I’ve been working on building Single Executable Applications in Node.js. It’s been a fun nightmare—something that I thought would take a week or two has turned into a month or two. I’m nervous that the patterns I’m working on are horrific anti-patterns, far too kludgy and brittle to ever be useful. But it has been gratifying that whenever I ask whatever flavor of LLM is du jour for help, they do the AI equivalent of looking at me with a blank stare before hitting my kneecaps with a tire iron: actively harmful advice only. The training data definitely does not contain this stuff… yet.
Anyway, I need something else to think about. So I’ve been thinking about the peasantry, as one does. There’s a whole through-line that I feel fairly strongly about related to the death of the second estate (aristocracy), followed eventually by the monarch, and what that means for us today—especially when lefties talk about oligarchy being the government we live in. Maybe I’ll flesh it out more someday, but basically the story is along the lines of: God killed the nobility so the peasants could all LARP as nobles, and then the peasantry thanked God by murdering him in cold blood and are currently trying to summon Satan with their calculators.
I want this journal to be more practical, forward-looking, and useful for prediction rather than pure intellectual masturbation (there’ll still be some strokin’). So, how do you convince peasants to do things that are bad for them?
I’ll argue that if you can convince the peasantry that the horrible thing you want them to do is actually something that normally only aristocrats get to do, then it becomes an exciting thing for them.
Examples? Last names—now mostly only useful for creating marital conflict—were originally only a thing for nobles, to clarify issues like succession/inheritance. Peasants, not actually owning anything, did not need to worry about such fancies. Eventually, though, with exciting technologies like the printing press, a more effective administrative state started to take shape that wanted to do things like expropriate taxes and conscript the peasantry to die in pointless continental wars. To do this, they needed unique, fixed identifiers. So all the peasants got to LARP as nobles! They all got surnames and could forevermore be more easily taxed and conscripted.
Oh, is that too old for you? Okay, let’s go to the 20th century in America, where income tax is being rolled out for the first time—but uniquely, it’s initially framed as a tax on the top 2% of income earners in the country. It’s a tax that only the rich pay! As time goes on and the USD continues its terminal nosedive into worthlessness, eventually almost everyone gets to earn this privilege, and being able to consider themselves rich makes the medicine go down easier. Hell, today more people than ever can brag about a “six-figure income” and owning a million-dollar home! Just don’t think about it too hard.
Basically, the more you can leverage peasants’ delusion identity of being a noble, the more you can get them to sing your praises while you abuse them. Practically, if you want to anticipate what new abuse you will likely agree to in the future as a peasant, look to what nobility—if they existed anymore—would be doing.
Humanlike “staff” / “help”
Humanoid robots that can do household labor are going to be the biggest consumer product category. Its marketing only needs to do some wink-wink, nudge-nudge reference to owning slaves, and the peasantry will be furiously taking out high-interest loans to buy one (the dream of every slave is to own their own one day). These robots will respect your privacy about as much as the rest of your technology currently does—which is to say not at all.
Help today is almost entirely done under some very onerous forms of non-disclosure and non-disparagement agreement. I remember a girl I met who worked for [redacted] doing “event planning” for him. My impression was that a large part of the job was being an attractive twenty-something with other attractive friends she would invite to his parties. Suffice it to say, I was never invited to any. There's lots of neuroticism around keeping your privacy as an ultra rich peasant with hired help. The pitch to this group will be that instead of dealing with real humans, and the emotions that go along with them, they will instead be able to enter into complicated and enforcable privacy agreements with robot slave companies.
So many different types of help that the ultra-rich hire—and that means so many product verticals for AI. But for us normal peasants the power dynamic will be completely inverted. Instead of the help fearing for their financial (or physical) well-being, with the possibility of being deported or sued hanging over their head, it will be trillion-dollar tech companies with 200 lb. humanlike surveillance machines in every peasant’s home.
The AI help won’t even have to be embodied to have these problems. AI tutors, AI publicists, AI personal assistants, AI executive coaches. These will all have the same pitfalls, where the nature of the power dynamic being inverted creates perverse incentives. If you disagree with this, you need to have a coherent argument why this will work out differently than the last two decades of tech—where the model that resoundingly won was surveillance capitalism. If not, you need this to be your base case. Hope for the best, prepare for the worst.
