The Governor and the Maximizer
Reflections On Attending LessOnline And Fluidity Forum
At the end of the month, I'll be at LessOnline 2025. So will David Chapman!
One of his favorite pastimes is explaining what he’s not doing. Not philosophy, not rationalism, not post-rationalism. Apparently many things are best expressible through a via negativa.
I founded Fluidity Forum in no small part as an attempt to create an in-person scene grounded mostly in his work. But not grounded in his part of the country. So that's also something he didn't do.
If one wishes to focus on one's own local in-person scenes, I support that. I can rarely attend events outside the Eastern and Central time zones. When I found out he's going to LessOnline this year, that clinched it for me. I stretched my budget and booked a flight to California, and rented an Airbnb for an entire contingent of Fluidity Forum attendees to attend LessOnline together.
In an important sense, LessOnline is a rationalist conference. The venue, a delightful retreat center named Lighthaven, is owned and run by the same org that owns and runs LessWrong.com from which the name “LessOnline” is derived. It formed around Eliezer Yudkowsky's “The Sequences”: an enormous set of posts from the mid-2000s trying to ground a theory of rationality in Bayes, decision theory, and cognitive psych.
In another sense, LessOnline could be thought of as a blogosphere conference. Lots of the invited headliners are not rationalists, but come from an intermingling culture that formed through blogs, including the rationalists (like Yudkowsky and Scott Alexander), post-rationalists (like Venkatesh Rao or Sarah Perry), meta-rationalists (like Chapman), and so on. Often they actively reject being grouped together, and there are important reasons for that.
But there is a recognizable style. Their posts are longform, interdisciplinary, allergic to bullshit, and not easily monetizable. Those in this subculture might be the only people in my life capable of reading long form text from me such as an email, and composing a thoughtful response. I consider that one of the most meaningfully useful ontological groupings I have!
Filtering for the ability not only to read and write, but the tendency to hold one's attention on it thoughtfully, leads to amazing in-person conversations. Last year, one of the Fluidity Forum staff paid my way to LessOnline 2024 so I could act as their wingman socially. They wanted to be introduced to lots of strangers, and knew I'm very comfortable doing so in environments like these. I found LessOnline possessed that trait of my favorite events: I was in constant conversations, and 9 out of 10 of them were absolute world-class. Just like at Fluidity Forum.
Creating Your Own Local LessOnline
You might be thinking "I can't go to LessOnline. How can I create an event like that where I live?" I kind of did and kind of didn't. While drumming up attention for the first Fluidity Forum in 2023, I posted polls to gauge interest in various things. Here's the image I made summarizing the results.
I launched the website with a list at the top of the homepage, of various people named on that chart, who had expressed interest in attending Fluidity Forum. All of them canceled, except for the tis.so bloggers, who are in the Midwest. Yes, you can create a local event where the conversations are 10 out of 10. It does not mean you'll attract the same headliners.
The question is, do you need to? Well, if you're in the SF Bay area or New York City, you need headliners. If you're not on the coasts, you need everyone else, which is lots of us. I recommend you start small, and host an Astral Codex Ten “Meetups Everywhere” event in your city. That's what I did for a few years.
Not Folk Localism
Compare “Inventing The Future” with “Bowling Alone”.
Inventing The Future
In "Inventing The Future", Nick Srnicek and Alex Williams critique localism as an aspect of "folk politics". Suppose what you want to do is change the world. What you need is a scalable, strategic, goal-oriented approach to changing what is widely considered to be common sense among the population. Every movement that wanted to pose a threat to the status quo in the past few decades has been doomed by an ephemeral feel-good focus on the local. It's not that Srnicek and Williams oppose the small and local; it's just a part of it. They describe in detail how they status quo was forged in the mid-20th century by a coordinated global strategy of think tanks, symposiums, networks of experts, and publicity. In Scott Alexander’s review of Inventing The Future, it’s summarized like this:
Foster intellectual talent
Seek long-term academic influence.
