatever it takes to better control its usage. If we had known a decade ago how chips and software would be enabling slaughter in Ukraine today, I'd like to believe we would have tried harder to prevent it. I would encourage all participants to work with each other and governments to tackle this increasingly important issue for the sake of humanity and the planet.
✌️-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.