SITALWeek #401
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: alcohol consumption may be going the way of tobacco; facing the uncertainty of LLMs leads to more questions; gravitational waves everywhere; the power of hot wings; US high tech manufacturing boom; KYC for cloud; and, much more below.
Stuff about Innovation and Technology
I, LLM
The iconic 1979 book Gödel, Escher, Bach, by cognitive scientist Douglas Hofstadter, makes the case that knowledge and cognition arise from the way the human brain operates, and that meaning can come from seemingly meaningless things. Hofstadter, who previously believed machine intelligence approximating humans was not possible until far into the future, recently disclosed in an interview that he felt his deepest beliefs are being challenged by LLMs. In the interview with an inquisitor from Game Thinking TV (who appears to be in over their head), Hofstadter proclaims: “It's a very traumatic experience when some of your most core beliefs start collapsing”. He also professes a “terror that an onslaught is going to catch humanity off guard” and compares AI to the discovery of fire, but fears that the world may already be burning down. He went on to admit: “The human mind is not nearly as complex as I thought when I wrote G.E.B....it makes me feel diminished...extremely inferior.” Like almost everyone else who has a copy of GEB, I too have picked up the book several times only to meet my match and place it back on my shelves. But, from what I have read, I have two reactions to Hofstadter’s cracking belief system: it’s both reassuring (because I too have experienced this feeling of ungrounding) and terrifying, knowing that no one knows what’s in store for the human race as we enter the AI Age. Hofstadter also distills GEB down to the following question: what is it that enables a human to say “I”, i.e., identify as a self? Hofstadter believes GPT4 is demonstrating some degree of “I”, and I think this sense of self will be greatly advanced when AIs have long-term memory. All that said, in an ongoing thread here at SITALWeek where I am reminded that there is no longer any ground truth when it comes to people and what they say, just yesterday Hofstadter penned a long editorial in The Atlantic that seems to refute much of the content of the interview I referenced above. Hofstadter took extreme issue with ChatGPT's poor impersonation of him. This is a response we've seen before: an intelligent and/or creative person has their sense of self threatened by AI, and they react by trivializing and/or criticizing it. It's strange to me to criticize an artificial intelligence of being bad at impersonation, when most people are bad at impersonating other people, and we don't say that makes them dumb, worthless, or dangerous. I think we have a choice here to be John Henry, and lay down and die trying to beat the steam drill, or step back and recognize that maybe LLMs represents the start of something new and useful that will change the future in ways we can't yet predict. If there were a Pascal's Wager for LLMs, I think the safer bet is to remain open minded and intrigued by them, rather than react in anger and fear.
Hofstadter has thus seemed, at least for the moment, to pivot to one of the common criticisms leveled at LLMs: that they are mere tricks, i.e., their apparent manifest human-like behavior isn’t real – they are just cleverly designed search engines (AWS’ head of AI recently called LLMs parlor tricks). This criticism, however, needs to be inverted: the actual trick is our own human consciousness! Afterall, human consciousness, like that of LLMs, appears to emerge from a highly complex form of autocomplete (combined with memory and senses). Once you internalize this idea, it becomes much easier (but still not easy!) to accept how intelligence can emerge from both biological and computational systems, and perhaps simply from language itself. Nearly ten years ago, I wrote a paper attempting to explain that having a sense of self is indeed natural selection’s greatest trick to date, and that paper has been foundational to my understanding of LLMs and the advancements just around the corner that will come from incorporating long-term memory and embedding LLMs into “bodies” with sensors via robots (a PDF of the paper The Funny Thing About Reality is available on my personal website). It’s a time to hold all of our beliefs loosely more so than ever and to ask as many questions as possible. For more tips on asking questions see More Q, Less A.
Miscellaneous Stuff
Celestial Gravitational Wave Detector
The NANOGrav team of scientists released a series of papers in The Astrophysical Journal Letters providing support for the existence of gravitational waves. One prediction derived from Einstein’s general theory of relativity is that space is being constantly churned at a minuscule scale by a large number of gravitational waves emitted from various sources, including black hole interactions. The scientists monitored signals from pulsars, neutron stars that spin hundreds of times a minute and emit regular information (similar to an atomic clock), to look for the subtle signs of distortions that would prove the “subtle swells” of space-time. This study, based on over 15 years of collected data, complements the LIGO project, which was the first to observe gravitational waves using an earth-based (as opposed to a celestial) sensor array.
Hot Ones: CEO Edition?
One of my favorite shows, Hot Ones – “The show with hot questions and even hotter wings”, recently hit 300 episodes on YouTube. The show is a cultural phenomenon that was created outside of the Hollywood studio system; yet, ironically, it’s driven by guests that are largely Hollywood and pop culture stars (e.g., a 2023 episode with actor Jenna Ortega garnered 16M views). The format and venue for the show has allowed it to go somewhere that late night talk shows can’t, with the deep-diving questions by host Sean Evans eliciting open, genuine responses from capsaicin-addled guests. I’ve previously suggested that I would love to see a Hot Ones show with CEOs of public companies braving extreme Scoville levels while answering questions about their company culture and latest products! The show also sells the custom assortment of sauces featured on each episode and even has aspirations to become the most popular wing restaurant in the US. Hot Ones is part of First We Feast, which is owned by Complex Networks, which is owned by Buzzfeed (which is attempting to sell Complex but retain First We Feast).
“There Is No Safe Level of Living”
My potential (albeit very distant) relative, Edward Slingerland, makes the case that the alcohol industry is naively going the way of tobacco by meekly succumbing to an onslaught of efforts to tax, regulate, and scare customers away. Slingerland, a philosophy professor and author of the excellent books Trying Not To Try and Drunk: How We Sipped, Danced and Stumbled Our Way to Civilization, makes the case that alcohol is perhaps different from other regulated substances: while some vices “are evolutionary mistakes—brain hijacks or ancient adaptations gone wrong—our desire for alcohol has had important adaptive functions over the history of our species…It is an essential tool in reducing stress, enhancing individual and group creativity, and facilitating social bonding.” Or, maybe, we should just replace alcohol with high-index hot sauce. Is anyone working on getting LLMs buzzed?
Stuff About Demographics, the Economy, and Investing
AI + Investing
Brinton and I did a webinar on LLMs and investing with Harbor Funds in June. This story covers a few of the topics we discussed.
US Tech Boom
The US Treasury released a report on the historic boom in growth of manufacturing, with an industry breakdown highlighting that the vast majority of growth is coming from the high tech sector. Further, the report notes that the growth, for now, is unique to the US.
Policing the Cloud
The Biden administration is looking to restrict Chinese companies from accessing AI on cloud computing platforms like AWS and Azure without prior government approval. I highlighted this issue last October in #367, proposing a KYC (know your customer) for cloud computing. With VPNs and other various techniques, it may be impossible to keep bad actors from accessing Western-developed AI systems, but it’s worth a try. The important part of a solution will not be weeding out Chinese customers, but knowing exactly who is using what AI services at all times via verification and revoking rights on a case-by-case basis. I have proposed something similar for chips via gated use of programming tools.
✌️-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.