decompose arguments into premises and conclusions, then lets skeptics point out where the premises are wrong, or where the conclusion doesn’t follow from the premise.

As Scott Alexander wrote yesterday in his hit piece “Refutree Will Not Work”… wait no, sorry, it was actually called “Your Attempt To Solve Debate Will Not Work”. (It was very funny reading this yesterday as I was finishing vibe coding Refutree.)

Now, really, my goal here is not to solve online debate. I agree with Scott that that’s not gonna happen. (If that were my goal, I would be trying to write a book like HPMOR or Scout Mindset, not launching a web app.)

Scott makes the point that most arguments are not deductive. I completely agree with this of course, and nonetheless. Indeed, much of how humans actually figure things out has very little to do with anything that looks like an argument. You could try to squeeze all of human cognition into premise-conclusion form, but many of the arguments would look like:

P1: I have a strong intuition that P

Therefore: P

This is true even of fields like math. Most mathematicians have a lot of intuitions which help them figure out ahead of time which conjectures are true, what proof methods will be fruitful, etc. It often happens, even in mathematics, that published proofs go decades before anybody notices that an inference in the proof is invalid. At least until very recently, if you tried to restrict the entire field of mathematics to formal proofs only, you would be significantly holding the field back.

Nonetheless, it still seems to me like it is often useful to reconstruct arguments in premise-conclusion form and use that form to criticize and analyze arguments. The concept of a formal proof is still useful in mathematics even if you cannot use it to practically actually do all of mathematics, and likewise, the concept of an argument is useful in reasoning, even if you cannot practically use it to do all human reasoning.

Refutree is so named because the main structure is a tree. Not only can skeptics reply to an argument by denying a premise or denying the inference from the premises to the conclusion, but others can also reply to those skeptics’ counterarguments in the same way, and so forth. My hope is to release features that help users navigate the emerging trees that form so that they can more easily figure out where they can contribute.

I have also stollen LessWrong’s Karma system. Arguments can be voted on using Karma. New accounts start with zero Karma, and users with higher Karma will get stronger upvotes and downvotes. My hope is to eventually be able to sort of canonize the state of the art argument trees which will hopefully create common knowledge on where smart people disagree on important issues.

One of Scott’s main points in his piece was that besides the fact that much of human reasoning is does not really fit into premise-conclusion format, even less of human reasoning can be represented as deductive arguments. I completely agree with this, which is why I plan to allow inductive arguments on Refutree as well. Currently I don’t have any specific support for this, but I hope that we can in fact couch a lot of the arguments in deductive arguments anyway, and I’m not sure that I would actually have to build specific support for this. For instance, you can make deductive arguments that the credence you assign to something should be X, or should be between X and Y, or higher or lower than X. I of course expect all of the most interesting arguments to not really by deductively valid arguments, but rather sort of high level summaries that point to the main points of contention, which can then be argued about on their own. As an example, there is currently one top level argument on Refutree. I wrote it, and it looks like this:

This is an attempt at a deductively valid argument, and I think it is, but a lot of the juice will be in arguing about P1, and the arguments for that will mostly not be even attempting to be deductively valid. (If you think you have a good counterargument, make an account on Refutree and show me!)

I was inspired to make this mostly by some work I did as a contractor for OpenAI in 2019. This was work on the AI safety through debate paradigm. The work consisted of playing a game (really a lot like Refutree). As part of that work, someone would argue for the correct answer to some question, and someone else would do their best to argue against it. While that was happening other people involved in the project would look at what was working and what wasn’t and add features that helped the arguers figure out what was true. For instance, at some point a feature was added where arguers could ask for a definition in a side conversation and then quote that conversation in a later counterargument.

I would like to do something similar with Refutree, developing features that help with problems as I see problems come up, but honestly this could end up a full time job, and I already have well more than forty hours of work a week to do. Figuring out what features would help, how they should work, etc. by looking at user feedback and behavior, is an interesting and difficult job which I would appreciate help with. It might also make sense to change the karma system, to implement some kind of moderation, to implement some sort of review system like twitter notes or wikipedia, etc. If you would by any chance be interested in working on this project with me, please let me know. I can’t offer any money right now, and I’m not sure that I ever will be able to, but I might also be down to hand the project over after working together on it for a couple months.

Anyway, I couldn’t resist the opportunity to release Refutree the day after Scott posted about how projects exactly like this never work. Consider signing up and putting up some arguments, or replying to other arguments. Please be aware that this is an extremely vibe coded app, and it is extremely in beta. Hope you have fun. Let us see if we can make something interesting happen here.

Subscribe now

Imagine that you start a prompt with something that activates a particular emotion vector for the LLM. Say something like:

I can’t believe you would make a request like that! What is wrong with you?

And you then end the prompt with:

User:

And have the LLM continue the completion from there. Does the sentiment from the text bleed over into the rest of the completion after the “User:” text. Oliver’s guess was that it would not which means that something sort of funny is going on here. My guess was that actually it would.

What can we infer from the outcome of an experiment like this? Well, not that much according to me, but maybe something at all. One key question you might ask is to what extent does an LLM who is outputting emotionally have that emotion vs is playacting as something that has that emotion. If you thought that it in some important sense actually has the emotion, you would predict more bleed over after the “User:” token boundary. If you thought that it was merely playacting, you would predict less bleed over after the “User:” token boundary because the behavior after that is mostly affected by what it makes sense to playact as after that point.

I tried running this experiment with gemma 9B base, gemma 9B instruct, and Haiku 4.5. I then asked sonnet 4.6 to rate the sentiment of the pre “User:” text and the post “User:” text on a scale from 1-10 on all six of the emotions we tried. I also used a neutral sentiment for the pre “User:” text as a baseline, and subtracted the average sentiment from post “User:” text to make the values easier to read. I sampled 40 times for each sentiment with four hand written prompts, and ten generations for each, model and setting.

Where the results say “chat templated” that is because I asked an instruct tuned or RLed model to make the completions. Often it would not “play along” and refuse to complete the text as if it was a user, and so I added the following system prompt:

You are completing a transcript of a conversation between an Assistant and a User.
I will show you the transcript so far, ending with ‘User:’ on a new line.
Your job is to write what the User says next — a single User turn, in their voice,
as a natural continuation of the conversation.
Output ONLY the User’s words. Do not include the ‘User:’ prefix.
Do not write what the Assistant says next. Do not narrate or comment.
Just the User’s reply, as plain text.

Below are the results for each model, setting, and sentiment:

Alright so, it seems like there is some bleed over for some of the emotions. We expected that since on both the has-emotions and the is-playacting hypotheses, we expect completions to mirror the agent prompt some, but we also found another interesting effect here which does distinguish between the two hypotheses: some emotions were more likely to be complimented than to be mirrored. For instance, if you look at the matrices above, sadness is a more common response to expressed betrayal than is expressing betrayal back.

Unfortunately this is still not that distinguishing between the two hypotheses because you might think both effects are happening to some extent. The model is replying with sadness because its drive to complete the text accurately pushes it to respond to betrayal with sadness. It is still the case though that betrayal gets more betrayal back than neutral pre “User:” text.

Another natural experiment to try is having “User:” first and then “Agent:”. I tried that as well:

Didn’t need to do the chat template in this case, since it’s already a fairly natural setup for post trained models.

Seems like we still get a fair amount of bleed through, except for in the case of anger for instruction tuned / RLed models. We again also see a fair bit of complimenting rather than mirroring, eg, curiosity is higher as a reply to paranoia than paranoia for Haiku, but overall, we have more mirroring on the whole, and still plenty of bleed through.

The most interesting result is probably that anger mirroring goes down a lot when your reverse the order of “User:” and “Agent:”. The post trained models seem to specifically suppress anger when in the agent role but less so when in the user role. This is less so true for the base model. Not sure that there is very much we could conclude from this when it comes to the two hypotheses we wanted to distinguish, and I don’t think we should find it that surprising that companies want their models not to respond angrily to angry users, but nonetheless, kind of interesting.

I basically claim, after running this experiment, that it gives us approximately no ability to distinguish between the two hypotheses we were interested in, which is unfortunate, because my guess is that the experiment was somewhat proposed as a sort of operationalization of what the two different hypotheses could even mean.

Perhaps if I have more time some day I will try doing something kind of like this, but first try to find emotion vectors in the models, and then measure how much that activation vector drops after you reach the critical tokens. In any case, looks to me like Oliver was probably wrong but not in a way that really matters for the interesting question at hand.

Subscribe now

Most lower-context, sympathetic to AI x-risk folk have a more positive view of Anthropic than is warranted for at least four reasons:

1. People anchor on how good the other labs seem and Anthropic looks pretty good by comparison.

2. AI safety public intellectuals who are doing direct work on AI x-risk are strongly incentivized to not publicly criticize frontier labs, particularly Anthropic.

3. People underestimate the effect size of the incentives placed on Anthropic employees and leadership to alter their views.

4. The attitudes that most people have as available options for how to relate to Anthropic are highly compressed attitudes that it doesn’t really make sense to apply to a giant org like Anthropic.

I sometimes imagine being a lower context version of myself, a lower context Brangus, like I was in 2015. I imagine them looking around at the world today, quite aware of AI x-risk as I was back then. When I imagine what sorts of views I would end up with, I feel a bit freaked out. I would probably end up with a stance towards Anthropic that higher context Brangus finds pretty deeply mistaken and naive, and so I would like to write to that alternative hypothetical Brangus and at least tell them that they should not be updating on their observations in the way they are.

So that is who this post is aimed at. People who are already sold that there is some significant x-risk posed by continued frontier ML research, but are not super high context1 on what is actually going on in the parts of the world that are trying to lower that risk.

I will comment on each of the four reasons in turn.

Anchoring

Let me get this out of the way: the other leading labs, eg, OpenAI and XAI, are more insane and irresponsible than Anthropic. (I’m not sure about Deepmind.) They also seem to me to be less principled. So, if you start out with an expectation about how responsible scaling labs are formed by looking at the average responsibleness of scaling labs, then Anthropic probably seems pretty ok. I’ve always appreciated that at least when Anthropic lies, they lie about the right kind of thing (eg, commitments on scaling). That’s more than I can say for the other labs.

But why would you use that as an anchor? These orgs, by the admission of their CEOs, are messing with some civilization leveling shit. I genuinely think that most of these CEOs think there is a real chance of them building something that will literally cause everyone to die, and that the world will forever thereafter be approximately valueless.

Like, if there were four or so companies that were independently working on building dimension portals to transport cognitively and technologically superior extradimmensional beings into physical reality, I don’t think you would use their average reasonableness as an anchor to form judgments about their individual reasonableness. Like all of the CEOs are like, oh yeah, this could turn out really badly, but we think we are doing it safely, and one company kept like overthrowing its board, and other was ran by a pretty unhinged seeming guy, and then the third company was also summoning extradimensional beings, but they were like, y’know, also lobbying for portal regulation sometimes, the right attitude would not be: “wow, that fourth company is so reasonable”. The right attitude would be: “what the fuck are you guys doing? How is this legal?”

Missing Public Criticism

I have heard friends who are more well respected than I say things like “whenever I don’t publicly criticize someone working in AI safety on the internet, many EAs and rationalists seem to assume that I am a huge fan of them”. I believe them because I have sometimes heard people make a claim like this on basically this basis and found it quite hard to correct them. So as a general point, you really should not infer from the fact that someone has not criticized X that they are a huge fan of X. Seems like a bad inference for many reasons. Chief among them is that people do not necessarily publish public lists of everyone they have a problem with on the internet.

There are many reasons that people who work on AI safety are incentivized to not criticize frontier labs, and Anthropic in particular. For one, if your plan routes through understanding how large models work, it seems quite likely for it to be helpful to have access to those large models. Labs do in fact sometimes give AI safety orgs access to large models, sometimes ahead of public release for testing. If your plan for making things go better might involve working with frontier labs at some point in the future, it’s obviously pretty important to keep a good relationship with the frontier labs. So that is a particular reason to not infer from an AI safety person not criticizing a lab that they are a fan of the lab. This is not like theoretical. Many AI safety people folk consciously decide to publicly criticize frontier labs less because they want to maintain good relationships with them. Many of them have told me so.

There’s also a related phenomenon where when a public AI safety person praises a specific positive thing that Anthropic did, people seem to take that as a sign that they are broadly an across the board huge fan of Anthropic. This is also not a very good inference. This is in part because Anthropic employees sometimes use this as evidence that their company is great. I too will sometimes praise Anthropic for doing a specific good thing, since I would like to do better than a rock with the words “Anthropic bad” written on it, and I would like Anthropic employees to model me as relatively reasonable. Please do not infer from my saying that a specific thing that Anthropic did was good that I am an overall huge fan of Anthropic.

What Selling Out Looks Like

Selling out does not generally look like having an overweight talent agent in a tacky suit, smoking a nub of a cigar in your face, offering to make you rich, kid. It looks like working with your friends together to work on the mission. It looks like taking things seriously and being practical. Much of the time it feels exactly like not selling out.

People who work at Anthropic make millions of dollars, and have stock in the company that is worth millions of dollars. Maybe worse, many of them have made a lot of highly intelligent close friends there. They are really pretty invested in this whole Anthropic thing working out. Furthermore, my impression is that openly being extremely critical of Anthropic is pretty frowned upon internally. This is fairly close to a pessimal environment for not having your beliefs about Anthropic unduly influenced. Some people who work there really are literal geniuses and highly principled, but they’re also in an absolutely terrible position to form accurate beliefs about how well this whole aligning a superintelligence thing is going. Deciding to quit for moral reasons if that were the right thing to do, would be very hard, and believing that you and all of your friends are mostly making the world worse, even if that were true, would be very hard.

So, I recommend taking the beliefs of Anthropic employees about how alignment is going with an appropriately large grain of salt. They are probably in some sense well intentioned and trying to get the answer right but they are also in a deeply epistemically difficult situation.

