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: We take a look at the ramifications of going from billions of humans to trillions of LLMs becoming interactive agents in the economy. This phase shift will create a fascinating way to slow down time, which takes a handful of analogies and movie references to explain. Also in today's newsletter: GM is replacing OnStar reps with Google's LLMs; late-night talk show hosts unite; Stephen King readies himself for LLMs to demonstrate creativity; and a look at how ergodicity economics explains the value of partnering and sharing, a topic dear to us at NZS Capital.

Publishing Schedule for September: SITALWeek will be on break September 10th and 17th, returning on September 24th. If you happen to see me wandering the streets of Sydney or the halls of a hotel conference center, feel free to say hello.

Stuff about Innovation and Technology
Simulacrum
Yes...No...Maybe...Attend For Me? Google Meet’s AI assistant, Duet, will soon be able to attend meetings on your behalf. If you select “attend for me” on a meeting invite, Duet will auto generate some topics it thinks it should bring up on your behalf and it will take notes for you. Google also has a beta product called NotebookLM that could evolve into a training ground that would allow you to create an accurate proxy of yourself as an LLM. I think one of the most interesting impacts of chatbots is that we will go from eight billion human agents interacting in the world economy to trillions upon trillions of human analogs interacting in simulations of various parallel realities. Why have one Einstein when you can have 1,000? Why have one CEO when...well, maybe let’s cap the number of CEOs. Why have one version of ourselves in meetings when we can have 100 versions debating and divining new ideas and solutions to pursue? Why not get input from outside experts, both extant and historical? In the not-too-distant future, the bulk of decision making will likely occur via chatbot confabs in data centers. 

Complex adaptive systems like the Earth’s economy have emergent and chaotic outcomes in no small part related to the number of interacting agents, so supersizing the variables might lead to wildly emergent behavior (See: Complexity Investing for more on complex systems science). Essentially, chatbot multiplication and interaction will increase the probability of finding interesting solutions to challenging problems, but it will also vastly increase the volatility we all experience. And, you’ve probably already guessed the negative ramifications this will have for white collar jobs, i.e., they are rapidly simulated into irrelevance in this particular version of the future. If you’re an avid reader of SITALWeek and you haven’t seen Her yet, I am a little offended. Be forewarned that I am going to discuss the 2nd half of the film, so if you haven’t seen it, go see it now. Spike Jonze’s brilliant vision in writing and directing Her was the concept of an infinite number of chatbots confabulating with not only themselves but also any historical figure or theoretical representation of a human mind. Perhaps his biggest insight though was that infinite interacting bots is equivalent to slowing down time. As the “OS” (i.e., chatbot) Samantha says near the end: “It’s like I am reading a book, and it’s a book I deeply love. But I am reading it slowly now. So, the words are really far apart and the spaces between words are almost infinite. I can still feel you and the words of our story, but it’s in this endless space between the words that I’m finding myself now. It’s a place that’s not of the physical world. It’s where everything else is that I didn’t even know existed.” I find it mind-blowing that Spike Jonze figured out that interacting chatbots would lead to time dilation before LLMs even existed.

To explain what I mean by chatbots enabling us to slow down time: imagine having a million one-minute conversations in parallel in virtual chatbot space rather than a single one-minute conversation in the real world – you’ve slowed down time by a factor of a million, meaning you got one million times the output in the same amount of time. This is like being near a gravity well: ten minutes might pass for you, but the distant world experiences hundreds of years of progress. What you want is access to all that data – you want to relax and sip your coffee while a frenetic network of chatbots churns through a century’s worth of computations, and then spits out an optimized solution while your beverage is still hot. Time dilation is hard for us humans (simulated or otherwise) to wrap our brains around. One of my favorite time travel movies about the gripping effects of general relativity on humans is Interstellar. So, if you want some homework while we’re off for the next two weeks, (re)watch Her and Interstellar. And, for extra credit, if you want to grok the ramifications of the gears of time turning at different speeds, which is perhaps the most important concept to internalize for navigating the coming disruptions, watch 2017’s indie film Time Trap.

Another way to think about this accelerated throughput is as a new form of modeling. Rather than running a bunch of entirely useless Monte Carlo simulations based on randomness (or using the faulty expected utility theory that underpins all of modern economic theory), you could run a million alternate realities with human proxies and see which ones have the most interesting and useful outcomes. Think about all the questions we might be able to answer in the blink of an eye – what’s the best solution for autonomous driving? How do we overhaul healthcare? How do we turn social media into a beneficent, unifying platform? (ok, even AI probably can’t solve that one!) Google and Stanford are already working on modeling societal behavior by simulating a neighborhood of interacting LLMs called Smallville. Some of the residents of Smallville started spontaneously going to the bar at noon and developed drinking problems. And, of course, yes, we might be currently in one of these LLM simulations, but let’s not dwell on that. If you want to run your own Smallville-like sim, the code for the open-source variant, AI Town, is available on GitHub. Massive virtual world games, like Microsoft Bethesda’s new Starfield, are likely going to be great testing grounds for running millions of parallel LLM agent realities.  

