Welcome to Stuff I Thought About Last Week, a personal collection of topics on tech, innovation, science, the digital economic transition, the finance industry, and whatever else made me think last week.
Click HERE to SIGN UP for SITALWeek’s Sunday Email.
In today’s post: machine learning audio models make music; VR in the OR; new digital health records requirements; Jonathan Haidt on the fracturing of shared culture; introducing randomness into algorithms; drone-based 3D printers; Lorne Michaels; demographic puzzles; comparing the post-WWII reallocation and prosperity to the rebuilding of the West in the coming decades; and, much more below...
Stuff about Innovation and Technology
AI Singer-Songwriters
Google has developed a learning model called AudioLM that does for songs what language and image transformer models do for conversations and images/videos. Inputting just a few seconds of audio allows the system to expand the snippet into a full song. It won’t be long before we see a movie – soundtrack and all – developed entirely by various transformer models with just a few small prompts.
Increasing Patient Comfort with VR
When I saw the headline of this article: “Taking a child's fear out of the OR using virtual reality”, I assumed kids were being equipped with VR headsets before being wheeled in for surgery, knowing that (as we explored last week) VR reduces pain sensation and even the amount of anesthesia needed. It turns out the idea is to let kids experience the operating room through VR before they go in so it’s not as unfamiliar. VR will eventually become a standard practice for surgeries. If I go in for surgery, I’d like to be exploring space in a VR headset the entire time, and I imagine many kids would rather be experiencing something different than the harsh lights and cold sterility of an OR. I can imagine a future where anesthesiologists are also VR tour guides.
Rx for Digital Records Access
Starting last week, the Federal government is requiring all medical providers to grant patients complete digital access to all of their medical records, as STAT reports. The major change in the way people can access health data should finally open up a variety of apps and services to manage – and ultimately improve – health outcomes. I expect Apple and Google will want to offer medical record aggregation and data collection in combination with their smart watches. There will be many opportunities for innovation as well as increased privacy risk.
Social Media Waterfall
Last week, I watched two appearances by social psychologist Jonathan Haidt. The first was a recorded session for Long Now, hosted by Kevin Kelly and Stewart Brand, and the second was at a Santa Fe Institute symposium. Both presentations left me admittedly a little bit depressed. I covered Haidt’s excellent “Structural Stupidity” essay published in The Atlantic in SITALWeek #342 back in April. Haidt makes many compelling arguments that support what I think is now fairly conventional wisdom that social media has been the biggest (but not the only) enabler of the fracturing of society and erosion of trust. Since The Atlantic article, Haidt seems to have more fuel to back up his claims, in particular the specific circumstances facing kids born after 1996 – especially girls who were teenagers around 2011-2016 as the front-facing camera emerged, Facebook went public and bought Instagram, and the “like” and “retweet” buttons took over collective behavior. Haidt has noted regarding the founding of the US: “the key to designing a sustainable republic, therefore, was to build in mechanisms to slow things down, cool passions, require compromise, and give leaders some insulation from the mania of the moment while still holding them accountable to the people periodically, on Election Day.” Social media of course does the opposite, speeding up time and heating up arguments. Haidt is also honing in on the idea that a lack of common stories is the biggest problem we face as a global society. This topic is near and dear to me, as I wrote in Digital Tribalism last December:
I had hoped this year would bring a meaningful shift from political tribalism to more (generally) benign forms, like professional sports spectating, that might have deescalated some of our more violent, hateful tribal instincts. I worry the trend towards increasingly fractured online tribalism comes at the expense of our sense of community, both locally and more broadly, as well as our ability to treat people as individuals rather than categories. With fewer shared experiences, it becomes harder and harder to find common ground and cultural identity among individuals with increasingly polarizing labels and divergent life experiences and perspectives. Even families are fracturing more and more. I am not sure what unifying celestial force might cause humans to get back on the same page, or at least slow the fractal fracturing. Algorithms are increasingly influencing a larger portion of the population to the point where people are changing their behaviors to feed the algorithm that is subsequently changing their behaviors (this Vox article on the perils of the Spotify year-end review and its identity-creating algorithm is a good example of this phenomenon). The logical progression is towards a planet with eight billion tribes of one, with the individual, defined and crafted by algorithms, believing in a unique set of views that no one else shares. Instead of us vs. them, it becomes me vs. the Universe. Is this just a cultural evolution of the definition of human identity, or is it something much darker? I could see an improvement in leadership from both governments and corporations helping us out of this trajectory, and anything that can slow down the infinitely amplifying tribalism motivators of social networking and media is sure to bring some relief.
