Hello everyone! This is going to be a bit of a housekeeping post and a welcome to new subscribers.
My writing can be intimidating. There is a lot of it, and it’s often dense. As always, choose only the parts relevant to your interests, do not be afraid to make cuts. I attempt to make every post accessible as an entry point, but I also want to build up a superstructure over time. This seemed like a good time to recap some of the very best of my old writing and talk about what I’m up to.
Over many years, this blog has morphed from focusing on rationality to COVID to AI.
But not only those things. I’m interested in almost everything. I write periodic updates about housing policy, childhood, fertility, medicine and health, gaming and grab bags of everything else.
In addition to writing, I also run a small 501c(3) with one employee called Balsa Research. Balsa is dedicated to laying groundwork on a few key issues to make big civilizational wins possible, starting with repeal of the Jones Act. This link is to an update on that, and you can donate here. Your subscriptions here are also very much appreciated.
Underlying it all continues to be my version of the principles of rationality.
Rationality
A lot has changed since my last best-of writeup six years ago. One thing that has not changed is that I consider myself part of the rationalist community. No specific interest in rationality or its modes of thinking are required, but I strive to embody my version of this style of thinking, and to illustrate and hopefully pass on this mode of thinking throughout my writing.
What is rationality? This post is one good answer. It is believing, and updating on evidence, so as to systematically improve the correspondence between your map and the territory, and using that map to achieve your values.
To me, a rationalist continues to be someone who highly values, and invests in, the version of this process and the art thereof that they believe in, both in themselves and others.
If you're wondering why anyone would think this way, my best responses to that are Responses to Tyler Cowen on Rationality and Why Rationality?
If you're interested in going deeper, you should try reading the sequences. You can get the Kindle version here.
I think rationality and the sequences are pretty great. The sequences were created by Eliezer Yudkowsky, in the hopes that those who learned to think well in general would also be able to think well about AI. Whether or not you have any interest in thinking about AI, or thinking about it well, I find it valuable to think about everything well, whenever and to the extent I can.
While I do consider myself a Rationalist, I do not consider myself an Effective Altruist. That is a very different set of norms and cultural constructs.
The Evergreen Posts
These are to me the ten posts most worth reading today, along with a pitch on why you might want to read each of them. Only one is directly about AI, exactly because AI moves so quickly, and my top AI posts are listed in the next section down.
The top ten are in alphabetical order, all are listed again in their appropriate sections.
If you only read one recent post and are here for AI, read OpenAI: The Battle of the Board.
If you only read one fully evergreen older post, read Slack.
An Unexpected Victory: Container Stacking at the Port of Long Beach. This is still highly underappreciated. How did Ryan’s boat ride and Tweetstorm cause a policy change? Could we duplicate this success elsewhere in the future? How?
Asymmetric Justice. A concept I wish more people knew and understood. Many moral and financial systems effectively reward only good effects caused intentionally, while they punish any and all bad effects. This creates a strong bias towards inaction. If fully applied with enough knowledge, no one does anything.
Book Review: Going Infinite. A world tour of everything Sam Bankman-Fried and FTX, starting with the book by Michael Lewis, and going on to explore lots of questions including the foundational problems with Effective Altruism, how Lewis could fail to realize Sam was doing all of the crime, and how the whole incident is a metaphor for accidentally misaligned artificial intelligence with maximalist goals that could kill us all.
How to Bounded Distrust. Important practical advice on how to process incoming media information in an era when you can trust them in some ways, but not in other important ways. This is especially hard because the rules they play by continuously change.
Mazes Sequence Summary. This is a proxy for the Moral Mazes sequence, which is effectively an entire book about what happens in large organizations and to those caught in their webs.
Motive Ambiguity. Another key concept to understand. People care quite a lot about what motives others attribute to them. Often, when faced with choices on how to accomplish a task, people intentionally make the worse choice, so that others know that the central task was the central task motivating them, and they didn’t instead care about something else.
On Car Seats as Contraception. My top post by hits to this day. It turns out we have enough information to model this, and yes car seat requirements do reduce fertility so much that repealing them would be an efficient intervention for that reason alone.
OpenAI: The Battle of the Board. I put together, in real time, a picture of my best guess of what happened at OpenAI in November 2023. There are details we do not know, but reports continue to confirm the core story here, and to contradict mainstream media reports and widely expressed viewpoints on Twitter. It was widely linked as the best explanation of events. I was asked if I had private sources on this. I didn’t quite have zero, but all the key information was public and I put it together. As Michael Samuel once said: My guesswork is better than your intel.
