SITALWeek #289
Welcome to Stuff I Thought About Last Week, a collection of topics on tech, innovation, science, the digital economic transition, the finance industry, hummingbirds, 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: creator confusion; manufacturing exits China; Salesforce and Slack's new collaboration tools; AI from robots; SoCs, Intel progressing, and Phoenix rising; Amazon’s expanding healthcare initiatives; a mystery particle; EVs on the road; and lots more below...
Stuff about Innovation and Technology
Cloud Spending Outpaces Legacy
Enterprise spending on cloud computing reached nearly $130B in 2020, passing on-premises for the first time, with 35% growth in cloud compared to 6% decline for on-prem to $90B.
Digital Ad Power-Law Winners
Google, Amazon, and Facebook garnered 90% of digital advertising in the US in 2020, up from 80% the prior year. That tally of around $117B amounts to half of all US ad spend, according to GroupM estimates as reported in the WSJ.
Streaming’s Surge and Carbon Footprint
Streaming video subscriptions grew 26% globally to 1.1B in 2020. Nielsen pegs Netflix as having 19/20 top original and acquired streamed series, while Disney+ had 7/10 top streamed movies. In related news, an hour of Netflix video streaming uses about the same amount of carbon as driving a car a quarter mile.
Watching Wearables
Smart wearables grew 27% in Q4 2020 to 154M total units and 28% for the year to 445M. IDC included earbuds (hearables) in the tally if they also contained some sort of fitness tracking or smart assistant.
Airborne Virus Vigilance
Thermo Fisher Scientific’s AerosolSense Sampler is a $5,000 device that can sense the presence of viruses like COVID in the air. One expert said it’s like “looking for a needle in a haystack in a field of haystacks”. The device isn’t real time – collection cartridges require removal and laboratory analysis to obtain results – but potential applications include compliance for hospital safety protocols. If we’ve learned anything in the last year, it’s that we’ve been underestimating air as a vector of viral spread. I'd expect an increased focus on indoor air quality.
Robots Can Accelerate AI
Jeff Hawkins, creator of the Palm Pilot and founder of neuroscience company Numenta, has some intriguing ideas around the importance of embodiment to developing artificial general intelligence. Hawkins believes researchers should focus on robots interacting with the real world and learning to build artificial intelligence instead of deep learning in the data center. Robot progress has been slow as this WaPo article points out, but that is an overly cynical take. We see this mistaken analysis quite often when a group of companies or experts is stuck in an Industrial-Age mental model, unable to see how the consumerization of a sector brings innovation, scale, and simplification. I’m not saying Rosie from the Jetsons is around the corner, but let’s not forget what happens when an industry shifts into the Information/AI Age like robotics is about to do – I covered the robot revolution in more detail at the top of SITALWeek #257.
Salesforce Advances Collaboration Tools
Salesforce has announced an AI overlay on Zoom that gives you information on conference participants and helpful advice, like suggesting to slow down your talking speed. It’s hard NOT to imagine the world five years from now with Terminator-style, augmented-reality AI overlays. The vast potential this technology has to eliminate cognitive bias in the decision making process is hard to wrap your head around – sensors and AI will easily be able to spot bias when we can’t spot it ourselves, and eye tracking and health sensors might know our preferences before we are conscious of them. It breaks my brain to think about it. In other Salesforce news, its pending acquisition target, Slack, announced you can Slack to Slack with other companies, an effort to remake the tool into a full competitor of email. From one perspective, buying Slack is part of a larger arc in the Benioff vs. Nadella battle. Benioff got a little irritated when Nadella snaked the LinkedIn deal away from him five years ago, so he paid $750M to hire Bret Taylor for his would-be Office-killer Quip. Taylor is now President of Salesforce. Benioff chose AWS as Salesforce’s preferred cloud as the LinkedIn deal fell apart. Benioff is perhaps a lover scorned after the rumored merger between Salesforce and Microsoft came to naught the year before the LinkedIn deal was inked. Does his revenge plot now have a chance at success as Slack iterates quickly on features and will have the full Salesforce sales force backing it soon? What else does Salesforce need to storm the walls at Redmond?