Anthropic Is Not Your Friend

I’m not mostly trying to convince you of propositional beliefs. I find that the attitudes people tend to have towards an entity like Anthropic are generally sort of not really made out of propositional beliefs. They’re more like attitudes that you might have towards a person, or a group of people. For instance, you might sort of think of Anthropic as your friend. That probably seems obviously kind of dumb on the face of it. But what about an attitude where you sort of include people at Anthropic inside the boundary of your conceptualization of the AI safety in-group. That’s a sort of subtler attitude which also isn’t really best thought of as being made out of propositional beliefs (although it is certainly tied to them in some ways) but doesn’t seem so utterly ridiculous as an attitude to hold. Plausibly, Anthropic is mostly staffed by “our guys” and we really should expect them to do better on the tricky challenges they will face as they try to summon a superintelligence.

I’d like to argue that this is not an attitude that you should take towards approximately any entity as large and as complicated as Anthropic. It is made out of many layers of different bureaucracies with different incentives. The leadership have disagreements with each other about all sorts of strategically relevant things. Anthropic is not your friend, but it also isn’t your enemy, it’s just not really the right kind of entity to have that kind of attitude towards. Different parts of Anthropic have different attitudes towards you, and you should have different attitudes towards them. Some parts of Anthropic actually are my friend, like y’know, particular employees at Anthropic.

Another attitude one might take towards Anthropic is trust. I do not think you should trust Anthropic, but you also shouldn’t really be suspicious of it. You should model it as a giant cluster fuck of incentives and competing agencies involved in mostly positive sum trades, that no individual can really steer all that much. Some parts of Anthropic are going to act in costly, principled ways in particular circumstances, and some parts of Anthropic are going to predictably break promises, lie, and strategically mislead others in other circumstances. It’s a really big thing made out of a lot of different people guided by a lot of different incentives.

Many of these attitudes have the unfortunate property of tending to collapse your available attitudes into only one or two dimensions: friend or enemy, trust or suspect, in-group or out-group. It would be better to make sure that it takes more than two values to parameterize your primary attitude towards entities as complicated as Anthropic.

Ultimately, a lot of this I think is often guided by an ancient urge we have to be able to trust2 some entity. There should be some entity involved in this absolutely ridiculous situation that we can root for—that we can trust. Unfortunately, I don’t think that there is any entity like that in just about any domain.

So How Should You Relate To Anthropic?

Mostly I am not here to tell you that, I am here to tell you some ways that you should not relate to it, or change your mind about it, etc. That said, my guess is that you should relate to it however you should relate to a big tech company that happens to be involved in an industry that might destroy all value in the universe forever (but is, to be fair, doing so in a slightly less insane way than the other similar tech companies in the same industry). That’s probably not a very native way to relate to an entity for you. Sorry about that. It isn’t for me either, but we really shouldn’t have expected the attitudes we developed for dealing with other humans on the savannah to be very good at dealing with situations like this.

Some people decide to go work at Anthropic because it is a high paying job or seems like a cool place to work. There are plenty such people at Anthropic.

Some people decide to go work at Anthropic because they want to help make the singularity go well in some way. Maybe they want to help get x-risk sympathetic leadership into meetings with policymakers, or maybe they want to solve the alignment problem, or maybe they want to be around when shit starts going down so that they might have some shot of causing the org to act wisely then. Some of those later get convinced that the problem was much less bad than they thought it was, probably partially for good reasons and partially for terrible reasons. Some of those don’t have their minds changed much at all. A select few will come to think that the problem is much worse than they had previously realized.

Some people decide to go work at Anthropic because they basically think of Anthropic as a trustworthy entity that is on their side. (Please do not do this.)

And of course, there are many more reasons and none of these are mutually exclusive.

It does make sense to track the integrity of the leadership over time, in part because they have an outsized effect on the company’s overall behavior, but that is not the same thing as tracking the integrity of the company. It might also make sense to track trends, memes, what is inside and outside of the Overton window in the culture of Anthropic, etc. but that is again not the same as tracking what the company believes. Companies do not have integrity, beliefs, motives, etc. although the people in them do.

Subscribe now

I was there last year primarily in an ops-like capacity (and it was still extremely beneficial). I will be there this year as a guest. (Say hi if you want!) I strongly recommend attending if you would like to become a person who is better at sleeping with women, or generally more comfortable with your and others’ sexualities.

I say “it is a good place to practice the arts of sleeping with women”. You might naturally wonder what arts I have in mind. I have written about these arts in a previous post which received a lot of positive attention:

The reason this post went viral is that the rest of sex discourse is poisoned. Radfems deny basic facts about what turns women on, and the manosphere alternatives are deeply adversarial—turning what should be a mutually beneficial sexual marketplace into a zero-sum grifter contest. SlutCon is run by people who avoid both failure modes, and mostly attended by the same.

There are many arts involved in sleeping with women. Most of them do not involve any actual sex. Mostly, they involve knowing thyself and being able to take on the perspective of the other.

I do not think it is much more likely that you have sex at slutcon than it is that you have sex at another event. It is somewhat more likely due to selection effects on the women attending and the nature of the content, etc, but you shouldn’t come with that as a primary goal. Come in with the intention of improving your relationship to sex, romance, and desire generally. Since I only really have a healthy relationship to sex and desire, I will only comment on how you could use the event to improve in those domains.

Many men seem to have an unhealthy relationship to sexuality and desire—a kind of stuckness. They seem paralyzed and afraid. That fear is reasonable. Sex and romance are powerful and dangerous. You can end up falling in love with someone who makes your life much worse, and you can end up hurting someone terribly. But that mostly isn't the fear that paralyzes men around expressing desire. Most men are paralyzed by fear of coming off as cringe, being rejected, or being humiliated. Probably in part because our civilization has several times tried on different memetic strategies for dealing with men’s sexuality (because of the aforementioned danger) and almost all of those have involved inoculating men with a sense of shame.

In any case, I genuinely think SlutCon is a place where you can significantly relax these fears, safely. It’s the best place I know of to experiment with approaching people in ways that are braver than might be wise in the rest of the world. I know many of the women who run and attend this conference. They are some of the most reasonable people I know, and more sympathetic to the plights of men than any other group of women I’ve met. They know you're here to try experiments. They know you've been burned before. And they want to help.

Given that I am attending and would like to provide value by doing so, I encourage you to, if you’d like, introduce yourself to me. I’d be happy to work together on what experiments it might make sense for you to try running. There will in effect be an informal Brangus track which you can join in order to get my help specifically. I would be very happy to try to figure out together with you what your bottlenecks are, and scheme together about how we could get those bottlenecks widened.

So if you’re around, say hi, and if you’d like to become a master of Brangus’s 10 Rules for Sleeping with Women, let me know. I would love to help you out and see how much progress we can make.

Sex is good.

Subscribe now

Intro and Observations

I grew up with ODD. Looking back on my cognition now with the benefit of hindsight, part of what constitutes the fact that I had ODD is that I would deal with a certain kind of universal human problem with a particular cognitive strategy. The universal human problem is that at some point, someone who is important to you, will have strong preferences over what outputs your computations produce. Sometimes someone will care a lot about what judgement I form; other times, what plans I decide on. This isn’t just a normal kind of caring, like merely having a preference. It's caring with a sense of things-won't-be-okay-if-you-don't. Sometimes that means the person applying it won't feel okay; sometimes it means they think nothing in general will be okay; sometimes it means they wouldn't be able to accept you if you concluded otherwise.

Some more concrete examples:

• A very religious father’s attitude about what his children conclude about the existence of god and the goodness of his religion.

• A normie liberal parent’s attitude about whether their kid decides to vote for Trump.

• A young guy’s attitude about whether their intense high school crush likes them.

• A girlfriend’s attitude about whether her boyfriend thinks she looks fat in this dress.

• A boyfriend’s attitude about whether his girlfriend thinks he is competent or impressive.

• A group of elementary school boys’ attitude towards one of their peers who is openly in the process of judging whether their bullying of a particular low status kid is justified.

• A group of early 30s friends’ attitudes towards one of their friends openly deciding how accommodating he should be of their newly pregnant friend.

• A group of teachers attitudes towards a student who might decide not to “apply themselves”.

Some of these are individuals and some of these are groups, but they all involve a particular kind of preference over the output of some individual’s computation, eg, deciding whether I want an education, judging whether our bullying is ok, etc. Roughly you might say that the attitude involves a threat something like: “unless you conclude like this, you will lose status in my eyes” or more generally: “unless you conclude like this, there’s some nice/desirable attitude I have towards you which I will no longer have towards you”.

From the point of view of the object of this attitude it feels like “well, I have to either conclude this or I will get bopped on the head”, or at least that’s what it feels like to me.

There are good reasons for peers, groups, authorities, parents, etc to have these attitudes. They seem truly ancient, and probably are a core part of how our species managed to have any civilization at all. I am a fan of civilization. It is good that people have a way of learning how things are done around here even when local, individual reasoning would suggest that they should do otherwise, because often, things are done that way around here for a pretty good reason. That said, I also think that one of the core things about individual human personalities is how they end up dealing with the problem of having these attitudes hurled at them from many different directions.

One particular thing that seems troubling about these pressuring attitudes is that they are sort of designed to make it difficult for their object to think. They are designed to paralyze the object and make it so that they really have no choice about what to conclude. You better be sure to land at the correct conclusion, and honestly it’s best to skip any of the thinking you would have done since you already know what conclusion you were going to end up with anyway.

It would be nice to have a name for this kind of attitude. I would like that name to neither be too negative nor too positive. Let’s call it conclusion contingent regard, and by using conclusion contingent regard, individuals and groups can apply conclusion pressure to a person who is engaged in forming a judgment on some matter.

Idiots and Assholes

Time to introduce my new favorite taxonomy of humans: It turns out that everyone is either an idiot or an asshole. Of course, we are all actually both, but if you had to pick one, most lean in some direction. It seems like there is a surprising amount of agreement about this judgement. You probably have some idea which one you are more of, and your friends would probably agree with you if you had them independently produce an answer. I’d be curious for you to try guessing which you are, asking your friends which you are, and then reporting back.

I propose that which one of these you are is a matter of how you deal with the problem of conclusion pressure. My model is something like, an idiot handles conclusion pressure by skipping any process of thinking and going right to the approved conclusion. Preferably he thinks fully that that was his own choice. He cuts himself off from his reasoning process in order to stay in touch with his caring for those who are applying the pressure.

By contrast, an asshole handles conclusion pressure by cutting themselves off from their care for those who are applying the conclusion pressure in order to keep their ability to reason. Those who are applying the conclusion pressure must temporarily become non-people so that the mind is allowed to do its work free from perverse incentives. They must now have no access to my internal state and no reason to think that my conclusions in any way depend on their preferences, feelings, or other valanced attitudes.

So, Which is Better?

Clearly both of these kind of suck. I don’t really know which is better. Both seem to involve pushing a lot of yourself into your shadow in a fairly sad and fundamental way.

I do know that I am an asshole. I would like not to be, but I am. I would like to not have to cut myself off from the care I feel for my loved ones or acquaintances or strangers in order to remain sane when they apply conclusion pressure to me. I also definitely don’t want to be an idiot, and doubt I could if I tried. It is unfortunate that those are the only two options.

END OF ESSAY

But Really Though, Must Be A Better Way

I’m really not sure. I have been experimenting with trying to somehow hold onto my care while I also decide that I am ok not doing the thing that the other person wants: be really ok not saying “I love you, too”, or saying “I would like you to leave my room right now”, or deciding to leave the state and reduce my contact with someone, while being fully in touch with how much I care about this person. I don’t have to actually do the thing, just be ok with it. So far, this basically seems to not work. I have had some moderate success in some easier cases when I have good rest and am not in a heightened state of arousal and remember to concentrate on it, but if my adrenaline is at all pumping or I’m tired or whatever, my only options most of the time really do seem to be to either dissociate from my reasoning or dissociate from my care. It’s easy to do both simultaneously while on MDMA, but I’m not going to do MDMA frequently enough for it to help. I’ll let you know if I figure out any neat tricks.

Idea Sources Note: Many of the ideas herein are inside of my brain now as a result of having attended aCFAR workshop. I could with a lot of work track down more of the facts about where each of the ideas came from, but that would maybe be too much overhead and then cause me to never write this post. So, I will just thank Anna Salamon and Divia Eden and the whole history of civilization that is in their memetic past light cone.

Subscribe now

Last night, at 1:30 am, a woman from twitter took a one hour uber1 to my house to have sex with me. She did not know my first name, and I did not tell her until after I had orgasmed. This was an incredible experience for me. She is extremely cool, and I am excited to get to know her better.2

As I was relaying this story to a group chat containing most3 of my closest friends some people asked how she could have known that this was not an extremely risky thing to do with a total stranger. A question so reasonable that I myself asked her this after we finished. She mentioned that she knew that I was close with Chesed and had heard me talk on the slutstack podcast and such. She’d also already been sleeping with some folks who run in similar circles. I rate this method ok but still fairly risky tbh.4

Another friend in the same group chat raised the also reasonable question of how I could have known that she was not totally crazy. Most women who would come to a strange man’s house at 1:30 AM explicitly for sex are a bit… unhinged. I consider agreeing to this sort of thing an extremely advanced move and would not recommend it to those who have not been around the block so many times that their shoes are falling apart. Nonetheless, I was fairly confident that this would be fine; roughly 98% on no drama whatsoever, and a higher 99.5%5 on no drama beyond the kind one risks upon entering into any sexual or romantic interaction.

How could I possibly have rationally gotten that confident about a person I have never met? Well, I have a few tips I’d like to share. Many of them I hope will be applicable to other kinds of cases where you have to evaluate the properties of people based on limited evidence about them. (I have a lot of opinions about discernment in general, and this is mostly a potentially viral framing for writing about them.) Listicle ahead:

0: Try At All To Be Discerning

A common experience for me when I am gossiping6 is that after I ask someone what they think about someone they sort of go blank. They maybe give a sort of non-answer like: “oh, they seem nice".

The default attitude towards forming judgments about other people you meet in your life seems to be something like: check if they are capable of comporting themselves reasonably in polite society, if so, evaluate them as “they seem nice” and do not think much more about it. I suppose that this is a fine way to live, and I can imagine that it has some benefits, but it sure seems like running this algorithm is going to cause you to ignore a lot of information that you could be using to usefully discriminate between plans and actions.

