SITALWeek #396
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: the best AIs might be the ones with a good sense of humor; drones are proliferating; credence creeps up for fusion energy; using wearable data to login; the new GLP-1 diabetes and weight loss drugs also appear to boost important cancer-fighting immune cells; and, much more below.
Stuff about Innovation and Technology
Farmer Drones and Air-Pizza
The FAA has approved use of the first eVTOL (electric vertical takeoff and landing) drone, and its primary function is to autonomously spray crops. The drone’s maker Guardian Agriculture says that it covers a wider area with more precision at a lower cost than a human-piloted plane or helicopter. Guardian has over $100M in orders for the $300K drones. In other drone news, Zipline announced it’s on track to hit 1M deliveries this year and has its first restaurant partner for their new P2 drone system. Pagliacci’s Pizza has worked with Zipline on a new drone-compatible, double-decker box that holds two 13” pizzas. Pagliacci’s has multiple locations in Seattle, one notably just a few miles from Jeff Bezos’ house. If this new partnership gets off the ground, it shines a spotlight on the continued struggles of Amazon’s own drone business. I wrote more about Zipline and the broader potential for innovation in the infrastructure layers of the economy in #389.
Drive-Thru Bot
The WSJ reports that Wendy’s will trial the use of Google’s AI chatbot technology to replace humans in taking drive-thru orders. Just last year, McDonald's experiment with automated voice-based ordering effectively failed, but the recent commercialization of LLMs is likely to be game changing for many voice-based interactions. Google Cloud’s CEO Thomas Kurian noted: “You may think driving by and speaking into a drive-through is an easy problem for AI, but it’s actually one of the hardest.” Wendy’s sees 80% of orders come from drive-thrus, up from two-thirds pre-pandemic. The AI chatbot is also programmed to upsell you.
Comedic Brain
A few months ago, we had the new Bing-Chat (which turned out to be powered by GPT4 in creative mode) formulate a series of jokes based on specific dog breeds. And, they were actually pretty clever, which implies the AI has some understanding of the nature of comedy. This was one of the interactions that crystalized my views on LLMs being a new form of alien-like intelligence. There was quite a lot of press last week around the departure of Google’s “Godfather of AI” researcher Geoffrey Hinton due to Hinton’s concerns about AI. While Hinton generally praised Google’s methodic approach to product release (although at last week’s Google I/O developer conference, it appeared to me the slow rollout vs. OpenAI is more related to technological lag), I was most interested in Hinton’s comments about comedy in a Wired interview:
When asked what triggered his newfound alarm about the technology he has spent his life working on, Hinton points to two recent flashes of insight.
One was a revelatory interaction with a powerful new AI system—in his case, Google’s AI language model PaLM, which is similar to the model behind ChatGPT, and which the company made accessible via an API in March. A few months ago, Hinton says he asked the model to explain a joke that he had just made up—he doesn’t recall the specific quip—and was astonished to get a response that clearly explained what made it funny. “I’d been telling people for years that it's gonna be a long time before AI can tell you why jokes are funny,” he says. “It was a kind of litmus test.”
His second revelation concerned the substantial capabilities of only moderately complex systems:
Hinton concluded that as AI algorithms become larger, they might outstrip their human creators within a few years. “I used to think it would be 30 to 50 years from now,” he says. “Now I think it's more likely to be five to 20.”
There is something special about the comedic brain and the way it sees the obvious and connects dots in ways most humans tend to miss. In a recent interview with comedian Steven Wright, Conan O’Brien describes the comedic mind as broken in a useful way: “The thing that happens in evolution to keep us alive is our brain making connections that make sense...there's almost something like a miswiring that wouldn't work in evolution, but we're connecting things that shouldn't be connected.” This interaction was sparked by Wright’s famous joke: “I went into a restaurant that said they served breakfast anytime......so, I ordered French toast from The Renaissance.” I spent more time than I’d like to admit trying to get Bard and GPT-4 to come up with a funnier joke about restaurants that serve breakfast 24/7 with nothing to show for my efforts. However, I asked GPT-4 to explain Wright’s joke to me to see if it understood why it’s so funny, and here is the response:
This joke is an example of a play on words, or a pun, which is a common type of humor. The humor lies in the ambiguity of the term "anytime". The speaker interprets it literally, choosing a time period rather than a time of day.