Push a utopian vision, along with practical first steps within the Overton Window.
Be prepared to step in as saviors when a crisis arrives.
Bowling Alone
According to Robert Putnam's "Bowling Alone" and the recent Netflix documentary about Putnam’s ideas titled Join or Die, the problem isn’t that people stopped Maximizing Impact and Leveraging Alpha, it’s that they stopped joining things which form connections. Attending a potluck or a hobby club may not directly affect GDP or climate policy, but the loss of those small, regular, local connections has led to measurable collapse in social trust, public health, and democratic resilience. The local is the bedrock of anything that needs to get bigger.
Admittedly, both books concern themselves with politics. You might concern yourself with philanthropy, entrepreneurship, healing, or wisdom. But something similar will be at play. Nothing changes the world unless it has operations at multiple simultaneous scales. Whatever event you or I create, as part of an overall movement spreading from Berkeley and Brooklyn, will not be at the same scale as in Berkeley and Brooklyn. But look at it the other way around. Those events cannot play the part where we are needed at our scale.
The word "scale" conceals a better distinction: optimization vs regulation.
The Governor And The Maximizer
A recent post from Slime Mold Time Mold described two types of systems: One kind, the utility maximizer, is always trying to push a value as high as possible forever. The classic example is a runaway takeoff of artificial intelligence. This has been one of the main themes of the rationalist community since before its inception; I was already reading Eliezer Yudkowsky talk about it on the Shock Level 4 email list when I discovered it (and him) in the early 2000s. Ironically, the AI safety community which arose in response to this threat is, itself a utility maximizer, because there is a story in that social environment in which the stakes for you could not possibly be higher. The more you hear the word “impact” in a social environment, the more you are hearing the siren song of a utility maximizer.
The other kind, a cybernetic governor acts only when there's an error, and stops once things are back in range. The book “Finite And Infinite Games” by James P Carse never says the words “cybernetic”, “governor”, or “homeostasis”, but nevertheless seems to be about them. It’s all about the infinite game (there is only one) in which the object is to keep the players in the game and continue to play. To do that, we cannot sacrifice all forms of human desires at the altar of only one desire, as utility maximizers tend to do. The subcultures from which Burning Man sprang were explicitly influenced by Carse.
In "The Scene That Became Cities: What Burning Man Philosophy Can Teach Us About Building Better Communities", Caveat Magister has a section titled "Efficiency Third." (Here’s a similar article by him.) Unlike a maximizer, Caveat deliberately calls Burning Man "pointless". I think this means they don't want to scale up one goal, converge on one goal, or purify all but one goal. Instead, he says to give people space, trust, and responsibility, and let them bring the best of themselves, and they’re less likely to burn out. Each person will see something in front of them that needs doing, and do whatever that is in that moment.
But here’s the thing— when the thing they worked on is done, they’ll stop and hang out on a beach chair, just like Slime Mold Time Mold described in their article. The AI safety community, prevalent among rationalists, has often assumed intelligence has a “utility function”. That term means to the degree that you have power and intelligence to do so, you always drive toward your goal until your tile the entire universe with that monomaniacal goal. David Chapman pointed out humans do not have a “utility function”. In my opinion only a few humans act like they have one, and they are usually drug cartel kingpins with a fleet of private jets, a pet tiger, and a feud with James Bond.
If your idea of efficiency is to optimize a singular metric, you lose what Venkatesh Rao calls "slack" and “mediocratopia”, the capacity of systems to self-correct without shattering.
The ten principles of Burning Man contradict each other in a generative tension. When Burners pursue all ten principles, the principles act as checks and balances on each other. The principles act like a ten thermostats sensing errors and pulling back, gently, from ten directions, toward a center.