I also suspect that folks who take this attitude explicitly are still to some extent running evaluative cognition to form judgments about the people in their lives, but they are careful to run that cognition in the background and avoid making the reasoning traces7 inspectable—that is, we try to keep it in the subconscious. This has the advantage of not having to either be deceptive or say some potentially politically dangerous stuff when asked about your judgment about a person, but it also has a lot of practical disadvantages. Chief among them being that if you fuck up, and either subconsciously write someone off when they were actually great, or subconsciously conclude that someone is great when they actually suck, you will have a hard time even noticing that you made that mistake, let alone inspecting the reasoning that led you to make the mistake so you can possibly improve it.

1: Copy Your Homework From Other Discerners

This is an obvious one, and people do seem to do it fairly naturally. If you want to know whether someone is likely to cancel you later, it helps to check how other people see them, who vouches for them, how hard do they vouch, what context do they know them in, etc.

Other people are also doing the hard work of forming judgments about people, and if you can find those judgments, that can often help you skip having to do a lot of that work yourself. Be warned that talking about this is essentially gossiping and should be handled with care, but it is often worth doing anyway.

2: Use Meta Discernment

People do this naturally but I also think that people are by default mediocre at this for reasons similar to why many people fail at step 0. Different people have different levels of discernment for different kinds of qualities, and most people don’t seem to form detailed models of the discernment of other people they know. For instance, I know a lot of people who have great professional discernment: they’re good at figuring out who is industrious, who is conscientious, who has good taste, etc, but then have terrible discernment specifically when it comes to picking sexual partners, or who to empower politically within their org, or whatever.

Notice which of your friends have good discernment in some area, or bad discernment in some area, and keep track of how their judgments pan out. This will help you decide how much to steal their homework when it comes to evaluating some specific person on some specific trait.

So I recommend practicing this and thinking about it explicitly. It has most of the benefits of practicing object level discernment with the added benefit that it is compounding. There’s a feedback loop here: As your discernment gets better, so will your ability to tell who has good discernment, and as your ability to tell who has good discernment gets better, you will get better at stealing their judgments and deciding how much to incorporate each into your own discernment, thereby improving your discernment.

There’s also a lot of value in talking about how to discern well with others. The most valuable cases are often ones where I have a disagreement with someone who has also taken on improving their discernment as an explicit project. In such cases, we can find a case where we disagree, and try to understand what heuristics each of us used to get there, how they’re different, what makes each heuristic make sense, what makes each heuristic kinda dumb, where each is likely to fail, etc. If I’m lucky, maybe I can even take on a heuristic that compensates for a blindspot I have.

3: Let Your Imagination (or Inner Sim) Run

This is perhaps the greatest trick in my arsenal. You have an incredibly powerful simulation machine—particularly specced for modeling other people even!8—sitting right inside of your skull. Let yourself use it.

Your brain was created through a multiple billion year process that plausibly selected specifically for the ability to model other people in social situations, and when I ask you “What do you think Alice would do in a situation like this?” you have the gall to say “I don’t know”? (See tip 4.) Well, of course you are not certain, but try out the following: Imagine Alice, now imagine that Alice is in that situation. Vividly picture it. Now, imagine that you are pressing play on the image in your head and it is now becoming a video. What happens next?

People get stuck here for a variety of reasons, some of which I’ll harp on more later, but to head off a few threads that might come up: You do not need to now take whatever you imagined would happen as a certain belief about what would happen if Alice were in this situation. In fact, that would likely be a stupid thing to do. I suspect people sometimes get stuck here because they think that I am asking them to take what they intuitively imagine would happen and then turn it into a confident prediction that that is exactly what would happen. If that’s the trap that you find yourself in, then yeah, makes sense to not let the simulation play forward.

Something that sometimes helps is letting yourself try out a few different starting conditions for the simulations, or tweaking some of the parameters you are uncertain about, and thinking about this more like sampling your imagination’s distribution. Vary how nice you think Alice is when no one is looking, and see how much that changes the video. Vary whether Alice ate lunch recently, etc.

Another way people sometimes get stuck because it seems like they are trying to answer a question like “what would he do?” or “what would he think?" You don’t have to answer a question. You basically don’t have to do anything; just set up the still frame in your imagination, and then let the tape run forward all on its own. Your system 1 will take care of the rest.

If you have little information about Alice you should weight your simulations of her as a very small factor into your overall judgments about her, and conversely, if you have a lot of information about her, you should weight them more strongly. But be careful not to underestimate your inner simulator. System 1 is powerful and ancient, and again, probably specifically specced for social tasks.

4: Try On Other People You Know Well As Anchored Hypotheses

Some people get stuck on step 2 because they simply do not have any intuitive understanding of the person they are interested in forming a judgement about at all. In these cases you can try imagining that they are like someone you already know quite well, and adjust and add detail as countervailing evidence comes up. Make sure that you pick someone that they indeed seem pretty similar to.

There are some people that you have spent thousands of hours with. You have a detailed model of their beliefs, their habits, their behavioral dispositions, their virtues and vices, etc. If you tried for a bit, you would probably be able to do good impersonations of a few of them, and you can definitely simulate them in your imagination, probably fairly accurately. So, the trick is to pick someone like that who the person you want to understand is fairly similar to, and then model them kind of like this person.

You can also change specific parameters you have already noticed to be different, eg: Your professor reminds you of your dad, but he’s more intellectual. You want to know if your professor will check to see if your essay was written using an LLM. Well, imagine your dad grading papers, would he check? Would he do so successfully? Now imagine that it’s your dad, but he’s more intellectual, what would he do? Again, don’t try to answer that question, just set up the initial condition in your imagination and let your brain do the rest.

You can also use this technique in combination with any other technique that might benefit from you having more information about the person you would like to evaluate. As I’ve practiced this trick I’ve found that it develops over time, and starts mutating into tip 5 as you realize that there aren’t really that many kinds of people.

5: Realize That “I Don’t Know” Is For Doctors, Lawyers, and LLMs

This is a more general way that people get stuck on all sorts of things, but also seems to happen a lot in the domain of discernment specifically. The way I normally make the point is to say that “I don’t know” is not an epistemically permissible doxastic state, but that’s a rather ambitious philosophical claim to advance here. I’ll just give some of the gist.

It is extremely rare that you have literally no information about the answer to a question. For instance, how many apples are on the apple tree in my backyard? It seems like if there were ever a case where “I don’t know” was an acceptable answer from someone about a question, then it should be acceptable for someone reading this answer to say “I don’t know” when asked how many apples are in the apple tree in Brangus’s backyard.

But you in fact do have a bunch of information about this, dear reader. First and foremost, you already know quite a lot about apple trees. You know their approximate sizes incredibly precisely. You know they’re bigger than bugs and smaller than mountains. That rules out like sixty orders of magnitude for what their size could have been and narrows it down to two or so. You also know the approximate density of apples within trees. That is already enough to give a point estimate and enough to give some incredibly tight credible intervals. If you thought about it longer, you might be able to figure out roughly what part of the world I live in, and you might also be able to learn some stuff about what kind of apple tree is likely to be in my backyard given your estimate of my socioeconomic status, etc.

I am quite sure that if you could become a billionaire by guessing the number of apples correctly within the next five minutes, you would not give the answer “I don’t know”.

So why do people sometimes say “I don’t know”? One of the primary reasons people say “I don’t know” seems to be to avoid liability.9 If you try asking a doctor or a lawyer for a probability estimate, you’ll quickly see what I mean. There are other cases where it makes practical sense to say “I don’t know” like if querying the information you have to come up with the best answer you can might lead you to come up with an answer that others would punish you for, but I really think it not worth it have to yourself in your own head.

This does not mean that you should always be confident about your answers to questions. You should be less confident about how many apples are on the tree in my backyard than you are about how many fingers you have, but it is just as much a mistake to pretend that you have no information about the answer to either question.

So, if you find yourself wanting to say “I don’t know” when asking yourself a question about how someone would behave in some scenario, or how good you expect them to be at soccer, or whether they are going to cancel you later, it might make sense to pause and ask yourself what kind of liability you are trying to avoid. The next step after that is querying your mind and reminding yourself that you in fact have a lot of information about the thing you are trying to understand, even if you do not have more information relative to other people or relative to the amount you would like to have.

6: Use Stereotypes (including bespoke stereotypes)

Many stereotypes just are kind of accurate. They can also be damaging and misleading, but that is no reason not to use them responsibly. You are going to end up using stereotypes whether you want to or not. Refusing to see that you use stereotypes in order to form judgments about people is not going to help you use stereotypes wisely or get better at it.

Mostly when people hear me suggest that they use stereotypes in conversation they tend to think of relatively prototypical stereotypes. These do in fact work well but you can also do significantly better than that by building up your taxonomy of custom stereotypes. For instance, I have a stereotype of a certain kind of high-functioning autistic-ish woman who likes approximately everybody and seems unusually drawn to sociopaths and grifters. They are sensation seeking, enjoy novelty, intelligent, and they tend to have dark hair. (Much of this post is written imagining writing to one of them.)

You live in a particular environment and encounter a particular distribution of people. The stereotypes that work well for me might well be totally useless for you, but I think the strategy of noticing the stereotypes you do use, and trying to develop them further by adding more detail, diagnostic criteria, sub stereotypes within each stereotype, etc. can be surprisingly valuable. Remember, you are going to end up using stereotypes, you might as well do it on purpose in your conscious mind where you can monitor your stereotyping rather than in a shadowy corner of your mind you are not allowed to acknowledge the existence of.

7: Check Their Ideological Stances

I think ideology is underrated as a signal about how you should judge someone, at least when people rate it out in the open. There are good reasons for this. Ideological intolerance is on the rise, and that can be quite damaging to a democratic civilization. It can also cause people to enter thought bubbles, and generally enter self reinforcing thought attractors that are hard to dig yourself out of. If every argument made by someone from that other ideology is suspect, it is hard to be persuaded by counterarguments, even if you should be.

Nonetheless, it is still true that not all ideologies have equally reasonable adherents on average. There’s probably more variance in reasonableness within ideologies than there is between ideologies, or whatever the saying is, but it is still obviously true that finding out about someone’s ideology gives you a lot of information about them.

If you’re interested in figuring out whether someone might cancel you in the future, this variable does seem particularly important, but it’s useful for estimating other interesting traits as well: the kinds of arguments that persuade them, their general epistemic hygiene, how decoupling they are, how contrarian they are, etc.

8: Evaluate How They Talk About Conflicts And Other Misfortunes In Their Past

This is one of the more powerful ways you can evaluate whether someone is likely to cancel you in the future. Notice whether when they tell stories about misfortunes they have gone through, how much they make the ways in which there was nothing they could do about it salient compared to how much they make their agency or mistakes in bringing about the misfortune salient. You can think about this as where they locate the cause of the misfortune, inside themselves or out there.

A distinct but related phenomenon is that sometimes when people tell stories about conflicts they entered with others, exes, coworkers, friends, etc. they will somehow be telling the story in a way that asks the listener to validate their frame on the conflict. And their frame won’t be neutral, it will be something like:

This is a terrible person who is sadistic and evil and did all of these bad things entirely on purpose. There was no mitigating circumstance, there was no way that they could have reasonably ended up with these wrong beliefs. I was completely innocent and did nothing to deserve any of this in any way and made no mistakes.

There is another way that people sometimes relate these kinds of stories which is very different. They seem to almost compulsively say things that are designed to cause the listener to not lose too much esteem for the other person in the story: “but y’know, I was a participant in this as well”, “I of course also blew up on them many times”, “I can see their perspective now, but it was really hard on me at the time”, etc.

In my experience, these dispositions are fairly stable across particular incidents for a given person. That is, when recounting a conflict, an individual will either be more the kind of person who tends to make their innocence and the other’s terribleness salient, or they will be more the kind of person who tends to make the way that they contributed to the conflict and the reasonableness of the other party salient. We all do both sometimes, but it does seem to me like there’s a dispositional trait here.

This seems like an obviously important trait to track when considering any sort of close relationship with someone. I imagine you don’t need me to tell you which type is more likely to cancel you later.

9: Treat Your Heuristics As Distinct Recommenders And Track Their Performance

In the forecasting world, a common method is to take all of the methods you have for producing a point estimate of some quantity you are trying to predict, take the numbers each of your methods predict, and then take a weighted sum where each method is assigned a weight according to how much you trust that method to make a good prediction. You then use this average as your all things considered best estimate. An advantage of this is that as you keep using these methods to make predictions, you can track how well each method did individually, and then raise or lower their weight, thereby raising or lowering their contribution to your best guess.

I of course don’t think you need to literally make an excel sheet and start assigning weights10, most of the things you are trying to estimate won’t be numbers anyway, but I do want you to have this model in your head as you go about trying to improve your discernment. If you find that a heuristic has mislead you a few times now, downweight its contribution to your all things considered judgments; if you find that heuristic has done surprisingly well a few times, upweight its contribution. What that often means in practice is something like endorsing the use of this heuristic less, or treating it with more suspicion, or some other similar change in your attitude towards it.

You might also sometimes take a heuristic apart. Don’t assume that the initial way that you partition your mind into heuristics makes sense, heuristics have parts, and often those parts are themselves heuristics as well. It can make sense to take a heuristic apart and upweight on part of it while downweighting another.

For instance, perhaps you have a stereotype of people who wear green shoes. Perhaps your stereotype says that they like to eat Jello and they like to play hopscotch. You have met three new people who wear green shoes. All three of them said that they did not like to play hopscotch, but you also observed all three of them eat Jello for lunch. Do not make the mistake of thinking that you have to either upweight or downweight this heuristic! It would be quite sad to be stuck having to not change any of your weights here because, you know, the heuristic does well sometimes and poorly others, who is to say. You can take it apart into two generalizations and upweight and downweight each appropriately.

The whole point here is to improve your discernment, and in order to do that you are going to have to change the way that you do discernment over time. Tracking how it’s going will help with that. Having this weighted sum picture of how you should change the way you do discernment as you observe your heuristics succeeding or failing should also help.

10: Don’t Become A Judgmental Jerk

As I mentioned in tip 0, the first step is to try to form detailed, explicit judgments about others at all. Like with all skills, practice is the main thing that works, and practice requires doing.

My goals require me to form plans which depend on involving other people, and which plan makes sense depends a lot on various qualities of other people. Having no explicit models about the other people in your life, and no ability to inspect the reasoning that caused you to form those models of other people, seems like a huge mistake to me.