In the context of the restaurant, "breakfast anytime" is meant to imply that you can order breakfast foods at any hour of the day, even if it's typically past the traditional breakfast hours.
However, the speaker humorously misconstrues "anytime" to mean any time in history, thus ordering "French toast from The Renaissance" – a time period in history (14th–17th century) rather than a time of day.
The humor comes from this unexpected interpretation and the absurdity of ordering a meal from a specific historical era, something that's obviously impossible in a restaurant context.
When I asked the same question of Google's Bard with its upgraded PaLM 2 LLM, it sort of got the joke, but not nearly as well as GPT-4, and it made up several erroneous explanations. Perhaps Bard's weak sense of humor explains some of its other shortfalls in wisdom.
Betting on Fusion
Microsoft signed an audacious agreement to purchase 50MW of fusion-generated electricity starting in five years from Helion, a company backed by OpenAI co-founder and Microsoft partner Sam Altman. There are many gigantic hurdles to the holy grail of energy, which replicates the process of how stars generate their fuel, and there is no shortage of attempts being made. Google’s DeepMind has been using AI to help control the volatile plasma fields with some success. One of the biggest hurdles is sourcing tritium, a key fuel for most fusion projects. Currently, tritium is sourced from fission reactors, most of which are being phased out, and since tritium has a half-life of 12 years, by the time fusion works (if it works) we won’t have much of this fuel left. While researchers are working on self-sourcing tritium from fusion reactors, that appears to be a long way off. Helion, which doesn’t use the conventional method of lasers to control a plasma field (which is DeepMind’s focus), also avoids the need for tritium by heating deuterium and helium-3 into a plasma and controlling it with magnets. Helium-3 itself (an atom composed of two protons and one neutron vs. tritium’s one proton and two neutrons) is very rare, but Helion claims to be able to produce it by fusing the hydrogen isotope deuterium (one proton one neutron), which is relatively abundant on Earth. This YouTube video demonstrates how Helion’s fusion system works in more detail. While it might remain a safe bet to assume large-scale fusion energy production won't take place anytime in the next couple of decades, it does appear we should be inching up our credence that it could happen in the next 5-10 years. This creates a series of interesting thought experiments with abundant, cheap, green energy. The entire thrust of today's green push would pivot dramatically. Further, if AI is critical to solving fusion, and advanced AI is only available in the West, there are going to be significant geopolitical ramifications to an early arrival of the stellar energy source.
Wearable Passwords
Health wearable maker Oura acquired identity company Proxy. Proxy is said to be working on tools that use biological data to confirm identity, which would theoretically enable health monitors like the Oura ring to act as a tool for payments, access, or, ultimately, perhaps password replacement.
Miscellaneous Stuff
GLP-1’s Immune Boost
The popular GLP-1 class of weight loss and diabetes drugs was also shown to boost the activity of cancer-fighting immune cells known as natural killer (NK) cells in a small clinical study. The increase in function of NK cells was independent of weight loss, suggesting an independent mechanism.
Truman’s Jury Duty
Jury Duty is a new bingeable show on Amazon’s Freevee channel (I watched it via Prime Video). The show follows a fake jury trial where only one member of the jury, Ronald Gladden, isn’t in on the act. Gladden, a 29-year-old solar installer, spent weeks in fake sequestration under the assumption he was part of a documentary on the judicial process. It’s a bit like a real-world version of The Truman Show, and it has many delightful moments throughout. I of course loved it because it plays with society’s increasingly blurred lines between fact and fiction (see Cinéma Vérité). In yet another reality-defying twist, Amazon plans to submit Ronald Gladden for a best-actor-in-a-comedy-series Emmy, despite him not even knowing he was in a TV show during filming.
Dolly '24
After watching this video of Dolly Parton premier the first single, "World on Fire", off of her first rock & roll album, Rockstar, last week at the American Country Music Awards, I had only one thought: why isn't Parton running for President of the United States?
And, lastly, a happy Mother's Day to all our maternal readers!🌹
✌️-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.