On the other hand, I was inspired last night by a conversation at the Burning Man Community Forum here in Detroit. Greg (playa name “Daisy”) told me that one advantage of having local small regional Burns (instead of just the original eighty-thousand-person Burning Man festival in the desert) is safe experimentation. There can be a lot of experiments that go to a wild extreme and fly completely off the rails, and it’s okay because all our eggs are not in one basket.
This lesson applies to LessOnline and Fluidity Forum, and the events that I hope will follow.
Cities And Non-Ambition
Any form of change that works at multiple scales probably needs something like both a massively distributed cybernetic governor system, and some crazy maximizer experiments off in little pockets or in big cities such as San Francisco. Where you or I fall in that will depend on how you or I are situated.
I remain wary of being captured by the optimization drive so predominant in the Bay Area. Long before it was famous for startups that eventually exceed a billion dollars, California has always been famous for a proliferation of cults. Why? Because ever since the gold rush, it disproportionately attracts those who seek an outsized meaning in life.
Paul Graham noted in Cities And Ambition that what New York City is to money, and Washington DC is to power, and Hollywood is to fame, the Bay Area is to one's personal value. According to Graham, Berkeley and SF say "you should live better". Silicon Valley just cares about "how much effect you have on the world".
The Bay Area ended up also getting money, power, and fame anyway.
I think about Graham’s essay whenever I visit the SF Bay area or Brooklyn NY. I can feel what Michael Valentine calls the roles in the social web. I feel it acting on me more noticeably than those in my familiar environment.
What does Detroit elevate? A refusal to adopt elite aspiration. Ever since the ribbon-farms, Detroit has attempted to resist optimization schemes imposed from the outside. There’s a pervasive skepticism of grand systems, which have historically failed. Think of it as a strategic preservation of sovereignty.
As a result, a lot of our resources are deliberately not publicized. A couple of years ago, a young reporter moved here from a much larger city, and wrote effectively the first story of her career. She was writing about a generative and beautiful scene, but one which the participants didn’t want to spread past word-of-mouth. She brought an understanding of success based on scaling up to as many strangers as possible as quickly as possible. Joe Edelman has a name for it: piling up strangers. But that scene’s understanding of success was based on homeostasis of something personalized to their specific participants. Her article inadvertently threatened what she was reporting on, which then moved much farther underground. This city embodies temporary autonomous zones, almost down to the level of neighborhoods; you can walk a mile in any direction and feel like you’re in an entire different city. A lot of the greatness that happens in Detroit flies under the radar. And that’s all I’ll say about that here.
Change The World Without Doing World Changing
If you’re waiting for me to judge between cybernetic thermostats and maximum impact leveragers, and declare a winner, I'm not just dodging coming down on an answer. It depends. When the world changes, it involves those who are Doing World Changing but also millions of active people who are not Doing World Changing.
As Chapman put it in Meaningness, when it comes to personal value, the eternalizing stance is "I have a distinct and superior value given by the eternal ordering principle", and the nihilizing stance is "My value comes from being like everyone else", but the complete stance is "Developing all my abilities in order to serve others.” Everyone can do that.
By organizing events like LessOnline or Fluidity Forum, we are doing what Tyler Alterman calls worlding.
A definition of worlding could b something like: the art of creating habitats, i.e. immersive social and physical environments that ppl live inside. These habitats can be either ephemeral (eg LARPs) or enduring & ubiquitous across many aspects of life (eg a punk scene). Their cultural structures (eg shows) will often be afforded by physical structures (eg venues). These physical structures support cultural norms & activities through both architecture and ornamentation.
…
You could say that this sort of thing has already been done by social animals for millions of years. However, what matters for worlding as an *artform* is that it's done from a particular stance: that of passion and imagination. It's about beauty, not just a sense of necessity or responsibility Furthermore, what you get from regarding worlding as an *artform* is more of a bias toward experimentation. I want to see ppl rapidly innovating new ways of life together. And by treating it more like a theatre production than an “intentional community” we’re less likely to drink our own cool-aid and end up with cults
Sounds more like the cybernetic governor than the maximizer.
-Matt