In crypto/VC spheres there was this semi-realized fear of debanking from 2017–2024, and the looming boogeyman of central bank digital currencies(CBDCs), whereby the state would control how you accessed and spent your money based on their political whims. Scary for sure, but after the Trump crypto legislation it seems dead. The MasterCard/Visa payments duopoly is secure! A core part of the American experience is allowing two decrepit companies from the 60s to skim several percent from every transaction you make. Who needs public payment rails anyway, communist bullshit and all. Instead, we’re getting private money (stablecoins) and an exciting new wave of wildcat banking. We will have AI agents who’ve been trained/reinforced/fine-tuned to opt for certain products when you ask them to go grocery shopping for you, or plan a vacation, organize your kid’s summer schedule, etc.
There is no need for CBDC capital controls when people stop spending their own money without using AI agent abstractions. Welcome to your future, peasants. As time goes on, it will even be difficult to enter certain markets as a non-AI. Merchants will offer cheaper prices for the AIs; good deals will be snatched up by them before humans can even realize they existed. Have you tried buying a new GPU in the last 3 years? It’s going to be like that but for buying eggs.
Cool story but how can I personally profit from this
The Georgist econ people like framing American economic history as a series of successive frontiers that are opened: the initial colonization, the Ohio River Valley, the Mississippi Purchase, the West, the invention of the car opening the suburbs, remote work opening up second- and third-tier cities. Basically, the main way to make it in our system is getting in on the ground floor on one of these frontiers that end up resembling pyramid schemes due to how tax and property law works in the US. Robot slaves are likely going to open up a new frontier—or, put in other words, it will cause a repricing of the existing asset stock. Opportunity is calling! Are you listening?
If the peasantry writ large is successful in their quest to summon Satan with GPUs, you probably have bigger problems to worry about. Don’t let me stop you if you’re desperate to minimize humans, to view yourself as a pattern-matching meat computer. But if you’re willing to go out on a limb that we don’t have everything figured out, and that maybe there’s something divine about your existence, you might be able to entertain an AI future that doesn’t end up with human obsolescence and artificial superintelligence. Personally, I’m bullish on the quantum mind theories. Anyway, AI Satan is pretty uninvestable IMO.
Ok, so assuming we end up with JARVIS-like AIs that aren’t super-intelligent demons. What then? Well, my friend, we are back in “bicycle for the mind” territory, bigly. Big up mind-bicycles. Already every curious thought you have can be expanded on by chatting with an LLM. What humanlike robots will do is expand it to the space you live in. Instead of tokens per second, it’ll be about actions per minute. You’ll have AI agents drawing up work plans and orchestrating tool/material acquisition to do things like: do my laundry, build a playground, build a guest house, mow the yard, feed the chickens, etc.
Commercial real estate will have significant dislocations, a bifurcation between spots that cater to robots vs. spots that are more human-centric (or ban robots altogether). Robots take up space, so residential real estate will favor larger spaces that are easily accessed and traversed by robots. Walk-up brownstones will be out, McMansions will have a moment. Most of the current housing stock is either 100+ y/o buildings that were meant to only last 30 years, 100–60 y/o toxic dumps filled with lead and asbestos, or more recent builds already rotting and filled with mold due to cheap cardboard-adjacent building materials. Humanlike robots are going to bring about a renaissance of new high-quality construction in the US. Like how the car necessitated building garages, robots will demand architectural changes from housing. You’ll just need to be living in an area with a reasonable zoning/planning board.
The labor dislocations will put massive strains on state and city governments. The default will be that most cities will do their own version of Detroit in the ’80s. I expect the current property tax structure to be massively adjusted in these areas as they try to cope—only to further push people out of them. Look for well-run cities: functioning policing, low debt, large assets. It’s likely that most cities will fail to avoid disaster due to the incredibly powerful public worker unions. You will probably see new AI-native city incorporation—possibly even new U.S. territories. These AI-native cities will have a small group of talented, high-agency orchestrators running most of the city. Instead of relying on property taxes alone, you could see them requiring citizens to donate robot-time—where you delegate your robots to do work for the city for a few days, a Robot Conservation Corps. The point is that AI-native cities aren’t just like our cities today, but with robots instead of humans doing all the public worker jobs, they will be very different to the 1960s relics we inhabit.
The ultimate release valve for the economic/labor dislocations will be the USD, which—by way of greenlighting private stablecoins and wildcat banking—is going to crater in purchasing power. Eventually, the US government will need to do something drastic to prevent people entirely abandoning the dollar for gold/bitcoin. What exactly they’ll do is anyone’s guess. I’m personally betting that a strategic bitcoin reserve, bitbonds, and arm-twisting trade deals will be at least one attempt at stabilizing its value.
So practically, as a peasant:
- Stay liquid
- Have exposure to AI either from your career or from investments
- Own bitcoin
- Don’t buy real estate in a badly run city