Cooperation in order to compete with rivals is one of the most powerful unifying and motivating forces for creativity among humans. We seem to need an enemy in order to make forward progress. Unfortunately, every common enemy that pops up (e.g., climate change) now becomes a fracturing rather than unifying presence thanks to social media algorithms. And, our lack of quality leaders across business, politics, and religions also creates a roadblock to teamwork and common goals. Even if a leader were to emerge, it’s likely social media would undermine their ability to impact people and behavior. It turns out, as Robert Wright points out in Nonzero, that even war can be a positive-sum event on the whole (although that’s certainly no consolation for anyone suffering through conflicts). As media consumption and the speed of information fracture the population into smaller and smaller slivers – or, to use Haidt’s analogy, social media is a waterfall, turning a river into mist – I mourn the lack of shared stories that humans need to function. We cannot even agree on the facts of things that happen. Haidt notes in the Long Now interview that the 9-11 terrorist attacks were one of the last times when a majority of people generally agreed on what happened at the time of the event (although, not that long after, beliefs began fracturing for some people). He suspects that humans are no longer capable of agreeing on facts for anything. Indeed, this seems to be the real legacy of Steve Jobs: he created a device that ended shared storytelling, dissolved culture into smaller and smaller droplets, and put a generation of phone-addicted teenagers into a state of permanent fight or flight (or “defend mode” as Haidt called it at his SFI presentation).
Of course, all technology has good and bad uses, and, in the case of smartphones, we’ve veered toward more negative outcomes than positive. However, as we learn from prior mistakes, we could easily shift back to leveraging technology to make the world better. In some ways, Tim Cook and the Apple of today are trying to reverse the damage with privacy changes that negatively impact the social-network business model. But, short of banning (or, as discussed on the Long Now, de-anonymizing and attenuating the virulence of) TikTok, Facebook, Instagram, and Twitter, the unfortunate and tragic legacy of Jobs’ iPhone and similar platforms continues. It’s ironic because Jobs was one of the best storytellers out there. He seemed to truly understand the value of a story – from Pixar to his infamous reality distortion field. But, now, we’ve lost that ability to craft unifying stories toward common goals. It’s very hard to see a path toward positively channeling our ingrained tribalism and tamping down our reactionary fight-or-flight instincts as screens continue to speed up time rather than slow it down. Is there any way to convince humans to slow down time, or is the path toward increased use of technology that displaces our shared culture a one-way street? As I’ve noted in the past, my favorite form of benign tribalism is professional sports. Sports appear to be about watching two teams dress up in costumes and perform a simulated war, but the real popularity lies in the stories behind the players, coaches, and rivalries. It’s sports’ inherent storytelling that makes them so popular. Thus, it’s perhaps encouraging that the NFL is setting ratings records this season, with some viewer numbers returning to the highs of twenty years ago.
Regulating AI Regulating Us
Algorithms are increasingly likely to be subject to regulatory oversight. A recent example of why oversight is needed is the faulty Epic algorithm that hospitals were using to determine the risk of sepsis. STAT reports that the FDA plans to have oversight of software used in medical decisions. I’ve covered the impacts of algorithms and the transparency problems in Algorithmic Threat to Illusion of Free Will and AI Is the New Dotcom. It seems reasonable that over the next decade, as algorithms increasingly influence most decision making by corporations and professionals, that the government will attempt to gain broad oversight. The problem is that the algorithms are increasingly writing themselves via machine learning, and their opacity is built in. One of the biggest risks of algorithmic control of human behavior and preferences is the loss of randomness. The more algorithmic control exerted over us, the more we are forced to become a caricature of ourselves, without even realizing it. One counter strategy could be to alter the reward systems in reinforcement learning algorithms to favor chance as a desired outcome, a strategy researchers at UC Berkeley have attempted. What if every 20th Tweet or Instagram post was from someone, or was on some topic, you've never expressed interest in before? Of course, that wouldn’t work for a medical algorithm, but it would make life more interesting. Whatever you do, try to make time and space for randomness in your life.
Miscellaneous Stuff
Building by Swarm
Inspired by collective hive building by bees and wasps, researchers equipped drones with 3D printing capability in order to build large-scale structures. Researchers had to tackle a number of technological problems involving coordination, turbulence, and materials selection. For example, to compensate for the oscillations of the flying drone, the printing heads adjust in real time. Employing a lightweight material, with high structural stability when it solidifies, is also key, and I am envisioning a future of homes built entirely out of spray foam!