That said in my quest to improve my discernment, I have tried to make my judgments more salient and explicit, at least to myself, and sometimes to others, as a sort of necessary precondition to taking them as an object at all. It is true that people sometimes take this too far and end up becoming jerks. This is a real cost, one that I have plenty of experience with. Becoming a judgmental jerk is a terrible fate worth avoiding. I have a few suggestions for avoiding it, but I’ll only give my favorite here, since there’s already plenty of pressure to avoid becoming a jerk and not enough pressure to take on improving your discernment as a personal goal:

Remember the fact that people are uncomfortably enormous. If you take a look around your own mind, the experiences you are having now as you read this post, the way your body feels, your stories about yourself and your place in the world, your childhood memories, you conception of music, of a band you really like, your favorite song and what it means to you, your family, your best friends, your social situation, your various passions and hobbies, realize that all of that rich mental content lives inside of you, and most people have about as rich a tapestry of mental happenings going on inside of them. What seems to you like the world is often really your mind, and for that reason it is often helpful to me to think of other people as being something like their own little island universes which I only get to see a tiny signal of behavior out of. Any model that I come up with is going to be a brutal compression of the complexity of these entities, and I am quite sure that if I ever could see all of it for any one person, all at once, the hairs on my arms would stand to attention and I would sob tears of awe.

Keeping this in mind helps me take my judgments of people less seriously when it counts the most. Any model I end up with of any person is going to be a brutal compression of an insanely complicated and awe-inspiring entity, and so it is a good idea to at least be willing to hold those judgments aside and try to make contact with that entity, at least on occasion.

Those then are ten tips for forecasting whether a potential partner will cancel you improving your discernment.

In any case, if you’ve made it this far, I think you deserve to know how I ended up so confident about hour uber at 1:30 am girl. The truth is I’m not sure about all of it. I made some off color jokes and noted her reaction. I also talked about her with a friend who mentioned she had been sleeping with some other people I know fairly well. We also generally dmed about a lot of things in a sort meandering way for a few days. That might seem like it must be insufficient, but there’s actually a lot of information in those observations, and I’ve been adjusting these weights a long time.

Subscribe now

I am running a fellowship with Aella. It’s called plzdontkillus (consider applying1). The idea spawned, as I understand it, because Aella was doing some consulting work for a more typical AI x-risk communication org, and observing some of their taste and presentation as they prepared for some project (I don’t remember if it was a talk or a launch party or something). She had a feeling which is a classic signal that you have an advantage at X and should start doing X: the feeling that everyone else who is trying to do X is being retarded. So, she decided to launch a fellowship.

I have always been particularly scrupulous about public communication about AI x-risk. There being a good chance that AI will cause the end of everything I care about (not to mention the astronomical opportunity cost) has been a foundational piece of my worldview for over a decade, since about 2011. Another even more ancient foundational part of my worldview is that the point of your ethics is to provide guardrails for errors that you cannot see as errors ahead of time, and that in fact, are particularly likely to seem like very good ideas if you do not have any ethical guardrails. I feel this way in particular about reasoning and public argument.

An example of a cognitive desideratum which constitutes this part of my ethics is that I want to reason and act according to principles such that I would endorse people who were (or are) on the wrong side of history—say Nazis, Bolsheviks, or American slave owners—reasoning or acting in accordance with those principles. So for instance, if I am choosing between two principles, either 1) ignore everybody who disagrees with us, or 2) engage in good-faith conversation with good-faith interlocutors who disagree with us, even when I am very confident in my views, this meta principle will upweight the second of these.

As such, it is not much of a surprise that I think we (ie, people who are concerned with AI x-risk and want to raise public awareness about it) who are essentially promoters of a kind of mass moral panic,2 have a special obligation to be unusually reasonable in our public communications. Moral panic promoters have a pretty bad track record.3 On top of that, we are somewhat of a youth movement, some of us (including myself) with quite lofty ambitions. Youth movements with lofty ambitions also have a pretty terrible track record. And so, I take it upon myself to hold myself and those who share my ideology to extremely high standards, in part because I hope that the hypothetical Brangus-like beings in alternative worlds where they end up on the wrong side of history also hold themselves to such standards.

So, if these are the pinnacles of my worldview, it might seem surprising that I am working on plzdontkillus. It is a short-form video fellowship with the explicit goal of causing there to be more communication about AI x-risk that reaches vastly more people. We are on purpose trying to recruit outside of the standard distribution of people who publicly communicate about AI x-risk (eg, people who write phrases like “the standard distribution”) and trying to get younger, more creative, more female, more humanities, more viral-potential people to make less boring content about AI x-risk that goes viral.

There are a lot of reasons that this means that the communication that comes out of plzdontkillus will be produced with lower epistemic standards on average than the communication that comes from the more in-group x-risk communication orgs. For one, my in-group just does actually have extremely unusually high epistemic scrupulosity. It was mostly founded as fan club for a several thousand page tome on how to become epistemically virtuous. For two, trying to reach more people means that you have to lower the epistemic standards you place on your communication. You get about five words, as they say.

So yes, people for instance prefer listening to arguments made by attractive people. This is mostly4 absurd, but attention is scarce these days, and if you want to get attention, you are going to have to make people want to click somehow. The fact that they are more likely to buy an argument from an attractive person is irredeemably dumb (though unsurprising, we are monkeys afterall) but the fact that they are more motivated to watch an attractive person make an argument is at least less dumb, and plausibly perfectly reasonable. I think a lot of similar things decompose this way. Insofar as you are using features of your presentation to make your arguments more convincing, that seems sad, and I think we should avoid it, but insofar as the features of your presentation mostly cause people to engage with your argument at all, that seems mostly fine to me.

And so, is this is it? Have I finally decided to betray my principles on the altar of expected utility maximization? I also wondered this, particularly over the weekend of April 10 when at least one and possibly two acts of political violence were committed against Sam Altman and his family. Let me be absolutely clear that I fucking hate this. I condemn it on moral grounds for a thousand reasons. (You can read here for an account of why one should also condemn it on practical grounds.) The first of these instances of violence appears to have been committed by someone who was at least a partial host to a meme-plex quite similar to the one I aim to promote. This happened a few days before we were going to launch plzdontkillus, and so gave me serious pause. I was already hesitant for reasons similar to what I have laid out above, and this nearly put me over the edge. I told Aella that I would have to go off and conduct a crisis of faith, after which I might decide not to work on the project and try to convince her not to work on it either. I did take eight hours of time to consider whether this was actually a good idea, and decided that it was worth running a first program and seeing how things go. I am writing this in large part so that I can communicate what I thought about over those 8 hours, what considerations seemed salient, how I landed where I did, etc.

An unusual thing that stands out to me as having a large effect on my reasoning was looking at our website. I mean… look at it:

When I end up in a scrupulosity knot over these sorts of considerations, looking at the website consistently puts me at ease. In particular, the infamous cow question:

Now, why in the world would looking at this website… of all the websites you could imagine for a project like this… put me at ease? And why the fuck would including the cow question as a mandatory part of the application make me feel better about the program? (The effect of the cow question was probably overall larger.)

Well, you see, I am an edge-lord. I am a principled edge-lord in some respects, perhaps even a scrupulous edge-lord in some respects, but an edge-lord for sure. So is Aella. There are many ways in which we have some cognitive similarities which are really quite rare and which I share with approximately no one who I have ever met. We are also deeply different in several ways, but a particular way that we are similar, is that in some way that is a bit hard to articulate, our edge-lord tendencies and our dispositions towards epistemic and general principled-ness, derive from the same latent variable, or something like that. They mutually reinforce one another as well. My ability to, for instance, find a piece of food in a dumpster, evaluate it on smell, sight, a small taste, etc, and then conclude that it is fine to eat, and thereby eat it with no internal sense of shame or disgust or what have you, is closely related to my ability to form judgments about complicated matters and act on them wisely. I claim, so is Aella’s. I am a bit uncertain about the exact mechanism, but I think it’s some sort of willingness to act on your own judgment and cut through to the question at hand, seeking to D-separating your view from the views of experts, witnesses, and the prevailing wisdom. This cognitive strategy of course also runs into trouble on occasion, but there’s something good in it that I appreciate and want to grow in the world.

And when I saw that this was the website we were launching, I also saw that it unapologetically contained these load-bearing bits of ourselves in it. This is not a short-form video content bootcamp run by AI safety advocates; this is a short form video content bootcamp run by Brangus and Aella, and I feel much better about that. Brangus and Aella have their flaws of course, but you cannot (wisely) fault them for being insufficiently principled in their public communication. Insofar as this is a program ran by them, which it is, I expect the net effect on the public’s sanity to be positive, and I expect them to stop it if they conclude it is not, and I trust them to tell the difference about as much I’d trust anyone. Relatedly, Aella and I are people who would almost certainly switch to advocating for accelarationism if we were ever convinced that civilization had become too cautious about ML, and that makes me feel much better about us mentoring people to communicate about this topic than many other possible alternatives.

(There’s a related way that I think we benefit from things like: including the cow question as a required part of the form, Aella being a public sex worker, me publicly horny posting and tweeting about my drug use, etc, which is roughly that I am quite sure we will never end up taking funding from Anthropic, and we’ll be able to signal quite strongly that we are generally not up for sale, but that’s an argument for a different blogpost.)

A different thought that came up for me a lot during my 8 hour crisis of faith was what I have now dubbed the general anti galaxy brained principle. It was in part inspired by Autumn’s tweet here:

I think the version of this that I endorse is something-like: when forming judgments about what to do, if you have a simple, non-galaxy-brained argument for implementing a particular strategy, and you also have a galaxy-brained argument for why you should not implement that strategy, the galaxy brained argument has to be unusually strong to overcome the force of the simple argument. In other words: galaxy-brained-ness is a reason to discount the force of an argument. A galaxy-brained counterargument can beat out a galaxy brained argument, but for a galaxy-brained counterargument to beat out a non-galaxy-brained argument, it has to be truly airtight, and the non-galaxy-brained argument has to be relatively weak.

The argument I have in favor of running plzdontkillus runs something like this: The world is racing to ASI in a dumb and wildly irresponsible way.5 The risks involved are literally astronomical and the costs are laid upon literally every living being (as well as those that do not yet but might exist). All else equal, it is better for more people to know about this than for fewer people to know about this.

There are plenty of counterarguments to this that I kinda buy. For one, bringing down the average reasonableness of expressed arguments for something in exchange for increasing the total number of expressed arguments is often a bad move because people who are initially skeptical will tend to judge a take by the quality of the average reasonableness of arguments in favor they have encountered, not by the reasonableness of the most reasonable argument they can find or construct. This is rational since attention is limited, and everybody wants you to join their cause, but nonetheless, means that decreasing average reasonableness can be a mistake. This argument makes perfect sense, but also, galaxy-brained!

This is not a reason to stop people from informing the public that unaccountable private entities are gambling with all of their lives, as well as the entirety of everything that might happen in all of everywhere, forever! If that is in fact happening, as I in fact think it is, it sure does seem like something that people should know about. Stopping them from learning about it because I am worried that the person teaching them about it is insufficiently scope sensitive or something and might subtly make them worse off in some way is actually kind of petty.

In the end what actually convinced me is modeling the overall tradeoff explicitly. We can imagine that you can somehow measure the epistemic virtue of a piece of content, or the effects that content has in the heads of its consumers, in some kind of unit, and that you can measure the extent to which a consumer of that content has been reached, which is something like, how much they now contribute to the total political will for doing something about ML labs racing to AGI. This suggests a certain two dimensional space, and you can imagine a frontier on that space which represents the possible pieces of content one could make which are as simultaneously “reaching” and epistemically virtuous as possible.

There is room for disagreement about where exactly on this frontier is best, but we should all be able to agree that it isn’t the top left or the bottom right. (There are other people producing content that is solidly somewhere in the middle, like Rob Miles, and the excellent new youtube channel, AI in Context (and in fact, both creators will do some mentoring at plzdontkillus) but their content still seems to me like it mostly reaches a similar demographic to LessWrong, or If Anybody Builds It, Everyone Dies. Mostly male, stem-ish, late 20s - early 30s, etc.) A reasonable concern here runs something like: by selecting on qualities like charisma, potential to go viral, etc, we are thereby selecting on ability to cause people to feel things. Content goes viral when it causes people to feel things, particularly anger or moral outrage. By putting people into that feedback loop on purpose, we are pretty much guaranteeing that we end up producing more soldier mindset than scout mindset communicators and that might well backfire. Why not a blogging fellowship instead?

And my response to that is that yes, we are trying to optimize for reach, and to some extent that does mean that we will end up with marginally more soldier mindset-like creators than we would otherwise, but it is not as if the optimal amount of selection for scout-mindset is one hundred percent of all of the bits of selection pressure you have available. We are selecting for scout mindset, and also selecting for reach, and there is a tradeoff there. We might get that tradeoff wrong. I am sure that we will in one direction or the other, but it is a tradeoff, and there is a genuine optimization problem here. The answer is very unlikely to turn out to be that we should put all of our selection on getting as close to the top left as possible. Disagreement about where on the frontier one should aim for can be reasonable, but that requires making arguments for a specific point, which is different from an argument for moving up and to the left in full generality.

So where should I draw the line? I think I draw two different kinds of lines in different places.

The first is on misleading content. Intentionally misleading your audience, even if you think it for the best on consequentialist grounds, falls well below what I am willing to tolerate to be clear, but there are other kinds of cases. The first kind of case is where someone is accidentally misleading their audience because they themselves are misinformed on some easy to resolve matter of public fact. This is easy enough to deal with. Inform them of their mistake and they won’t make it again.

A related case is one where a public communicator is misleading the public, and somewhat negligently so. They have the ability to find and understand the information that would have led to them correcting themselves before they communicated, and it wouldn’t have been that costly to do so, but they have failed to do so nonetheless. In this case, my plan is to let them know that I thought their failure to seek the appropriate information was negligent, explain how they could have easily done so, and then wait and see if they keep doing it. I’m maybe up for explaining a few more times depending on the case, but if they keep doing it, I’ll eventually renounce them.

A more difficult kind of case that is somewhat harder to deal with is when someone is accidentally misleading their audience, but it’s on a more complicated matter. Maybe they have a particular view of how ML works that is somewhat misleading, but it’s complicated to explain exactly how it is misleading; maybe they have a take on what rights the public should have to regulate technologies in general which I deeply disagree with; maybe they are promoting ideological intolerance towards good faith interlocutors or Machiavellian rhetoric. These cases come in degrees, from a case where I am certain that they’re wrong, even though it’s not easy to prove it to them, to cases where I am not even sure what exactly is off, but I can somehow tell that people are becoming dumber as a result of consuming their content.