Fools Rush in Where Angels Fear to Tread
For Microsoft, first it was “Windows, Windows, Windows!”, then “developers, developers, developers!”, then “advertisers, advertisers, advertisers!”, and now it’s “creation, creation, creation!”. As Microsoft is rumored to be buying chat service Discord, Bloomberg reported Satya Nadella’s words from a month ago: “Creation, creation, creation—the next 10 years is going to be as much about creation as it is about consumption and about the community around it, so it's not creating alone...If the last 10 years has been about consumption—we're shopping more, we're browsing more, we're binge watching more—there is creation behind every one of those. But I see that phenomenon being much more democratized.” I think I might agree with the spirit, but I am having a hard time with the math. If everyone is a creator, then who has time to consume? What is the ratio of creators to consumers in today’s economy, perhaps 1:100? For every person working hard to create art, 100 people are sitting back and consuming it. Where could the ratio go – 1:10? A 1:1 ratio seems illogical without an absurdly densely overlapping Venn diagram of creators and consumers, but maybe I lack sufficient imagination.
Kevin Kelly famously wrote about 1,000 true fans in 2008 (and, Kevin is right about everything 20 years early!). A few years before that essay, Marc Andreessen started Ning to go after niche networks of connected people and super fans (it failed, and I am always surprised by how few people have heard of it!). Twitter is trying to build a business around Super Follows. What is the motivating force behind this new creator community building? Could it be that in our digitally isolated world, we’ve lost our sense of physical community? In a 1973 Playboy Magazine interview (which I reference frequently!), author and human-condition commentator Kurt Vonnegut had the following soul-piercing commentary on community:
"Until recent times, you know, human beings usually had a permanent community of relatives. They had dozens of homes to go to. So when a married couple had a fight, one or the other could go to a house three doors down and stay with a close relative until he was feeling tender again. Or if a kid got so fed up with his parents that he couldn’t stand it, he could march over to his uncle’s for a while. And this is no longer possible. Each family is locked into its little box. The neighbors aren’t relatives. There aren’t other houses where people can go and be cared for. When Nixon is pondering what’s happening to America—“Where have the old values gone?” and all that—the answer is perfectly simple. We’re lonesome. We don’t have enough friends or relatives anymore. And we would if we lived in real communities...For a community really to work, you shouldn’t have to wonder what the person next to you is thinking. That is a primitive society...Human beings will be happier—not when they cure cancer or get to Mars or eliminate racial prejudice or flush Lake Erie but when they find ways to inhabit primitive communities again. That’s my Utopia. That’s what I want for me."
Community is essentially based on tribalism, and tribalism usually leads to strife and zero/negative-sum outcomes. Humans have experienced a disturbing rise in us-versus-them tribalism over the last few decades. Is a world of mini tribes built around creators a solution or an even bigger problem? Can we have community without the rampant negative effects of tribalism? (e.g., organized sports seem to occupy a comfortable, largely good-natured middle ground; how can we recreate that dynamic?) Can these rising creative leaders stand the pressure? I love that new tools are democratizing access and making it easier to create and share. But, creating art is hard for a reason. Entertainers and artists make it look easy, but great art belies an immense amount of hard work. Creating art is vulnerable: “The poet stands naked before the world” is a quote attributed to beat poet Jack Kerouac. Should everyone be a creator?
Phoenix is the New Hsinchu
Is Intel being run for customers again instead of shareholders? White knight Pat Gelsinger has galloped in with a plan to spend on manufacturing and foundry initiatives instead of share buybacks. Recall back in SITALWeek #287, before Gelsinger was hired, I laid out five things Intel could do to get back on track. With the recent announcements, they are checking (or partially checking) off the first four items: installing an engineer as CEO, milking the x86 cash cow, outsourcing leading-edge chiplet components, and forming a fab consortium with big tech clients and TSMC/Samsung (the fifth point was the wild speculation they should acquire Arm).