“This is a year of reinvention. And change is exhilarating”
Longtime readers know I am a fan of Lorne Michaels, the producer of Saturday Night Live, which just kicked off its 48th season. Michaels recently did a rare, long, two-part interview with Dana Carvey and David Spade on their Fly on the Wall podcast covering a wide range of topics about the show. Asked if he (now 77) would ever write a memoir about his life filled with celebrities and high-pressure experiences, Michaels replied: “No, I think the hard part is you have to tell the truth. And, when you get to my age you go, did that really happen?” Michaels also discussed the latest season of SNL in the NYT: “We went through really scary times, the last four years. Hopefully we’re coming out of it and it’s just the old scary things like a depression or war.”
Stuff about Geopolitics, Economics, and the Finance Industry
Spinning a Tale of the Good Old Days
Last week, at the Santa Fe Institute’s annual symposium on risk, Ole Peters presented some interesting data on the reallocation rate in the economy. We’ve covered Ole’s work many times in the past – notably how one person’s debt is another person’s asset, which implies that over the long term declining interest rates might be a one-way street for our increasingly leveraged economy (i.e., rate hikes pose an existential threat to an economy with significant debt burden; see: The Point at Which Higher Rates Collapse the Economy). Rising rates and the strong dollar have the potential to throw the global economy into a recession, as the UN recently noted. In the data Peters presented, largely from around 1940 to 1980, wealth was being distributed and more people were lifted up. Then, around 1980, this trend flipped, and wealth has since been concentrated amongst fewer players in the economy. That mid-twentieth century period now feels like a nostalgic picture of the past: high birth rates, growing immigrant population, building of the middle class, strong manufacturing economy, etc. Then came the personal computer, software, globalization (labor arbitrage sending manufacturing abroad), and automation, all of which seem to have been instrumental in initiating/sustaining a forty-year decline in interest rates, increased debt levels across all aspects of the economy, weak real-wage growth, rising inequality, and, more recently, rising populism and a circus of political developments. It’s hard to say to what degree I am telling a story here comparing the 1940-1980 period to the more recent four decades, but storytelling is life. (Indeed, telling stories is how we make sense of things, and only time will tell whether we weave more truth than fiction.) If the goal is to revert to a distributive era where things tend to feel more equal, I think it points to the importance of reinvigorating some of those potential drivers of growth post WWII. A resurgence in manufacturing seems plausible, but increasing birth rates and immigration feel less likely. And, automation could accelerate with AI, making it even more important to create economic forces that favor wealth reallocation on the margin, rather than concentration. Additionally, if rates do indeed need to stay low to avoid existential asset collapse, then that too will be a headwind to reallocation as debt burdens rise. As I noted last week, a little bit of increased inflation and expansion in manufacturing jobs may be a good sign long term that growth is returning to the economy. Professor Steve Keen discussed inflation and interest rate targets in a recent podcast (paywalled on Patreon), pointing out the lack of data underlying the 2% inflation target used by the central banks (not to mention the wildly flawed concepts of utility theory and the false notion of equilibrium that underlie their policies; see Complexity Economics). Keen reckons inflation up into the high single digits is tolerable for growing economies. I would love to see the West rebuild manufacturing and invest in the green energy transition, and I would welcome mid-single-digit inflation over the next few decades if it brings the prosperity we saw following WWII.
Demographics, My Dear Watson
When it comes to mysteries, demographic puzzles trump any true-crime story for me. If you follow enough clues, you can usually explain unexpected behavior by digging into history and demographics. In January of this year, I puzzled over the paradox of slumping household formations at a time of record low vacancy rates. The explanation I posited at the time was the pandemic had caused enough people to simultaneously occupy two residences that it was distorting the data. A recent collapse in rental demand seems to support that theory. I suspect some folks were renting apartments to work out of during office shutdowns while others were roaming around to different cities while keeping their original place. Additionally, perhaps roommates split up during the pandemic or children moved out of their parents’ house. Subsequent inflation and a return to more normal home/work situations is decreasing demand for secondary housing and likely driving people to go back to roommates or living at home. Record-setting weddings may also impact things, with (in some cases) two people living separately forming one household (this uptick includes pandemic-delayed weddings as well as the bolus of Millennials getting ready to start their delayed families – topics covered last week). We also have to remember the backdrop of near-zero growth in the number of people entering the labor force, as well as greatly reduced immigration levels. All of this paints a potentially negative scenario for housing demand in the coming years. We still have regional housing imbalances, but, overall, there is no shortage of places for households to settle in the US.
✌️-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.