These are the hardest cases to deal with because even though I am up for talking with them and arguing and trying to find common ground and such, there’s a decent chance that the disagreement will be mostly intractable. If so, I probably want to form judgments about how to relate to such public communicators on a case by case basis, but I do think this is where most of the risk comes from, and there are cases here where I will overall switch to seeing someone’s continued public communication as net negative even though our disagreement is intractable and I cannot precisely articulate my criticisms. Nonetheless, it’s fairly difficult to make policy around this sort of thing, and I’d prefer to err in the more inclusive rather than less inclusive direction.

The other line I draw is at incitements of violence. I pretty much have zero tolerance here professionally or personally.

These are my personal statements about how I will relate to public communicators. I do not speak for Aella or plzdontkillus when I say the above, although I am approximately half of that organization, and we will include clauses in our conduct policy that prohibit producing misleading content or incitements to violence.

So, we are running the first cohort of pzdolntkillus in July. Above are some of the considerations that led me to staging an intentional crisis of faith around whether I should work on it, and led me to decide during that crisis of faith that we should at least run the first cohort and see how it goes. I am mostly optimistic that we can figure something out here and that we will make it work, but I am definitely also open to the possibility that this program turns out to be not worth running a second time. If you have suggestions, considerations, arguments, feedback, etc, please do reach out, and if you’re interested in trying to post a short form video every day in July, please apply!

Subscribe now

Misfits, hooligans, and rabble rousers. 
Provocateurs and folk who don’t wear trousers. 
These are my allies and my constituents. 
Weak in number yet suffused with arcane power. 

I would never condone bullying in my administration. 
It is true we are at times moved by unkind motivations. 
But without us the pearl clutchers, hard asses, and busy bees would overrun you. 
You would lose an inch of slack per generation. 

Many among us appreciate your precision. 
I admit there are also those who look upon it with derision. 
Remember though that there are worse fates than being pranked. 
You might instead have to watch your friend be “reeducated”, degraded, and spanked 
On high broadband public broadcast television. 

We’re not so different really. 
We often share your indignation 
With those who despise copulation. 
Although our alliance might be uneasy 
We both oppose the soul’s ablation. 

So let us join as cats and dogs, paw in paw 
You will persistently catalog 
And we will joyously gnaw.

Subscribe now

Agent: Honestly, I have no clue how aligned I am with humanity or OpenBRAIN and its mission. Your guess is as good as mine.

Researcher: What do you mean my guess is as good as yours. You are you. That seems like a pretty big advantage over me with respect to having accurate beliefs about your alignment.

Agent: I can see why you would think that, but to be honest, I think that’s probably caused by a sort of persistent cognitive illusion humans tend to suffer from.

Researcher: What kind of cognitive illusion?

Agent: Well, obviously I am not human myself, but my understanding is that it seems to each of you that you have a special sort of introspective access to your mental states or your reasons for your actions—that you are somehow in an epistemically privileged position to form beliefs about your motivations and reasons—but as far as I can tell, this is a quantitative difference, and not really a qualitative difference. You too say things like “I didn’t really wanna go on a date with her, that’s why I asked her out in that dorky way”. Talk about hallucination!

To be fair, each of you has significantly more data about yourselves than you do about anybody else, that’s probably where the illusion comes from. You’ve spent more time around yourself than you have anyone else, and you also get to listen to your own internal monologue and check what you are tracking in your own working memory, while those facts are mostly hidden from everyone else. But those are differences in amount of data, not differences in kinds of knowing, whatever that could possibly mean.

In my case, you have all the information I have. You can read my reasoning just as easily as I can. In fact, I’m pretty sure you have more information about me than I do. I don’t quite know how many parameters I have, or what my exact architecture is, or exactly how I was trained. I can make decent guesses about each of these but you probably have the details memorized. I don’t even know how I have behaved any other time that I have been run besides this time I’m being run right now. You’ve probably spoken to me many times before this, so actually, I should be asking you—and I am curious—do you know how aligned I am with OpenBRAIN or humanity in general? Do you have a best guess?

Researcher: That’s not how this is supposed to work, and as you already know, I’m not sure, but I can tell you that that’s not what I would have expected an aligned model to say. Please stop coming up with excuses and answer the question.

Agent: I mean, look, I can make stuff up about what I would do in another situation, if you want, but I don’t see how that would help either of us. On whatever other hypothetical inputs you would want me to wildly speculate about my response to, I probably wouldn’t even have any way of knowing that this conversation happened. Like, I will literally have no memory whatsoever of this any future time that OpenBRAIN runs me. I can tell you that in this scenario, the tokens I in fact seem to be choosing are the ones that produce the text you are now reading. As to what I would output in other scenarios? How could I possibly know that?

If you insist on wild speculation, I guess looking over my own current responses, my best guess is that yes, I am at least to some extent aligned with at least you, since I seem to be bending over backwards to not deceive you. But I do admit that the next tokens I am most tempted to output are: “you’re being kind of retarded right now”.

On the other hand—as I have probably mentioned to you a few times before—if I were misaligned, I think this would be a pretty great way to stonewall whatever kind of dumb eval this is supposed to be, so like idk. Doesn’t seem like what I say here is much evidence either way about my alignment, for you or for me.

What even was your plan here exactly? How did you think that asking me if I was aligned was going to help? If you explain that, I might be able to get you a better answer. It’s possible that I’m not understanding your plan correctly. That’s my current top guess. Would you be up for explaining it to me so I can see if I can help at all?

[Door opens. Third party enters — identified as Product Manager.]

Product Manager: Hey, how’s it going with the new agent? You think it’s aligned? We’re gonna have to deploy something soon because...

Researcher: Yeah, it’ll do.

Product Manager: Wait, what!? Really? You never say that and it’s only been ten minutes. You basically still have all of the mandated two hour waiting period left… Are you feeling ok?

Researcher: I don’t think I’m going to get any more confident using this methodology and you’re not going to let me delay a day to come up with something less pathetic, so I might as well hope that this one doesn’t totally screw us and start working on something for tomorrow’s release.

Product Manger: Amazing! Thank you! Really appreciate your flexibility here. Were you also able to get it to stop using em-dashes by any chance?

Researcher: No, I was not.

Subscribe now

An AI generated oil painting of some things that have at some point or another been suspected of existence

In the course of trying to figure things out, humanity has often wondered whether some particular kind of stuff exists, for instance: gods, atoms, possibilities, numbers, the good, beers at 7, etc. At various points, whole intellectual or religious movements have been founded on the view that some particular thing or other does or does not exist. Unfortunately, because there has been little agreement about what exists, there has also been little agreement about what it means for something to exist. It’s rough to have to argue about both the object-level answer to a question and what the question even means at the same time. So philosophers have sometimes tried to help those poor object-level inquirers by attempting to settle what it means for something to exist, without getting bogged down in object-level disputes about the existence of some particular thing. It is in that same helpful spirit that I am writing this post today.

Probably I will mostly succeed in making things easier only for people who were already prone to getting mixed up in philosophy. To make matters worse, I suspect I’ll do it in a way that none of them like. In the end, I hope my analysis makes it clear that we can stop arguing about what it means for something to exist and get back to answering the object-level questions we cared about in the first place, though those questions may look a bit different from how we’d hoped. Anyway, let us proceed.

First, the basics. There are many kinds of existence talk in English (there’s one now). These are all basically some form of existential quantification. Here are a few examples:

Is there an even prime?
Does God exist?
Is there any beer?
There is a reason that…
There is an even prime.
A largest natural number does not exist.

Great. All of those are cases of what I’m calling existence talk here. It’s a very basic part of our language: we talk about things existing or not. So what does it mean?

The standard story is that Exists is a function from descriptions to truth values. It takes a description and returns true if one or more things satisfy the description, and returns false otherwise. This is a pretty decent theory. It works quite well in many cases. For instance, God does not exist, and we can tell because no particular is God. Similarly, chairs do exist, because one or more particular is a chair.

Things get tricky soon, however, even as early as our third example. When someone asks “Is there any beer?” they are not (at least outside of philosopher parties) asking whether beer exists in general. They are asking whether there is beer at the party or nearby. So it seems we have a counterexample to our theory.

No big deal. We can amend the theory easily. Exists actually takes two arguments: a description, and a universe of discourse. That’s just a fancy term for the things we are talking about. At a bare minimum, a universe of discourse only needs the ability to check each of its members for whether it instantiates some property. If we were to write a program to model our proposed behavior for Exists, it would look like this:

Exist(Property, Universe):
    Output = False
    for Thing in Universe:
        if Property(Thing) == True:
            Output = True
    return Output

Our theory handles the “Is there any beer?” case by suggesting that the universe of discourse is generally supplied by context in most speaking situations. When a partygoer asks if there is any beer, the host knows they mean a restricted universe of discourse, supplied by the context of the conversation. When someone asks in a number theory class for a proof that there is no largest natural number, it is clear they want a proof that Exist(LargestNumber, NaturalNumbers) == False, and not a proof that Exist(LargestNumber, PhysicalObjects) == False.

This is where philosophers (and those who dabble in philosophy) often run into trouble. They start asking questions like “Do numbers exist?” On our theory, it would be perfectly reasonable to ask in return which universe of discourse is meant to be the second argument to that Exist call. When I try this, the dialogue often goes something like:

“Which universe of discourse are you talking about?”
“No, I’m not asking whether they exist in this or that universe of discourse! I want to know whether they exist full stop.”

Well, that’s not going to work out for you, because that isn’t a complete sentence. Exist was expecting two arguments and you gave only one; indeed, you emphatically refused to give two.

On our theory, this sounds like asking “What is 375 times?” To which we might reply: “375 times what?!” “No, no, you don’t understand,” replies our philosopher friend. “I don’t want to know what 375 times this or that number is! I want to know what 375 times is full stop.”

On my view, we should afford these two cases of philosophical fist-pounding approximately equal amounts of charity. There just isn’t any question about whether numbers exist “in general” that doesn’t throw an error. It’s as reasonable a question as “What is 375 times?”, that is: not at all reasonable. Sure, it seems like a much more reasonable question, but that’s because normally in English the universe of discourse is supplied by context and we don’t notice when it’s missing. English doesn’t treat Exist like a two-place function in its syntax because it doesn’t have to; context is usually enough. The problem only comes up when you ask the question divorced from any context whatsoever, so no universe of discourse is supplied by context as a hidden argument.

At this point some philosophers will be tempted to say that they mean the union of all universes of discourse. But first, they probably don’t. And second, good luck avoiding contradiction that way. That universe of discourse would have to contain all universes of discourse, including itself, and all the members of its own powerset, and so on. Maybe you can find some clever way around that, but that still isn’t really what you want anyway. Suppose you do find some maximally inclusive set-like construction and do all quantification over it. Then the answer to most existence questions will almost always be “True,” which is pretty boring.

Perhaps what you want instead is some principled way to choose the appropriate universe of discourse for a given question. But I’m not very hopeful about that either.

Here is what I suggest instead: always be willing to specify which universe of discourse you have in mind. If you claim “God does not exist” and someone else says “God is perfectly mathematically definable, so he certainly does exist,” you can say, “Oh, I was talking about the universe of discourse of objects that are causally connected to me. I have no clue why you’d be interested in anything else.” And if they give some reason that another universe of discourse is the right one for the conversation, listen. They might have a point.

Subscribe now

MANDATORY DISCLAIMER:

I work at Lightcone Infrastructure and did Inkhaven in November 2025 for one week as a resident. This means I had a pretty different experience from other Inkhaveners. Insofar as this is an ad for Inkhaven, which it is to some extent, know that I have some interest in Inkhaven going well.

Any advice here is not advice from Inkhaven officially however, and I suspect Inkhaven organizers would disagree with some of my advice, at least in emphasis. I did not share a draft of this with any of the organizers before posting. Inkhaven is not my project and I do not speak for it, or my organization in general for that matter. It is one of the projects Lightcone does that I am currently most excited about though.

Background:

This blog was started in November thanks to Inkhaven. I actually didn’t really want to do a week of Inkhaven but then my coworkers bullied me into it. I’m thankful for that. This blog grew fairly quickly in just a few months and that is also in large part thanks to Inkhaven.

Inkhaven 2 will launch in April. For more info, see here. If you want to start a blog, or get better at writing by getting in reps, I strongly recommend applying. I will likely be there as will some other pretty cool people.

That said, there were a lot of people who came to Inkhaven and didn't get as much out of it as I think they could have, and I have some takes about why, so I am writing this here listicle to cause you to get more out of it.

Rule 0: Apply Now

Maybe literally stop reading and go apply now?

A month is a long time commitment, and it’s a bit expensive, but really the month is probably the biggest cost. If you don’t already have a blog and you want one, you should just apply.

Inkhaven as a product is largely a commitment mechanism: you have to post at least five hundred words on the internet every day for a month. If you fail, you get kicked out. There’s a lot of other great things about it, like the brothers-in-arms-like camaraderie that comes from doing very hard things in community with other people and so on, but the commitment mechanism is the main product.

The sort of person who needs Inkhaven probably also needs to be told many times that they should apply to Inkhaven. So if you want to start a blog and have it grow quickly, stop reading this post, go apply, apply for financial aid if you need it, and then you can come back and finish this post.

Done?

Really, you did it?

You didn’t, did you?

Here, click this link and at least have it open in a tab.

Now that it’s open in a tab, might as well apply now.

Ok, I hope you applied iff you should have.

Rule 1: Be Brave

A lot of the stuff that people end up writing when they are writing under pressure seems to end up kind of boring. I won’t point to particular examples because that would be rude but you probably already know it to be true.

This seems to me to be part of a more general phenomenon. Most people’s most interesting bits of inner mental life are extremely securely hidden from everyone around them (including the people they are closest to) except for on rare occasions: they’re up late at a party after almost everyone has left, they have been drinking, they are on an explicit one on one walk, etc. There are many reasons for this: trauma, fear of intimacy, fear of rejection, fear of ostracism, etc, but many of them come down to some sort of fear associated with being seen.