We’re quite skeptical about Intel’s foundry aspirations. They have tried and failed. Being a services company for others requires a completely different set of DNA, and, more importantly for chip making, a large library of IP and processes for customers to use very early in the design process. One way around these hurdles would be for Intel to license a foundry “stack” from someone like TSMC. Huh? Why would TSMC do that? Well, TSMC has become a victim of its own success. The world has become very dependent on, and thus fragile to, TSMC. We’ve been spotlighting the dangers of Taiwan’s status as the semi supply chain lynchpin for nearly five years now (as we say: “we’re not always right, but we are always early”). While Samsung has a few foundry customers, it has struggled to compete with the full ecosystem and network effects of TSMC for a variety of reasons. The world now basically runs on TSMC...and Intel chips. Intel also needs TSMC until it catches up using ASM Lithography’s EUV technology over the next couple of years. Intel is so bad at being a foundry that TSMC is still making all the chips for companies Intel has acquired over the years going all the way back to Altera in 2015. Potential early partners for Intel’s foundry efforts would be the cloud platforms making custom SoCs (see also section below on Google SoCs); Intel could outsource part of these to TSMC and make other chiplets (as well as the advanced packaging) in their own foundry.
The world will have software running on Intel x86 chips 50 years from now, just like we have IBM mainframes still alive and kicking. Software almost never dies, and, historically, it has been coded for specific processor architectures (this is finally changing with serverless, containers, abstraction, etc. that are allowing Arm to gain traction). Intel is also too big to fail. The world’s largest companies – Microsoft, Google, Amazon, Apple, and more – would struggle to get by without Intel in the next decade. TSMC and Samsung cannot make enough alternative processors in the next 5-10 years to replace Intel, as the capital/time required would be prohibitively large (and, AMD also needs that TSMC capacity). Just looking at the near-term demand for advanced chips is daunting: smartphones and PC’s are growing at a double digit clip and Hynix expects hyperscale data centers to double in the next five years, and decades of software can’t be economically altered to run on Arm overnight. So, if Intel can catchup and perhaps pull a foundry business together, that is Out-Of-The-Money Optionality (OOTMO™ as we call it!). Right now, the global economy is undergoing a decades-long transition from analog to digital and is wildly fragile to the town of Hsinchu, home of TSMC, on the island country of Taiwan. But, a decade from now, Phoenix, Arizona is slated to be just as important. Between announcements and rumors, Intel, TSMC, and Samsung could be building over $70B of new chip fabs in Arizona. On top of the vibrant desert chips scene that already exists, this investment is big enough to recreate an entirely new supply chain for inputs and outputs of the chip industry – packaging, testing, engineering, etc. We have been calling for building in this resilience for a half decade, and it is amazingly gratifying to see it now becoming almost inevitable.
Staying Moore’s-Law Competitive with Collaborative SoC
The Next Platform has an article about the recent departure of a top Intel engineer to work on the System-on-Chip (SoC) effort at Google. In order to keep Moore’s law conceptually alive, Google sees a need for chip makers to come together with various bits and pieces on a single chip instead of across the motherboard or via PCI Express: “In a way, what Google really wants to do is teach the chip makers to cooperate in a way that they really do not, and have not historically. Imagine if you could take bits and pieces of technology from Intel, AMD, IBM, and Nvidia and make the right kind of specific compute device. This is the kind of thing Google is dreaming about, and maybe it can happen if Google buys some IP here and there and integrates it to prove it works. Maybe it will happen at the chiplet level first.” The trend for the last decade in hyperscale cloud computing is for the data center itself to become the unit of compute rather than individual servers (we’ve written about this topic a few times and highly recommend this interview with Nvidia’s founder). And, now we can see a variation taking shape with a fabric of SoCs creating the entire data center.
Amazon Care Baby Steps Toward National Healthcare
STAT reports on Amazon’s new Care initiative, which is expanding to provide healthcare services to other companies following the collapse of the Amazon-Berkshire-JP Morgan JV: “For its part, Amazon has been laying the foundation for what could be a massive, wraparound service for years. It launched prescription delivery service Amazon Pharmacy in late 2020, completed its pilot project with Crossover Health, and launched a health tracker called the Halo that lets users share data with electronic medical record vendor Cerner. Notably, Care is also advertising its ability to conduct virtual and in-person visits focused on sleep, joint health, and overall wellness — components that Trzinski said will play an increasingly important role in employer-focused offerings of the future...Should Amazon take all the beads in its sprawling health care treasure chest and fashion them into a necklace, the company could go even further than Care and create a comprehensive national, technology-enabled health care service that offers everything from prescriptions to virtual and in-person care.”
