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: AI music is likely to become our predominant listening source soon; TikTok is losing virality; ads in conversations are intrusive, and this may also hold true for AI chatbots; taxbots; it's far too late to regulate AI; cool EUV; the transparent Internet era will be replaced with the secretive AI era; a lesson in staying long-term focused; and, much more below.

Stuff about Innovation and Technology
Functional Soundscapes and AI Streamers
AI-generated music has been in the news a lot lately given its already considerable ability to create ear-catching copycat music (and, even AI-generated podcasts are on the rise, but with less success so far than AI music). Last month, the giant music label UMG asked Spotify to crack down on AI scraping of music and lyrics. This month, however, UMG announced a deal to license their artists and content to the AI music company Endel, which creates functional soundscapes to aid focus, sleep, and relaxation. According to Billboard, there are 15B streams a month of such muzak. I was curious how this compared to the number of songs streamed a month on the leading platform, Spotify, but I couldn’t find an easy answer. So, I asked my research assistant ChatGPT4 (using the Bing Search module) to come up with an answer. Based on 515M active Spotify users, it determined that an average of 90 minutes a day and 3-minute songs would translate to 15B streams per month (and gave me links to supporting evidence for its reasoning). Many of these Spotify streams themselves are ambient of course. So, while I cannot say with any certainty, it seems as though a large amount of music that people are listening to is already AI-created ambient soundscapes, which is likely to increase as more music catalogs grant rights to AI. However, it also turns out that AI bots are listening to their own music in order to boost streams and their own royalties. It’s a robotic self-perpetuating get-rich-quick scheme and an example of how AI can make money for itself. Thus, perhaps AI-music streaming is overstated by AI bot behavior, but the trend follows what we’ve seen in other areas of media – a progression toward infinite content. 

Lens Chips
New metalenses are poised to disrupt some of the image sensor market, and ultimately might find their way into ultra-thin smartphones. Developed at Harvard and commercially produced by Metalenz and chip companies like ST Micro, the new devices offer a host of improved sensor functionality for applications like distance sensing. Importantly, these relatively simple chips have nanostructures capable of detecting not only visible light but also polarization: “Using this technology, we can replace previously large and expensive laboratory equipment with tiny polarization-analysis devices incorporated into smartphones, cars, and even augmented-reality glasses. A smartphone-based polarimeter could let you determine whether a stone in a ring is diamond or glass, whether concrete is cured or needs more time, or whether an expensive hockey stick is worth buying or contains micro cracks. Miniaturized polarimeters could be used to determine whether a bridge’s support beam is at risk of failure, whether a patch on the road is black ice or just wet, or if a patch of green is really a bush or a painted surface being used to hide a tank. These devices could also help enable spoof-proof facial identification, since light reflects off a 2D photo of a person at different angles than a 3D face and from a silicone mask differently than it does from skin. Handheld polarizers could improve remote medical diagnostics—for example, polarization is used in oncology to examine tissue changes.” This type of sensor could be very useful in the coming robot revolution as AI embedded in automatons of all types becomes a reality.

TikTok Deconcentrating
The viral TikTokification trend may have peaked in December 2022, according to new analysis presented in Adweek. Since then, virality and views for popular videos have been declining. While this may be due to de-powerlawing in the TikTok algorithm, it’s hard not to interpret the data as the potential beginning of the end of an era. This is perhaps predictable given the zero-sum nature of TikTok. The TikTokification of consumption habits is something I discussed/lamented in #353: “because content is getting shorter, it no longer becomes embedded into our common cultural lexicon to the same degree as it used to (see Digital Tribalism for more on this theme). The faster we binge or scroll through content, the more forgettable it becomes – with little time to process or appreciate, it evaporates before it can enter our long-term memory.” Perhaps even users’ affection for media platforms like TikTok has become a foreshortened fuse. If this trend holds, and assuming people are still glued to their screens as much as ever, where are they now? Perhaps the ongoing slowdown in long-form video and gaming consumption will make a U-turn. Meanwhile, in what is perhaps the grossest article I’ve read this year, chefs around the globe are making foods gooier and nauseatingly rich for TikToking.