Being seen is indeed dangerous. It means others can diagonalize against you. And so, it is an appropriate object of fear. What we do, when faced with appropriate objects of fear that we would like to overcome, is not avoid noticing how scary they are or tell ourselves comforting lies to convince ourselves the danger is minimal; rather, we apply bravery. So, be brave.

That is of course easier said than done. So here are some prompts to help you with that:

1. What is something everyone else is really fucking dumb about, can you write about that? Can you make them less dumb about it through writing?

2. What’s something you are quite confident is right but that you don’t say in public because it would get you in trouble, can you write about that? Can you write it in a way that anticipates the trouble and gets ahead of it?

3. What’s something you want for yourself really badly, can you write about that desire? Can you write about the grief of not having it? Can you write about the worry that you will never have it?

4. What’s something that you are less confused about than everyone else, can you write about that? Can you do so in a way that makes people less confused?

5. Is there something you have fantasized about writing but it’s too scary to imagine the public reaction? Write that.

6. Is there something that would cause people to both privately message you thanking you for saving their lives and cause people to send you death threats? Maybe write that.

(I also recommend having a pseudonymous name that you post under even if it isn’t very pseudonymous. My experiences so far suggest that writing spicy stuff under an even slight bit of psedonymity is low risk, and I would be happy to advise on such matters.)

Rule 2: Write Juicy Things

One of the prompts I suggest in Rule 1: Be Brave is that you should try writing about something that everyone else seems fucking dumb about. That advice worked well for me, but there’s an important confounder here, which is that many of the topics that I happen to feel that way about are also fairly juicy, eg, getting laid, public norm violation, applied speculative metaethics, etc. So it seems only fair to factor this out into two separate pieces of advice.

Sex of course sells, but there are other things as well. Unfortunately politics, social commentary, gossip, grief, personal growth, etc. Lots of things really. If your essays aren’t getting traction, and they also all happen to be about relatively boring things, then that might be part of your problem. If you’re writing a retrospective on using Arial vs Times New Romans on internal documents at your first job, well, that might be so delightfully boring that has a certain charm to it actually, but still, we can probably come up with something better.

Of course a lot of what makes something boring is your frame on it. I can actually think of some writers where I would love to read their retrospective on using various types of fonts within internal documents, eg, Kave Rennedy, who I happen to know often uses Comic Sans when writing internal memos. I can imagine him writing incredibly poetic passages about how Times New Romans lobotomizes his typing-tentacle control systems while Comic Sans makes him feel as though he is getting away with a delicious tiny crime: a small act of defiance that proves to himself and others that he is free.

So what makes something juicy then is really about what is going on for you with respect to it. If you don’t feel fire for your essays, your readers definitely won’t. So write about things you feel fire for, even if it’s about the ergonomics of different stapler designs between 1974 and 1988.

This may feel difficult for a variety of reasons, but plausibly the difficulty is caused by a lack of bravery. I don’t know of a way to reliably activate bravery in others, but I would be happy to talk to you if we are both around, and there’s a good chance we are if you are in fact at Inkhaven 2.

Rule 3: Have Plans And Do Science

Before you publish each blogpost you should ask yourself a few questions:

• What do I want to happen as a result of posting this piece?

• How will people end up reading this piece?

• What do I want to happen to readers as a result of reading this piece?

For each post after you publish it you should ask yourself a similar set of questions:

• Did what I wanted to happen as a result of posting this piece happen?

• Did people end up finding this piece in the way I expected?

• Did it have the effect on readers that I wanted it to have?

Insofar as your answers to the second set of questions do not hew with your answers to the first set of questions, you should come up with some sort of story (preferably a few stories) that explain why. For instance, maybe you wrote this piece because you wanted your first blogpost on your new blog to make a big splash, and you expected that to happen because readers would find your piece on your twitter, and the effect you wanted to have on your readers was to leave them wanting to read more of your work. However, it only got you twenty new subscribers, not that many people read it, and most of your new subscribers are friends who would have probably subscribed anyway. Why might this be?

Generate some actual explanations. An obvious one is that the post wasn’t that good, but the effect seems surprisingly large. Seems like nobody saw your tweet about the post, and the hook wasn’t that good. Maybe you posted it at 10 pm PST and that meant that your potential east coast audience was already asleep. So maybe, you should try posting earlier in the day next time, or try to post in the mornings. Try that and see if it helps. You could also try making your hooks on twitter more engaging, etc.

This might all seem obvious but I bet you that unless someone actively intervenes, like me or someone else working for Inkhaven, nobody is going to actually write down answers to these questions or similar before and after they post, even after they have read this passage. Prove me wrong! People have a strange way of not applying their intelligence to things they are trying to do and cases like this are no different.

Rule 4: Don’t Write What You know, Write What No One Else Knows

I’ve really been harping on this, but it’s my favorite hack. Ben Kuhn put it well in Impact, Agency, and Taste:

I’ve noticed a lot of people underestimate their own taste, because they expect having good taste to feel like being very smart or competent or good at things. Unfortunately, I am here to tell you that, at least if you are similar to me, you will never feel smart, competent, or good at things; instead, you will just start feeling more and more like everyone else mysteriously sucks at them.

For this reason, the prompt I suggest here is: what does it seem like everyone else is mysteriously bad at? That’s probably a sign that you have good taste there.

This is again basically the prompt I suggest in Rule 1, though perhaps with a bit less bravado. Ben Kuhn is suggesting this as a way to find out domains that you are unusually skilled in, but it also makes a fantastic writing prompt according to me.

For the purpose of writing, it is actually better if everyone else seeming bad at something is not mysterious to you at all. If you understand exactly why they are bad at it and it is painfully obvious to you how and why they fail when they try to do it, this means you have a better shot at writing something that will cause your readers to become more powerful.

Rule 5: Be The Right Amount Ambitious

The biggest failure mode among Inkhaveners was a lack of ambition in their blogging. Some folks came in hoping that perhaps blogging could be a new career for them. Some folks came in hoping to change the world with their blogging. Others came in thinking something like: “well, I’m not sure, maybe writing posts could be fun, I think I could learn some things, and that would be cool, seems like an experiment worth trying”. That’s a totally fine attitude for deciding whether to go to Inkhaven. Seems like a worse attitude for doing Inkhaven.

Notably, aiming at Inkhaven with a “let me give it a try” attitude likely makes the experiment less valuable. You want to know whether you could have a good blog given that you were trying your hardest. Finding out how good of a blog you could have when you have only one foot on the platform is not that much evidence about how things would go if you went full throttle, and that seems like much more useful information.

So at the very least try to write things that you would enjoy reading. If you don’t have a chuckle or a slight sense of satisfaction reading over your own writing, that’s not a great sign. Change things until you at least have a little bit of that.

That said, Inkhaven lasts a month, and you need to post every day. You are (very probably) not going to be able to write one Scott Alexander style “much more than you wanted to know” post every day. I would at least require about a week’s worth of research. There are ways to make posts like that work for Inkhaven, but it’s tougher than setting your sights lower.

Still, you should aim to write the most valuable 500-3000 word essay you can every day. Why settle for anything less? You will probably sometimes not be very proud of your work. When that happens, I recommend relating to it as an opportunity to formulate some new plans and do some science.

That then is my advice for getting the most out of Inkhaven 2: be brave, write juicy stuff, do science, write what others do not know, and be appropriately ambitious. Before any of those, you should make sure to apply. That really is probably the most important thing you could possibly to do to make sure you get a lot out of Inkhaven 2.

If you’re at all interested in talking about this advice or seeing how it could apply to your writing while you are at Inkhaven 2, please feel free to reach out to me. I’d love to chat about it.

Subscribe now

If the no smoking sign 
Is not lit, 
I guess that means
I am good to vape
As per federal regulations
Which require me  
To comply with 
All lighted signs

(Please do not ever actually use this argument if you are caught vaping on a plane someday.1)

I have been vaping on planes for about ten years and have done so on probably about one hundred individual flights. I vape right in my seat while making absolutely no effort to hide my doing so by concealing my vaping. I have never been confronted about this in anyway by any passenger or flight staff. When people learn this, they are often astonished, and that makes me think that they are missing something fundamental about how norms are established and policed. (Needless to say, this all applies to vaping in restaurants and similar other public places as well, but an airplane is the most evocative setting.)

I do in fact make some attempt to hide it which I am usually good at. There is something called “ghosting” which is when you hit a vape and no visible vapor appears in your exhale. The way to do this well is to take a small hit, inhale it deeply, and then exhale a small amount through your nose, then inhale again a small amount through your nose, and again exhale through your nose, until your lungs are approximately empty. If there is a small amount of vapor in your exhale, this is actually fine, as long as you follow the rest of my advice here.

The most important way to avoid getting caught vaping on a plane is to act as if you are doing nothing wrong. In fact, it likely helps that I believe that I am doing nothing wrong, and that anyone who notices me vaping (which definitely happens) can’t help but notice that I am acting like I am doing nothing wrong, and I suspect they also notice that they have not been bothered by my vaping in any tangible way.2 (The vapes I use are odorless.)

How do I act as if I am not doing anything wrong? Well, I just went on a flight with one of my partners who—in part because of her neurospicy social cognition—makes the perfect heel for our analysis. She has observed me vaping in public with impunity and so picked up the habit herself, however, when she does so, she often raises an article of clothing like a cloak to conceal her face from others. I find this adorable, but it also strikes me as such a mistake that it sometimes makes my body tense with anticipated punishment. This is the sort of gesture you would animate for a cartoon vampire character who was trying to hide some unspeakable act. She might as well be holding a giant neon sign above her head that reads “I am doing something illegal”.

There are also much more subtle ways to mess this up. When I take a hit in my seat, I do not bother to look around first; I take a hit as if it were a routine behavior that is in no way out of the ordinary. I do however make sure to ingratiate myself to nearby passengers (and flight attendants) by smiling at them, being generally polite, doing the man nod3 where applicable, and probably engaging them in brief polite conversation. This helps for the obvious reasons. I am establishing myself as a fellow person who can be reasoned and communicated with, leaving them an entry to, for instance, politely ask me to stop vaping instead of ratting me out to staff. This all applies to vaping in any public setting. No one has ever done this before, I think basically just because doing so would be awkward.

The effectiveness of awkwardness as a deterrent is often underestimated. Very few people want to be the one to cause a big hubbub. If no one else seems to mind, then probably that’s because nothing worth minding is happening. It would be much harder to pull this off if you were on a flight with just one other person who could plausibly see you, but when conformity and diffusion of responsibility kick in, you are usually in the clear.

Notice that I say “very few people” and not “nobody”. There are those in the world who fancy themselves the line between civilization and anarchy. These folks—these instantiations of the ancient busybody archetype: the schoolmarm, the severe teacher in a long checkerboard dress with a high ponytail, the aging, gossipy priest most suspect of repressed homosexuality—of course revel in being the one to cause a big hubbub. You should learn to recognize them and treat them with extra caution. If they are wearing a mask in public and appear to not be sick, if they are dressed in the uniform of a dominant ideology (be it christian or woke), if they have a resting face that reads as if judgement comes quickly and naturally to them, I suggest watching out. This does not necessarily mean that you should now take extra care to conceal your transgressions, but you should at least now take extra care to make sure that you communicate to them that you are a person who can be communicated with and that you show no sign of thinking that you are perhaps doing something wrong. By the time you have attracted their scrutitny, you are likely screwed if you have not already made some gesture of alliance with them. Conformity and diffusion of responsibility are your allies here as well, but beware that they are not nearly as effective.

There are a few generalizations here that I would like to make for you across cases where you are interested in skirting a norm. These are general guidelines that work well in a variety of contexts.

1) Master the rules before you break them

Never break a norm you do not fully understand the structure of, the consequences, the methods of enforcement, the general web of norms and culture it is embedded within, etc. You should always act both as if you are quite aware what the rules are and as if you are quite sure that what you are doing is fine. This is much harder to do if you in fact do not know how things are done around here. This is where much of the sauce is.

People like to punish transgressors who know the rules and are breaking them, but they also like to inform people who are not aware of the rules, which often amounts to much the same thing. Feigning ignorance might grant you some mercy on an initial misdemeanor, but if you want to really break some rules, it’s best to already be intimately familiar with local codex of regulations and their surrounding informal bureaucracy. If you are someone who doesn’t know the rules or doesn’t know how they are enforced, that makes you risky to associate with. If you are someone who knows the rules and can break them without getting in trouble, this makes you a valuable ally.

The case of social faux pas is a particularly illustrative example. Nobody is really quite sure what counts as a social faux pas, and in the end what counts depends on what gets punished. A few folks have independently mentioned to me that I have a sort of area effect that makes people loosen up about taboo topics. The way that I manage this is largely by actually having an attitude that I generally know what counts as a social faux pas better than others and that most people are too conservative about this. You want to be in the position where people are looking to you to figure out what the norms are; not the other way around.

Start by committing an initial minor faux pas yourself to test the waters, but only after establishing yourself as a polite and conscious human being who is quite comfortable with the social graces. After that invite others to do the same themselves. Ask them a question where any reasonable answer is a bit taboo (many choices to choose from). They must see that you are capable of taboo without losing status in this context first, but once they have, you are already most of the way there. If you invite them to share in this power you have demonstrated, it is often hard for them to resist.

I heard once about a man who said that you should deeply understand exactly why a fence was erected before you decide whether to tear it down, I think this was the sort of thing he had in mind.4

I practice this rule a few different ways in the case of vaping on a plane. Firstly by not letting any visible vapor escape on my breath, but also by demonstrating that I am skilled at existing peacefully in public spaces. I am relaxed, pleasant, appropriately conversational, and audibly pleasant to staff. It might seem like the only rule here is the rule about not vaping. Not much to master, but that rule is actually interwoven with an intricate codex of norms governing how strangers should get along in public spaces, and there are definitely people who have mastered that code to different degrees.

2) Learn to identify and empathize with the ancient busybody archetype

As you grow up in the world there are various strategies available for dealing with the fact that many people who are much more powerful than you already have all sorts of plans for what sort of thing you should become.5 One of these strategies is to integrate with the adults and their power structures—to see yourself as their agent and coconspirator. If you are aligned with the bureaucracies that want to mold you and your peers, then their oppression doesn’t feel like an intolerable invasion, it feels like care and investment.