Simulacra Levels Summary. This is the shorter version of the series of posts on the simulacra levels, which I recommend reading in full if you are sufficiently interested. I consider this one of the key concepts I use to think about things.
Slack. Slack is the absence of binding constraints on behavior. Slack is also life. Humans need slack to survive, and they need slack to do anything worthwhile. You need to fight for it, and the fight begins by realizing its importance.
Honorable Mentions: The Dial of Progress, AI: Practical Advice for the Worried, Everybody Knows, More Dakka, In a world… of venture capital, A Guide to Pizza, Something Was Wrong, Repeal the Foreign Dredge Act of 1906 (the last cut), The Long Long Covid Post, Out to Get You, Play in Hard Mode.
AI
The problem with AI is that if we do not solve a bunch of impossibly hard problems on the first try, then we will all die.
The far less serious problem with writing about AI is that this is all happening so quickly it makes your head spin. That also means a lot of what you write is quickly deprecated.
What is still relevant and likely to hold up?
If you’re reading this right away, of course, you can check the recent weekly update.
Other than that, here are the posts I think best stand up in their entirety.
OpenAI: The Battle of the Board. I expect this to be relevant for some time. We will continue to learn details, and the saga will continue, but this is your essential guide to What The Hell Happened At OpenAI.
On the Executive Order. The Executive Order is the most important American government action on AI. This analyzes what it actually says.
The Dial of Progress. I continue to endorse this theory of the more well-meaning of those who oppose on principle any move that could slow down AI development, and dismiss all safety concerns. It all makes me sad.
Gemini 1.0. Until we get our hands on Gemini Ultimate this is what we know about Gemini, one of the two most important LLMs going forward.
On AutoGPT. It has been a while, but many of these thoughts remain relevant.
AI: Practical Advice for the Worried. Pretty much everything here still applies.
Response to Tyler Cowen on Existential Risk and AI. While a lot has happened since, I believe both sides of this argument think their arguments still apply.
On RSPs. Only for those who want to go deep on this question.
On the UK Summit. Seems likely to remain relevant for a while.
On AGI Ruin: A List of Lethalities. Technical questions in detail.
On AI and Interest Rates. Why there is no easy path to riches.
I owe a debt of one (1) post that outlines what I think about existential risk and what we should do about it, in plain language.
There are a lot of sections of weekly posts that are still highly relevant, but someone has to go through them and figure out which ones they are. There’s a lot to go through. Even when reading this week’s post this week, I expect and advise most readers to be selective.
General Rationality and Principles of Life and Thinking
Asymmetric Justice. A concept I wish more people knew and understood. Many moral and financial systems effectively reward only good effects caused intentionally, while they punish any and all bad effects. This creates a strong bias towards inaction.
Blackmail and Privacy. I argue why Blackmail should be illegal, and why Privacy is important, because others are remarkably confused on both these questions.
More Dakka. If it does not solve your problem, perhaps you are not using enough.
Everybody Knows. Do they? A concept I think about a lot.
The Twelve Virtues of Rationality. Full geeking out on rationality theory.
Altruism is Incomplete. My first attempt to explain some of my differences with Effective Altruism, from long ago. There is much to admire, but EA is very much not my culture.
Simple Rules of Law. Some puzzles about how our preferences on rules and laws work.
Zeroing Out. It is much easier, and often more useful or profitable, to find a bias in a complex model and work from there, than it is to create a more accurate model.
Motive Ambiguity. People often do the worse thing intentionally, that lacks benefits or imposes costs, so it is clear they do not care about such other benefits or costs.
Also see the Choices Sequence below, especially the very important Slack. Highest value otherwise are Play in Hard Mode, Out to Get You and Choices Are Really Bad.
World Modeling
The Thing and the Symbolic Representation of The Thing. Which do you seek? Which do other people seek?
In a world… of venture capital. This is still my essential model of much of Silicon Valley, venture capital and many other things. You back something because you expect other backers to expect other backers to back it later on.
How to Bounded Distrust. The media is trying to mislead you, but it mostly follows certain rules. How can you use that to take what they say and extract information?
Leaders of Men. You get the person, action or culture that solves the most valuable problems, even if they fail at other obvious problems. Focus on that which matters and which is scarce.
Paper is True and Rock is Strong. Everybody wants a rock with a simple principle to live by. Will that break down, especially once people notice? Will they be able to wind a string around it? When will it serve them better than actually thinking or considering?