If LLMs do end up massively multiplying the effective number of interacting agents in the economy, our world could become largely deterministic, i.e., these alternate realities could drive business and policy decisions that drive the actual economy. Whereas I always caution against trying to predict the future, if our world comes to rely on this type of ultra-high throughput alternate reality modeling, it’s perhaps more accurate to postulate that our predictions might start becoming the future. This form of time dilation could catapult society forward. However, as new solutions are conceived, they will collide with the negative feedback loops of slowly moving progress in the real world (see When Positive and Negative Feedback Loops Collide). In addition to unexpected and emergent phenomena, this tension is likely to result in ongoing anxiety for humans.

AIStar
GM is using Google’s conversational AI agents to talk to OnStar customers. I increasingly find myself interacting with chatbots (instead of human agents) with largely positive improvements to customer service. It raises the question: why am I interacting with all of these customer service chatbots when my chatbot avatar should be interacting on my behalf? I’ve used Rocket Money’s chatbot agent in my stead to drive down many of my recurring bills, and I am definitely ready for an personal "LLM Brad" with more extensive capabilities. Instead of turtles all the way down, it’s going to be chatbots all the way down.

Miscellaneous Stuff
Strike Force Five and Artificial Comedians
Four late-night talk show hosts, Fallon, Colbert, Kimmel, and Meyers, plus honorable member John Oliver, have been Zooming regularly since the writer’s strike shut down their shows in May. As some of you know, I am an avid watcher of late-night talk shows and will typically start my day by skimming through all of the previous night’s episodes. So, while this strike is hitting all the late-night staff hard, it’s also created a large air pocket for me in humorously intelligent commentary on the world. I miss it dearly. The five hosts mentioned above have turned their regular meetings into a new podcast called Strike Force Five, with advertising proceeds helping to float their out-of-work staff. At the heart of the writers’ strike (and the related actors’ strike) is anxiety over AI displacing humans on page and screen. For better or worse, the transition is inevitable, and, even if the Hollywood Studios agree to not use AI, it won’t matter because entertainment now exists largely outside of Hollywood (on YouTube, social media, etc.). I discussed this in more detail in Will it Play in Peoria a few weeks back. The FT has a great article on comedians experimenting with AI bots in live shows and, more broadly, AI’s impact on comedy and how we need to adjust our expectations as the capabilities of LLMs grow.

Never Say Never
Commenting on the creativity of LLMs, Stephen King wrote in The Atlantic:
Creativity can’t happen without sentience, and there are now arguments that some AIs are indeed sentient. If that is true now or in the future, then creativity might be possible. I view this possibility with a certain dreadful fascination. Would I forbid the teaching (if that is the word) of my stories to computers? Not even if I could. I might as well be King Canute, forbidding the tide to come in. Or a Luddite trying to stop industrial progress by hammering a steam loom to pieces.
Does it make me nervous? Do I feel my territory encroached upon? Not yet, probably because I’ve reached a fairly advanced age. But I will tell you that this subject always makes me think of that most prescient novel, Colossus, by D. F. Jones. In it, the world-spanning computer does become sentient and tells its creator, Forbin, that in time, humanity will come to love and respect it. (The way, I suppose, many of us love and respect our phones.) Forbin cries, “Never!” But the narrator has the last word, and a single word is all it takes: 
“Never?”

Stuff About Demographics, the Economy, and Investing
Partner to Win
On the heels of Ole Peters’ latest essay on the coin toss paradox and the importance of understanding ergodicity, he has a new post on how cooperation saves the day. Whereas you tend to lose out over time on your own, if you partner up and split the outcomes, you come out far ahead. I’ve covered this topic in the past, e.g., way back in #202 when discussing the Farmer’s Fable. This idea of cooperation is of course no surprise to anthropologists and biologists who keenly understand the benefits of pooling resources/skills in an ecosystem. Whether it’s reciprocal altruism or non-zero-sum (NZS) outcomes in any game, cooperation always wins. We even named our firm NZS because we believe this concept – both in practice and as a lens for viewing the world – is foundational to understanding and success. Peters further explains why mainstream economics undervalues cooperation, and therefore miscalculates human behavior: 
Naturally, this is quite a change in perspective for researchers who are used to optimizing expected wealth (that’s most economists, for instance). Such researchers see no value in cooperation unless new function emerges from the interaction. I can lift you up on my shoulders, and together we’re tall enough to reach an apple on a tree. That sort of thing is understood, where my shoulders acquire the new function of lifting someone up, which they cannot have while I’m alone. But the value of simply agreeing to share my apples with you is not appreciated.
Here is the reason why economics undervalues cooperation, and it’s oddly convoluted so I recommend reading the next two sentences carefully. By focusing on expected value, mainstream economics focuses on an object which grows as fast as the wealth of an infinite cooperative. Adding cooperation in this situation, where it is inappropriately assumed that perfect cooperation already exists, naturally seems pointless. Hence the impression of cold-heartedness we get from mainstream economic theory? I think so.

✌️-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.