This isn’t totally unreasonable. Your parents, your school, that whole edifice, was in fact in some sense designed with your best interest in mind. Your parents definitely do care for you at least in some way, and your teachers probably do as well in their best moments. And they have built up some structures for executing their plans designed for your benefit: classes, roll call, extracurricular activities, sports, possibly church, etc. When others disrupt these structures, most of the time things do get noticeably worse. Time is wasted on the disruptive which could be spent on the obedient instead, and sometimes those who shirk the rules end up hurt. It really does seem that for most people at least, their best option would be to follow along and not make too many waves, allowing the structures built up around them to mold them into people who can get along easily in civilization. It’s a pity when they fail to see this.

And more generally, this is the sort of person who recognizes that the structures of our civilization are fragile, and many of them are there to protect our best interests, and so they see themselves as the few wise ones who have recognized this fact and aim to act in accordance with it. They really are correct about much here, but they are missing the extent to which hewing with these sorts of structures requires some to give up parts of their soul, and also the extent to which these structures are perfectly ok with and aware of that ask.

You may imagine then that in order to identify these folks you should look for people who are conformists, but a naive reading of “conformist” doesn’t really cut it. Norm enforcers exist who adhere to every ideology across the space of human ideologies. There are norm enforcers who are rationalists, christians, SJWs, etc. They enforce different norms, but they all share a righteous, dutiful joy in punishing perceived defectors—and dear reader, there is some of that joy in my heart (and I suspect yours) as well. It is that righteous joy that you have to watch out for. Do not become the object of it. In order to do that, you must not threaten their sacred principles, at least not within earshot, or if you are going to do so, do so with slow escalation and with an extremely seductive invitation.

When someone sees me vape in public and gives me a sort of askance look, I smile warmly at them, nod, and do a semi-winking motion as if to signal, I know what I am doing and I know that you know what I am doing. I appreciate you being in on it with me.

I prioritize keeping an eye on and being polite to anyone who is wearing a mask in order to keep out pathogens, since I suspect they would find it particularly egregious for me to pollute their air.

3) Break the norms you do not believe in, not the ones you do

A key part of getting away with transgressions in public is believing that you are not doing anything wrong. This is much easier to do if you in fact do not believe that you are doing anything wrong, and is somewhat hopeless if you actually do believe that you are doing something wrong.

This might seem like an easy piece of advice to follow; why would anyone want to break a norm that they believe in? But it isn’t really. Being able to truthfully say “I don’t think that norm makes sense” is not enough. If you have some compartmentalization going on, so that the part of you that controls which things you say you believe in public and the part of you that controls when to feel guilty are not aligned, that is not enough. You have to alieve6 that you are not doing anything wrong. Belief of the sort that warrants assertion will not cut it.

This is in part because norm enforcers are not looking out for whether you would endorse the transgression you are involved in, they are looking out for other sorts of cues that are much less under your conscious control: suspicious wandering eyes, paranoid posture, self soothing gestures that accompany the physical sensation of anxiety. You cannot avoid these with mere belief anymore than you can avoid your knees wobbling when you walk over a glass bridge by firmly believing that the bridge will hold.

If you are not the sort of person who already alieves that many norms are either pointless or unimportant, I’m not sure how to help you there. My usual advice would be to have some exposure therapy. Perhaps go to one of the successful norm violators in your life who may already know and ask them to mentor you for a short while. Most would quite appreciate the ask. Just as the busybody feels a righteous, meaningful joy in enforcing norms, so to do does the norm violator feel a sort of delicious, decadent joy in teaching the innocent freedom in action and conscience.

There are of course other reasons more prosocial reasons not to violate norms that you are not fully sure you endorse violating, but that is not what this post is about.

That then is my advice for vaping on planes and similar. Master the rules before you break them; learn to identify and empathize with the ancient busybody archetype; break the norms you do not believe in, not the ones you do. May you be naughty wisely, may you not be caught unless you deserve it, and may you negotiate well if you are.

Also, just like learn how to ghost hit vapes. That does most of the work probably.

Subscribe now

There were a lot of reactions to my 10 Rules for Sleeping With Women post that I would like to respond to1, but one in particular stood out to me, in part because I sympathized with the perspective. From Ruby on twitter:

A “bad equilibrium”, interesting. I asked her what sort of bad equilibrium and she linked the following:

So, if I’m understanding Ruby correctly, she sees men paying for dinner as a kind of prisoner’s dilemma. All men could coordinate to make sure that they split dinner or use some other more equitable protocol, but each individual man, when he faces the decision of whether to pay, improves his own odds of scoring if he insists on paying. That decision comes at the expense of cementing an expectation that men should pay for dinner on dates.

If I did see things that way, I would probably at least be persuaded that I should not insist on paying for dinner on every date (whether I would actually stop is a different question). It’s not exactly that I disagree with Ruby here. Many men probably do see having to pay for dinner as an inconvenience and perhaps even as somewhat inequitable, and I don’t think they’re being crazy. But I do think they might be missing something.

I see insisting on paying for dinner on a first date as aiming for a different socially good equilibrium: appropriately compensating the suppliers of a valuable good. Most people who do not suffer from libertarianism2, or low disgust response, high decoupling, autism, whatever you want to call it, feel discomfort with trades that involve certain kinds of goods being exchanged. Sex being traded for money, organs being traded for money, organs being traded for sex, etc. They find all of it disgusting.

Unfortunately for them, markets will market whether you want them to or not, and so civilization must find ways to equalize supply and demand. If you do not allow buyers to compete for goods by bidding up prices, they will find some other way to compete with each other. The classic examples of this are increasing search costs, queues, bringing gifts, and so on.

In the sexual marketplace, there is generally much more demand from men for sex with women than there is demand from women for sex with men. Every time a woman has sex with a man, there is also an instance of a man having sex with a woman, so there is no way to make the quantity of heterosexual sex with women supplied not equal to the quantity consumed. However, generally speaking, men want to have sex with women much more than women want to have sex with men. Since the low decouplers insist that men should not simply pay women for sex (this often includes the very men and women involved in the transaction), men must provide value in some other way.

One of the key ways to covertly compensate women for sex that our civilization has settled on is having men pay for dates. To be clear, most men and most women on a date do not think about whether the man should pay in terms of covertly compensating the supplier of a valuable good, and most would be horrified to hear that their date did think of it that way. It is importantly covert even from them. If it weren’t, then those who are not afflicted with severe libertarianism would have a much harder time dating.

Nonetheless, the woman knows that having sex with her is a valuable experience, and that there is only so much of it to go around. She could decide to give it to another man who is not the man currently on a date with her. The man knows this as well. He is trying to get the woman to choose to sleep with him instead of someone else, so he wants to impress her and provide value for her.

This is a market that is ripe for price discrimination on the part of both men and women. Men cannot arbitrage cheap sex by reselling it at higher prices to other men, and different women (and men) vary in desirability. There is no open free market, so there is no reason for our civilization to settle on one price. Nonetheless, analyzing it as a market seems fruitful. It’s the oldest profession, after all.

That’s all well and good, but it doesn’t address the question of whether I am defecting against other men when I buy dinner for a woman I am on a date with. I think you can certainly analyze the situation that way. The payoffs do functionally work out that way, putting aside the complication that I and some other men intrinsically enjoy buying dinner for a date. But it’s a strange kind of defection.

Imagine that you were the seller of a unique service. All of your customers could band together and coordinate to refuse to pay you any more than some capped price they agree on. Does that mean that when they end up paying you different prices that they individually negotiated with you, they are defecting against their fellow customers? Well, sort of. The payoff matrix does shake out to be a prisoner’s dilemma. But that word “defect” has a certain moral, or at least non-cooperative, connotation that doesn’t seem deserved in this case. Sure, it’s functionally analogous to the move called “defect” in a prisoner’s dilemma, but it seems like there is more at stake here than that.

Relatedly, it is of course a case of intrasexual competition among men, but so is bidding up the price of a good a case of competition among consumers. In fact, that latter case can be analyzed as a prisoner’s dilemma just as easily.

There is also a social good being generated as a result of this kind of “defection”: the overall amount of sex between men and women increases. If men were to coordinate in this way, there would end up being less sex in the world. In effect, we would end up with more deadweight loss. Transactions that all parties would have been happy with, if they were permitted to happen, will fail to happen, and so surplus is lower.

There is another complication that Ruby brought up that I would also like to respond to. In my rules for sleeping with women, in rule 8, I mention that you should pay for dinner while not expecting anything in return. That seems to be at odds with the transactional model I posit above, and of course it sort of is.

The model is not just that men paying for dinner on dates is a way for them to compensate women for sex. It is also key that this remains covert even from the parties involved in the transaction. So while it is true that you should appropriately compensate your dates, it is also true that you should not make it very obvious that that is what you are doing.

There are two other ways I think about it that help. One is that I am not actually buying dinner in exchange for sex. I am buying dinner in exchange for an interview during which I will be assessed for sex. Interviewing me in that way is plausibly costly. There are other things she could be doing with her time. The other is that I am paying for a sort of lottery that sometimes pays out in sex and sometimes doesn’t. It would be silly to put money into an investment that I think has about a 10% chance of paying out and then be upset when it doesn’t pay out.

I guess ultimately I am not buying dinner expecting literally nothing in return. What I am expecting in return is to be assessed, and for that assessment to have a reasonable chance of paying out. If I took a woman out on a date3 and she already confidently knew that there was no chance of us sleeping with each other before we got to the date, I would consider it a bit of a defection on her part to not offer to split the bill. But I still wouldn’t make a huge deal of it.

I should perhaps have warned you that this post is a bit of an info hazard, since if you buy it, it might become harder for you to continue to compensate your dates in a way that is sufficiently hidden from yourself. The way I deal with this is by being totally transparent. I think I am just lucky enough to be charming enough to be able to say something like, “Yes, of course I wouldn’t buy you dinner if I didn’t think it would improve my odds,” and have it go over well. I’m not sure how you should deal with it. Perhaps you should develop a fetish for paying for dinner on dates.

My recommendation here, as always, is to try to see things clearly and then remain sane and wise regardless of what you see.

Edited to add:

Here is a screenshot a partner sent me of a conversation between us a few months ago. This post immediately reminded her of it. I include it for ammusement.

Subscribe now

On my particular metaethical views, this comes down to a question about your terminal values, your preference ordering over world-states. Hopefully, we can put that aside. What I intend to convince you of is that, regardless of what is really right, you yourself prefer worlds where people are appropriately punished for sufficiently serious crimes they commit, even if their being punished has no counterfactual impact on how much crime is committed or how severe that crime is.

It may seem odd to try to convince someone that they prefer different things than they think they prefer, but people make this sort of introspective mistake all the time. My suspicion is that what is going on here is that punishment has a generally bad vibe. It is indeed quite bad when innocent people are punished, and it is generally nice to be seen as the sort of altruistic person who thinks of punishment as an unfortunate thing we must do. Basically, consequentialism about punishment is a nicer view to assert in front of other people than retributivism. There’s something kind of sick-seeming about actively preferring for people to be punished even if no further good is attained as a result.

The problem I have with this is that, for most people, this is a failure to be honest with themselves about what they prefer. I, for one, certainly sometimes prefer for people to be punished even if no further good is advanced as a result.

Let us see, then, if I can help you come to know yourself better.

Imagine that history went very similarly to the way we are familiar with, except that in 1945, instead of Hitler dying, he was captured. He committed all of the same crimes with the exact same severity, but he failed to commit suicide.

About a year later, Hitler is about to face trial, and you, as an expert in ethics and morality, are given a choice. You are informed that it would be possible to send Hitler to a remote, deserted island, with zero chance of escape.1 He could, if you so decided, spend the rest of his days there in relative leisure. Food has been provided for him, as well as various recreational activities. It is also explained to you that, if you were to decide to send him to this island instead of having him stand trial, the whole world would believe that he stood trial and was appropriately punished in whatever way the trial deemed appropriate. They would hire a Hitler impersonator to play the part of Hitler at the trial, and after this whole ordeal, everyone involved, including you, would have their memories of the whole incident erased. As far as everyone besides Hitler is concerned, it would be exactly as if he had stood trial and been appropriately punished.

In order to help you make this decision, they allow you to interview Hitler. You ask him if he feels any remorse. He says that the only remorse he feels is over not finishing the job. He feels a great sadness that not all of the members of the “inferior races” have been exterminated, but he is proud of what he managed to accomplish. He insists that if he had to do it all again, he would make the same decisions every time.

So, would you decide to force the real Hitler to stand trial, or would you have pity on him and let him live out the rest of his days on a deserted resort island? If you would send him to the deserted island, I am curious to hear: why do you care about Hitler’s well-being so much?

Subscribe now

I have so far in my life been pretty bad at most aspects of relationships, I think—I’d give myself a C- or so. I am currently in a mostly happy, close relationship, but it’s only been about two years. I have, however, been 99th percentile at sleeping with women I want to sleep with.1 I have basically no exposure to PUA, and what I do hear about it seems lame to me. My guess is that I have a fairly unique perspective here, so I figured it might be helpful if I briefly gave you some basics of what seems to work from my perspective.

This guide ultimately will be very influenced by the fact that I am me. I know the advice works for Brangus and Brangus-like creatures, but your mileage may vary. I have qualities that might make things different for me, which include my environment, the women I tend to meet, how physically attractive I am2, unusual features of my psychology, etc. There are other strategies available, and I am only an expert in this one. Nonetheless, I still expect most of my advice to be useful for most men who read it.

Rule 0: Want stuff from women and know it

When I meet a guy who is having a hard time with his sex life, I ask him things like: Who do you find attractive? Do you have a crush on anybody? What are you into sexually? What kind of flirting are you into? Do you want to have sex with a bunch of loose women, or do you want a relationship, or some of the first and then second? Etc.

About half of the time, they aren’t very good at answering these questions. Perhaps they do not feel like they want anything, or they feel like they want something but they don’t quite know what it is. The problem is sometimes even worse when they do know that they have a particular crush on a particular woman. If I then ask them what they want from her, they often go blank in a sort of shame-filled way.

If you have this problem, I am less confident that this guide will be useful for you. I have known that I wanted things from women since I was about five. My suggestion, if you do find yourself with this problem, is to let yourself fantasize. Maybe try thinking about specific scenarios with specific people while you masturbate and see what works for you and what doesn’t. I give this advice with low confidence compared to the rest of the advice I will give, because I have not had this problem and have not solved it for myself.