How to Best Use Twitter. Some people need practical advice.
Repeal the Foreign Dredge Act of 1906. Not sure this is the place for it, but yes do that.
On the Cost of Thriving Index. I continue to think that, while the math is not great, the critics are largely missing the point, and that is the most important thing here.
Zvi’s Thoughts on the Survival and Flourishing Fund (SFF). Still relevant for those potentially seeking funding there.
An Unexpected Victory: Container Stacking at the Port of Long Beach. This is still highly underappreciated. How did Ryan’s boat ride and Tweetstorm cause a policy change? One of my all time top posts in its original form on WordPress.
A Guide to Pizza for those who need one, and a less vital one to Restaurants in general.
Oh, Yeah, That Happened
Sadly, FTX. My at the time summary of the events at FTX. Holds up well.
Book Review: Going Infinite. My look back after reading the book by Michael Lewis. A full attempt to solve the puzzle of what happened, why it happened, what systems drove it to happen, how this all relates to EA, and what we should think about all that. Also all the crime, and lots of fun. Highly recommended.
Gaming Fun
You Play to Win the Game. That is the rule.
Who to Root For: 2019 College Football Edition. A complete guide for the time. Who has moved up and down? Michigan moves down recently because cheating. Anyone who disrupted a conference moves down, especially when up against those they stabbed in the back. If you hire a coach you take that on for good and ill. Considerations outside football also start to loom larger recently, shall we say.
A lot of posts in the older archives are about computer games or about Magic: The Gathering. Some of those are great fun, especially if you like TCGs or rogue deckbuilders half as much as I do. Alas, I realized a while ago that when I put out gaming posts, I lose subscribers, and interest was low. So I almost entirely stick to short comments in the monthly roundups now.
Fertility, School and Childhood
These aren’t a sequence, but as a parent all of this is of great interest to me.
Something Was Wrong. Recommended, won’t say more.
The Case Against Education. Covers Bryan Caplan’s excellent book.
Against Active Shooter Drills. Somehow I needed to write this.
The Kids are Not Okay. Why? I do primarily blame phones and social media.
On Car Seats as Contraception. This went highly viral, I think for good reason.
Formula for a Shortage. The shortage of baby formula. Really about economics.
Unlike many other topics, roundups on these subjects remain relevant:
Covid
I did not intend to end up writing weekly Covid posts. That sort of happened.
While I warned my coworkers and friends earlier, I did not write anything in public until March 2, when I wrote Coronavirus is Here, because like everyone else I thought, who am I to say something? Then I realized I was anyone.
I wrote a Purim parallel, and then as my family moved a Passover one, then posts on Initial Viral Load and how I was thinking about R0, and how Seemingly Popular Covid-19 Model is Obvious Nonsense. My confidence grew that writing could help me figure things out, perhaps help others too, and I was going to be spending time thinking about things anyway. An online debate with Robin Hanson inspired me to step things up even more.
On 4/17/20 I did the first weekly Covid post, without realizing what I was doing. Then they kept going.
Now that it is all over, what is worth looking back on? Mostly I think it is time to move on, but here are a few.
Key Mostly Outward-Facing Facts From the Story of VaccinateCA, telling the story of the effort by Patrick McKenzie. Mind blowing.
The Long Long Covid Post, my analysis of Long Covid.
Omicron Variant Post #1: We’re F***ed, It’s Over. A good one to look over to think about what it was right to think at the time, with what level of confidence.
My Current Model (from May 31, 2020). Good place to look back and postmortem.
The Simulacra Levels Sequence
I find the Simulacra Levels to be a vital concept for thinking about the world. Different people and different communications are on different levels. By appreciating the differences, and what level or levels everyone is operating on, you can better understand what is happening, and model what people will do.
One definition that is quick but not precise is these four meanings for ‘there is a lion across the river’:
Level 1: There’s a lion across the river.
Level 2: I don’t want to go (or have other people go) across the river.
Level 3: I’m with the popular kids who are too cool to go across the river.
Level 4: A firm stance against trans-river expansionism focus-grouped well with undecided voters in my constituency.
I wrote a Simulacra Levels Summary for those who don’t want a more lengthy version. Dive as deep as seems worthwhile, I do think this is worth a deep dive. The four children definition is pretty amazing if you get it, but hard to get if you don’t have the Jewish cultural context.