Miscellaneous Stuff
Hummingbirds Offer Solution to Nuisance Noise
With 12 high-speed cameras, six pressure plates, 2176 microphones, and three years of mathematical analysis, researchers determined how hummingbirds create their distinct hum. Most would agree that, especially compared to buzzing and whining winged creatures, the hummer’s hum is a pleasing sound. This fact is a bit surprising since hummingbird wings beat at 40 times a second, and humans typically find 40-Hertz (40-cycles-per-second) sounds irritating. It turns out the movement of the hummingbird’s wings reorients the pressure wave in space, which is what we are hearing. The wing create bass frequency and higher frequency harmonics at 80 and 120 hertz, which make hummingbirds more like a violin or piano than a propeller. Aside from the cool science, it’s possible we could use these findings to engineer noisy machines like drones with more pleasing harmonics.
New Player in Particle Physics?
Recent experiments at CERN’s Large Hadron Collider imply there is a chance that electrons and muons act differently than previously assumed. The results imply the existence of some new type of particle or law at play in the universe not previously seen or assumed in the standard models. While there is still confirmation to come, odds are there is only a 1 in 1,000 chance the results are wrong (ideally more analysis would move these odds much lower if true). “Dr. Michael McCann, who also played a leading role in the Imperial team, said: 'We know there must be new particles out there to discover because our current understanding of the Universe falls short in so many ways – we do not know what 95% of the Universe is made of, or why there is such a large imbalance between matter and anti-matter, nor do we understand the patterns in the properties of the particles that we do know about.'”
Charting Elemental Origins
This rendition of the periodic table shows each element’s historical source – from the fusion during the Big Bang (hydrogen and helium) to various types of dying/exploding stars, with the outliers beryllium and boron arising from cosmic ray fission.
Stuff about Geopolitics, Economics, and the Finance Industry
Manufacturing Leaving China
Delta Electronics, the Taiwan-based maker of power components for customers like Apple and Tesla, is reducing its headcount in China by 90%.
“‘For China the problem is, even without the US-China conflict, China is no longer a good place for manufacturing,’ [Delta’s Chair] Hai said. Wages in the southern Chinese city of Dongguan have grown tenfold since the company set up shop there in 1992. ‘More discouraging than that is the [staff] turnover rate. In India, it is much more stable,’ he said. Delta has responded with automation. On its most advanced production line in the eastern Chinese city of Suzhou, near-complete automation has cut the headcount from 42 to just three.”
EV Beats ICE for Lifetime Carbon Costs
The WSJ has a really nice visual article comparing the carbon from EVs and ICE cars. You can drive down the article to see the higher upfront carbon for EV manufacture vs. the long-term cost of driving ICEs, with EVs ending up much lower carbon over the lifetime of the vehicle. Given batteries can likely be reused/recycled and power will increasingly come from renewable sources, the gap will surely increase over time. Perhaps more importantly, the vast majority of passenger vehicles only need a range of 100 miles or less, which would also lower the initial carbon footprint for EV manufacturing.
Road Maintenance Shortfall Ahead
Governments will need to find a way to replace gas taxes to maintain roads as the consumer and commercial vehicular fleets shift to EV (or perhaps hydrogen) in the coming decades. In the US, Federal taxes amounted to ~$36B in 2019 or 18.4c per gallon (under 1 cent per mile traveled for the average vehicle). Additional state taxes vary greatly, but average to around 34c for gasoline and slightly higher for diesel bringing the total to somewhere in the neighborhood of $80B per year, or ~2-3c per mile driven. This would amount to roughly $200+ for every 10,000 miles driven depending on the fuel efficiency of your vehicle. As a household of longtime EV drivers, we’ve paid negligible gas taxes for seven years, but we are still driving on the same roads. Some states are turning to vehicle-miles traveled (VMT) taxes, and the Federal government is looking into the idea as well. Given the privacy problems of monitoring mileage for taxation, and the need to incentivize a more rapid shift to EVs, governments best option is to directly shoulder the road maintenance costs rather than offset through new taxes. In the meantime, if you drive an EV, make sure to thank all the ICE drivers out there for paying for your roads.
✌
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.