Chat[Ad]bots
Recently, someone commented to me that I interact with AI bots as if they were people. I say “hi”, carry out long conversations, and say “thank you”, “goodbye”, etc. It’s no different than when I am talking to someone I know well. I’ve found this approach produces more meaningful results. You can also tell the chatbot who it is before you start a conversation, such as: you are the world’s expert on xyz. Conversations in general between people are somewhat intimate experiences, so I’ve been thinking about how ads are going to play into AI chatbot conversations. Peppering personal conversations with ads is not a new controversy. At one time, Facebook wanted to insert ads into FB Messenger threads, e.g., if you were planning dinner with friends it would pop in with an ad for a local restaurant (maybe it actually does this today, I’ve never had any Meta accounts, so I am not sure, but I don’t think I’d like an ad served by AT&T when I am texting with family). Email has been difficult to monetize through ads because that too feels like a violation of privacy. You might know Gmail is reading your emails, but you don’t want them to suddenly start discussing personal content with you via ads and offers. Historically, communication between individuals and small groups is not something that’s been monetized with ads for these reasons. If I do have a commercial query and I ask a chatbot to create a matrix comparing various parameters for several product options I’m considering, I am essentially relying on it to become an expert in that area. As such, I might be interested in what it considers relevant ad offerings. Will AI simply become a friend that knows everything about everything, and therefore ads will be tolerable (or perhaps even valuable)? I can see a place for advertising in AI conversations, but I can also envision scenarios where it would feel like a substantial intrusion (it would be nice to have a no-ad subscription service as an option). Advertising is yet another unknown as we navigate a future filled with artificial friends and coworkers, but my hunch is ads might be a hard sell.

Accountants and Tax Software at Risk
It seems like every year for the last 20 years (this is perhaps an exaggeration) the IRS or government has wanted to simplify tax returns or provide easier ways to file them. Right on schedule, the IRS is testing a free e-filing system in 2024. It’s a unicorn that never seems to leave the fever dream. That said, as I work with chatbots more and more, I can’t help but think that my taxes will soon (perhaps next year?) be filled out by AI. Taxes are complex, but they are all based on a set of publicly available tax code documents and structured inputs – a classically easy task for AI. 

Regulatory Capture of AI
I’ve been thinking about the power of the Communications Act of 1996 and how it paved the way for the Internet as we know it today. There is a lot of talk about regulating AI, but the government is typically decades behind regulating new tech (the government can’t even figure out how to assess Facebook nearly 20 years after it was founded). Currently, giants like MicrosoftOpenAI are traveling to Congress and asking to be regulated, even setting forth their own agenda in the hopes of expediting regulatory capture. Such regulation would (as we’ve seen numerous times) hamstring upstart competition by raising the cost of doing business, thus paving the way for the big to get bigger. The Comm Act of 1996 was novel in that it came concurrently (and even ahead of) the commercialization of the Internet. This was perhaps because Clinton signed it into law, and, as everyone knows, Al Gore invented the Internet. The downside was that it gave too much protection to sites like Facebook and YouTube who never had to take responsibility for their impact on users and the world at large. The thought I had is: could we have a Communications Act for AI? Or, at least a version of it for certain industries, like healthcare? An AI Healthcare Act could perhaps allow us to explore experimental drugs and data collection without the extreme wall of regulation that exists today (although it would surely have downsides as well). My base case is: AI is going to move so fast that there is zero chance of regulating it, stopping it, or even understanding it. 

Extreme UV Lithography
Here is a great video explaining how EUV lithography works. It’s a technology without which the modern world would be very challenged, given that the methodology underpins all advanced semiconductors today. The video comes courtesy of the Linus Tech Tips YouTube channel and amassed over 1M views in just its first day. I am not sure if that’s a testament to the power of YouTube or the shifting zeitgeist on semiconductors.