The Moral Mazes Sequence
In December of 2019 I essentially wrote a book about Moral Mazes, based on the book by Robert Jackall, where those that are in large organizations or otherwise disconnected from underlying reality end up abandoning all else to devote all their time and resources to manipulation of the system, seeing such actions as intrinsically good and worthy of reward, and forming a coalition with all other such players against anyone playing the game any differently. I later wrote Mazes Sequence Summary as an introduction, overview or shorter summary.
Later I wrote Motive Ambiguity, which I think is a definite must read. It takes a central concept of the sequence, the reversal of morality in order to make clear what else you value, and explains it as an independent concept. Vital to understanding the world.
I continue to find these concepts important, and recommend reading at least the summary.
(Not part of sequence, but note: Quotes from Moral Mazes)
I would still like at some point to turn this into an actual book. If you are a publisher and would like to do that, please reach out.
The Choices Sequence
This series of posts from 2017 explores many related concepts around choices, having slack and the importance of observing the Sabbath. I still think about and refer back to them often.
Change Is Bad. Almost all changes are bad.
Choices Are Bad. Choices have inherent disutility trading against potential gains.
Complexity Is Bad. All complexity is expensive, keep things simple at their core.
Choices Are Really Bad. Many more costs to having choices. It gets expensive.
Play in Easy Mode and Play in Hard Mode. Choose wisely.
Out to Get You. Some things are Out to Get You. Recognize. Know your options.
Slack. Slack is the absence of binding constraints on behavior. Slack is life.
Bring Back the Sabbath. We need a restful time, free from work and also choice.
Sabbath Commentary. Extended thoughts and tangents on Sabbath questions.
Game Theory, Gambling and Prediction Markets
These three posts sum up my views on prediction markets. Prediction markets are great tools, I especially love Manifold and the fast user-created markets but also Polymarket and its ability to run real money. But you have to do it right, or they won’t work.
If you do bet, here is my post on The Kelly Criterion. Also consider my method of Evaluating Predictions in Hindsight.
A fun sequence of three posts about a cool game I played in college: The Darwin Game, The Darwin Pregame and The Darwin Results.
Completely non-vital but related and fun are Simplified Poker, Simplified Poker Strategy and Simplified Poker Conclusions.
Bonus Content: A Newly Salient Question for 2023
Before we begin, my spoiler-free John Wick 4 review is: This movie lived its best life.
Now that the fourth movie is out, we can return to Laws of John Wick.
Spoils for entire series will follow.
In the second movie we have Rule of Man. In the third we have Rule of Law.
Which is it? John Wick 4 attempts the reconciliation. We have Rule of Law and also Rule of Man. Fealty is central to the world now. Also fealty is part of the Rule of Law.
In particular, like a marker, it is part of contract law. The law is thou shalt keep thy word.
You do not owe fealty because The Table said so. You owe fealty because you swore fealty. You spoke the arc words. Once you swear fealty, you get what was agreed to, but in exchange you are below the table. If you agreed to it, as many do, the table will have an open ended infinite set of markers on you. It will enforce consecration as part of that deal, but it can undo that consecration if you break the deal. Do not forget your fealty. Or, if you decide something else matters more, do it with eyes open, on purpose.
Ultimately, John Wick 4 confirms that The Table keeps its word. It is a system of laws. If you engage in a duel and win the right to have all obligations to the table lifted, or your position reinstated, then that is it. It shall be done. Strictly business. The Table only has power over those who give it that power.
I don't remember how I ended up here (might've been a link in an ACX post, or a random tweet, or an MtG-related post in the pre-Substack period of the blog) but I'm very happy that I did. I've been following since early Covid days because those posts aimed at understanding the situation rather than pushing any sort of agenda. Later I noticed that regardless of the topic, even when I'm completely out of my depth, the posts are never boring and help me understand this world a bit better. Yeah, the AI roundups are massive and I've given up on consuming every post, but that's the nature of that topic (and I don't envy the people who are constantly aware of the X-risk etc).
I guess that was a meandering way to say a huge thank you for your writing and for keeping all of it free.
Btw I do enjoy the MtG/other game posts (my first reaction to your blog was "wait, I remember this name. A Magic pro wrote all of that?") but it makes sense if you want to keep that separate.
Your writing is extremely educational, even for someone like me who starts out knowing nothing. It doesn't make me feel like I'm being forced into a particular set of beliefs.
Re fertility and car seats, this is easily fixed by redesigning cars and car seats.