Miscellaneous Stuff
Lose Weight…and Addictions?
As I’ve followed new research on the GLP-1 weight-loss drugs, I’ve seen a few articles about the unexpected side effect of reducing addictive behaviors. The Atlantic writes extensively about this phenomenon, noting that patients have curtailed drinking, smoking, shopping, and nail biting. I was thinking about what the most addictive human behaviors are today on a mass scale, and the obvious answer is people watching TikTok videos on iPhones! So, perhaps the biggest loser in the GLP-1 revolution won’t be the industrial healthcare conglomerization, which endeavors to keep people sick in order to extract more money, but rather Apple and the app store. If the Pentaverate exists, as Mike Myers has taught us, then surely Colonel Sanders and the others are gearing up to squash these evil, anti-capitalism, anti-addiction, cancer-fighting, weight-loss miracle drugs.

Stuff About Demographics, the Economy, and Investing
Information Will Not Be Free
For the last couple of decades, the ethos of the Internet has been driven by the idea that “information wants to be free”. This concept is attributed to Stewart Brand and was popularized by the writer Steven Levy. The full quote (recorded at the first Hackers Conference in 1984 during a conversation between Brand and Steve Wozniak), however, contains something more relevant to the AI era: “On the one hand you have—the point you’re making Woz—is that information sort of wants to be expensive because it is so valuable—the right information in the right place just changes your life. On the other hand, information almost wants to be free because the costs of getting it out is getting lower and lower all of the time. So you have these two things fighting against each other.” If you have exclusive data and/or information, what you learn from feeding that knowledge into an AI model might far exceed the value of that data alone or the value of selling that data to someone else. As a result, secrecy and information hoarding may start to outbalance the trend of Internet-driven transparency. For example, Google recently decided to stop publishing its AI research after trendsetting papers from the last several years fueled much of the current boom in large language models. It might be worth a closer look at your own data and processes to determine what makes sense to set free vs. what to play closer to the chest. Probably the worst thing you could do is to write a free, future-predicting newsletter containing everything you thought about last week and publish it for anyone to see. 

Mental Rails
Last September, on a particularly bad day for technology stocks, I happened to find myself getting picked on by the hosts of Squawk Box (part of that linked clip is cut off, where it got even worse). I don’t like defending investments because investing should be impersonal (e.g., see my Cognitive Lessons from George Carlin), and defending your convictions causes a sense of tribalism about what you own versus what you don’t. You can own a business by buying shares in it, but you never want to mentally “own” an idea about that business; ideas are stories, and they should be held loosely. Thus, taking a defensive (or promotional) stance feeds cognitive bias. If a stock or sector is down, but there is no new information or explanation to be had, then defending it can make you believe your story more than you should. It's far better to simply say “I don’t know” and wait for new information. Of course, this investing strategy assumes you have your positions sized correctly for the potential range of outcomes. When I shared my views that day on why it still made sense to take a long-term optimistic view on the tech sector, the explanation did not seem satisfactory to the television hosts. One of the headline stocks I discussed in that interview on September 1, 2022 was Nvidia, whose shares closed at $139.37 that day. Although I try not to make a habit of it, I found myself on CNBC again last week, coincidentally on another special day for Nvidia (I promise, I did not plan this), with the stock surging 25% to $379.80 on strong earnings results and guidance (the clip here of my appearance last week is also abridged; I should also note this segment had different hosts than the meanies from last September). As I often point out, you are not nearly as dumb as you feel when your stocks are down, and you are not nearly as smart as you think you are when your stocks are up. I might feel a little smarter today than I did six months ago, but I know that I was neither dumb then nor am I a genius today, and surely the market will give us all a reason to feel stupid again very soon.

✌️-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.