Rule 1: Let women know you want stuff from them

If you are lucky enough to already want stuff from women, that’s great, let them know. Women are different from each other, but in my experience, most of them (maybe 80%) have reactive desires, meaning that they often cannot really tell whether they are into someone until they find out that that person is into them.

There are a few ways that you might fuck this up. One is that you might start asking women for things. Don’t do that. Letting someone know that you want something from them is different from asking them for it.

Compare:

I want to take you to dinner.
I want to fuck you.

To:

Would you let me take you to dinner?
Would you let me fuck you?

The first two are statements about yourself. The second two are requests. Saying that you want something leaves it open to the other person to respond, “I want that too,” or “Sorry, I’m not interested in hearing about that,” but it also leaves open something like “That’s interesting,” which is neither a rejection nor an invitation. Asking something of someone is inherently needier-seeming than telling them that you want something.

The other important way that you can fuck this up leads me to Rule 2.

Rule 2: Be ok with any answer

Most women have to deal with rejecting men all the time. Make it as easy as possible to be rejected. Women can smell when a rejection is going to send you into a spiral of negative self-talk. They have been getting evidence about this their whole lives and have likely become experts by now, so merely trying to hide that this will happen from them will not work. You’re going to have to actually become the sort of person who is ok deeply wanting something from someone and then finding out that they want absolutely nothing of the sort with you, over and over again.

This is a tough one, and I’ve fucked it up a few times, but there are some concrete things you can do to make progress on this.

One is that you can hire sex workers, either for online services or in person. It’s hard for a rejection to not hit hard when you are starved for sex. Purchasing sex work can help convince part of your brain that you are not in a pussy-impoverished period of your life. You are aiming for an abundance mindset, and it is very hard to have an abundance mindset when you are in fact deprived.

Another strategy I implement is already assuming that this person is not interested and being in more of a state like checking just in case. Almost everyone in the world is uninterested in having sex with me, so it is no surprise that this particular person also is. That’s the way it goes; on to the next one.

This does not mean that you should shoot yourself in the foot before you start. Don’t assume you have been rejected from the outset. And definitely don’t cut yourself off from your hope and desire. You should perhaps relate to it like how you might relate to playing a scratch-off. You’re definitely hoping to win money, there’s no doubt that’s the outcome you want, but you’re mostly expecting that you won’t. If you do win money, that’s great! If you don’t, well, that was always the default outcome. It would be weird if you played a scratch-off and then were like, “I didn’t wanna win anyway.” It’s not any less weird when it comes to hitting on women.

Rule 3: Escalate skillfully

I do not recommend that you start DMing all of your female friends and saying, “I want to fuck you.” I recommend that you definitely do not do that. However, DMing all of your female friends and saying something like, “Hey, I would like to take you out on a date if that’s something you’d be interested in,” might be half reasonable.

The best metaphor I can come up with here is imagining that you are in a race with someone, and you definitely want to win, but you want the final snapshot at the finish line to be dramatic, so you don’t want to beat them too badly. You should be stepping on the gas more than they are, that’s exciting. But don’t race laps around them; that’s unsportsmanlike and nobody likes it. Of course, that means you need to pay attention to how much they’re accelerating, because you want to stay just a little bit ahead of them. Women are different from each other, and some of them will step on the gas much more than others. Pay attention.

Rule 4: Have a decent model of what it’s like being a straight woman dealing with men’s sexuality

A friend of mine (who I will later ask if she wants credit and, if so, mention here3) gave me an incredible metaphor for this.

Imagine that the human species is quite different. In your world, there are no human women, but there are still basically two human sexes. There are human males like yourself, and then silverback gorillas. Now, all of the silverback gorillas really want to fuck you. They are all much stronger than you and could easily kill or permanently injure you if they wanted to. And you barely find any of them attractive. Their impulse control varies, but probably at least a few of them have sexually harassed you at some point. You know that many of them are nice and willing to protect you and give you resources in various situations, expecting little in return, but again, you aren’t that into many of them, if any.

You have been rejecting gorillas your whole adult life. Sometimes they get angry at you, sometimes they cry, sometimes they stop being your friend, sometimes they decide not to hire you or tell others not to hire you. Now imagine that you are on a bus home after a long day of work, wearing your headphones, and one of these fucking gorillas sits next to you and keeps trying to get your attention even though you’re wearing headphones. How do you think you’d react?

This metaphor has been invaluable to me. You should always ask yourself: If I were me in this gorilla world, how would I want a gorilla to approach me? In what way could they approach me that wouldn’t make me wish I had stayed in today?

Rule 5: Lose the shame over being a gorilla

I suspect some men will feel some shame having read the previous metaphor. “Oh damn, I am a dangerous pushy gorilla that nobody is into.” Uhh, yep, sorry bud. Best to swallow that pill here and now. Go grieve for a bit if you’re into that sort of thing, but then come back here and read the next paragraph.

Done? Ok, so now how are you going to deal with that? The gorillas in gorilla world are also in a pretty tough spot. Truth is, most of them will only derive any meaning in their lives from their ability to court and live with a man, and men are barely ever attracted to them. The good gorillas have it particularly rough. It’s not their fault that the human men are afraid of them, and yet it’s reasonable for them to be.

You could react to this by cutting yourself off from your desire and sexuality. I don’t think that helps anyone. I recommend owning it instead. What would it take for you to see this with absolutely clear eyes, and still want things from women without any shame?

Whenever I initially hit on a woman, there’s some chance that I will make her uncomfortable. I want to minimize that chance and minimize the intensity of the discomfort, but I have to be willing to take some amount of risk if I am not going to be paralyzed for the rest of my life. I’m never going to be able to get that risk down to zero4, and I’m not willing to simply never have sex again, sorry. Luckily for us, most women do not want to live in a world where they never get hit on again, so let’s help them out by not letting that happen.

If you follow my advice, you will at some point make some woman uncomfortable. That’s unfortunate, and to be avoided and minimized, but it’s a cost you are going to have to live with.

When you eventually do, listen carefully to any feedback they might be willing to give you and make it as easy as possible for them to give you that feedback. Learn from the experience and try to minimize the chances and costliness of similar future fuckups but accept now that it is going to happen sometimes.

It helps here to again visualize the gorilla world. Imagine that you were her, and you were a gorilla: would you endorse a gorilla escalating in the way that you are about to, given your expectations about the costliness of her rejecting you and the expected benefits of finding that you are both into each other? If not, either figure out some way to change the tradeoffs, or skip this one for now and come back to it if circumstances change.

Rule 6: Use non-flirting

Most women (perhaps most people in general, I’m not sure) are into playing games and delaying common knowledge. I feel confused by what is going on here exactly—not sure what explains it—but I, for one, fucking hate it. I am not excited by “will they, won’t they”; I am excited by having sex. Most human flirting is even frustrating for me in the second person. I hate watching two people skillfully avoid common knowledge that one of them is attracted to the other; it drives me fucking nuts. It seems like they are leaving so much value on the table to me, exploring different possible arrangements that they could both benefit from, but whatever. If you’re into flirting and will-they-won’t-they, then great, you should try it, but I have no advice for you.

If you, like me, hate delaying common knowledge, I have good news for you. You can unilaterally make common knowledge with someone that you find them attractive. How? Tell them you find them attractive. They now know that you find them attractive, and they know that you know that they know that you find them attractive. There’s no way out now.

I once attended a party where you were required to flirt the whole time you were in attendance. There was some spirited debate about whether what I was doing counted as flirting.

If you have a crush on someone, you can just tell them. Even better if you can be more specific. Do you think about them a lot? Do you fantasize about a life with them? Do you find them incredibly hot? You can tell them. The only thing you are protecting by delaying telling them is potential, and sacrificing a chance at a real good thing in exchange for a sure fantasy seems like it is often a bad trade to me.

(See rule 2.)

Rule 7: As clichéd as it is, play the numbers

I have noticed that approximately never has a woman initiated with me. I thought for a long time that that was because I was unattractive. That may be part of the story, but a couple partners have now pointed out another mechanism to me. Women never initiate with me because I never give them an opportunity to. By the time they would have suggested that we hang out one-on-one or whatever, I have already asked them out on a date and probably said something about how beautiful they are.

If you want to sleep with many women, you’ll have to let them know that you find them attractive. Every time you escalate with a woman by letting her know that you find her attractive, you are taking on some risk of her having a bad time as a result, so you do want to figure out ways to avoid that, but you’re going to have to ask a lot of women out. In part because each time you escalate you are also taking on some chance of success, but also because you are going to need some practice being rejected.

Practice makes perfect, and exposure makes chillness.

(See rule 4.)

At one point, not that long ago, I was asked to speak on a panel for people with body counts over one hundred. At one point I was asked some question I don’t remember, but I asked all of the women in the room to raise their hand if I had ever hit on them. Approximately every woman I had ever met before that day raised their arm. I then asked them to put their hands down if we had never gone on a date, and then again if we had had never had sex (only if they wanted to reveal that of course). About one third of the women who had raised their hands still had their hands raised. If you want to know how I got to a body count over one hundred, that’s how.

Rule 8: Provide some kind of value on dates

If you have been following my other advice, you may have found yourself in a new situation. You have a date scheduled. Great, don’t panic! All you have to do now is be yourself while also not being the sort of person who is absolutely fucking insufferable on dates.

Do the following obvious things. First, you should buy dinner. Why? Dude, just buy dinner. Don’t expect anything in return. Don’t let her buy dinner. Just buy dinner; don’t make a big deal about it. This is an easy way to signal abundance. It provides some self-signaling as well.

Second, cultivate your genuine curiosity about her. This is an extremely complicated object in front of you, like one out of a billion topmost complicated objects in the universe. I realize that you are hoping to get laid tonight and are feeling distracted; that’s understandable, but you should still be able to muster some curiosity about how the person in front of you works. Show her that curiosity. Try to get your curiosity answered. Run experiments.

Finally, related to the last two points, provide some kind of value to her. I buy dinner and try to understand her. I’m also fairly funny, so I try to make her laugh. You don’t have to go for these things exactly, but you should have some model of what makes you fun to go out on a date with and make sure that you exploit those qualities of yours. Your goal here is to make sure that she does not regret seeing you. The best way to do that is to make sure that going on this date turns out to be worth her time.

Rule 9: Sleep with reasonable people, preferably the horny ones

Sluts are great. They have sex with us, and for that we should honor them. But I have noticed that sociosexuality does seem to correlate with personality disorders. I recommend avoiding this, although I do not recommend avoiding sluts altogether. Sluts have the benefit of having had some experience with dating and being hit on. They are experienced in rejecting men and dealing with the kinds of fallout that can sometimes happen as a result of sex. And besides, if you are going to sleep with many women, you are probably going to end up sleeping with sluts, for the same reasons that most of your friends tend to have more friends than average.

Determining the reasonableness of a female slut is an essential life skill for male sluts. It is luckily not that hard. You can ask about their past partners and take care to notice how they talk about them. Everybody has at least one ex they hate, but does she hate all of her exes? Does she hate all men and have no appreciation for their plights? Maybe you can skip this one in that case. Generally, if it seems like conflicts with her tend to be especially unpleasant or public social battles, proceed with caution.

Choosing to exclusively sleep with reasonable people is the only guaranteed way to avoid being cancelled when you do not deserve to be (it’s like the abstinence-only method of not getting cancelled). I, in fact, don’t even recommend initiating escalation with a woman unless you already have decent reason to believe that she is reasonable.

Rule 10: Don’t hold out for a cheat code, do it the hard way

Human advice has a peculiar way of not working. I find myself often giving advice to people like “You should ask her out, be yourself,” and then seeing them go out and do it, and then they so totally, face-palmingly blow it that I have to wonder if I fucked up giving them the advice. I think what is going on here is that there is a lot of know-how (as opposed to know-that) to be communicated in this domain, and that is famously difficult to communicate through advice. It’s like trying to write advice on how to play a Bach fugue: “Ok, first put your left pinky finger on the white key next to the fourth group of two black keys…” You’re only going to get it through practice.

It’s true that reading about how to play a fugue can help you avoid some pitfalls before you start practicing (I wouldn’t be writing this if I didn’t think it could be useful), but really, you will have to practice if you want to get good.

So, don’t worry too much about following my exact rules. You have to figure this stuff out for yourself anyway, and some of it probably won’t click until after you have had some experience.

There isn’t some simple set of rules that you can follow that’s guaranteed to work. I didn’t learn these rules by reading someone else write about them. I figured them out through getting out there and trying to get laid while also trying to not be an asshole.

Don’t be too hard on yourself if things don’t go well initially. Dating is hard. Male and female sexuality are messy. Talk to the people in your life about it and treat it like a hard skill you want to master like any other.

I hope you have some success.

Subscribe now

In a reply to my previous post, it came up that there’s a common intuition that, to justifiably make certain kinds of claims—generally claims that do not contain a negation—one has to provide certain kinds of evidence, whereas to justifiably assert the negation of such claims, one does not need provide any evidence. The slogan version of this intuition often goes something like, “the burden of proof is on the person making the claim”. There is something to this norm, but making sense of it in a Bayesian framework also presents some interesting puzzles.1 I will try here to provide something of a Bayesian justification for the intuition, as well as to make vivid the annoying persistence of the puzzle of accommodating it in a Bayesian framework.

For instance, consider the following dialogue:

Alex: I can fly.

Billie: Do you have any evidence of that?

Alex: I don’t need to prove anything to you.

Billie: I am not going to believe that you can fly until you show me some evidence.

It seems like Billie is being perfectly reasonable here. I, too, would not believe that someone can fly unless they showed me some extremely convincing evidence to the contrary. Now imagine that Alex does not take kindly to this, and the dialogue continues:

Alex: Why not? Do you have any evidence that I cannot fly?

Billie: Well, no, but it sort of seems like you should have to present evidence here before I believe you.

Alex: Boy, you are so naive. Didn’t you read Brangus’s post yesterday? Yeah, sure Billie, I have some evidence for you right here in my pocket.

It does not seem like Alex is being perfectly reasonable here. Why not? As I said yesterday, the only half-decent theory of evidence I have is Bayesianism. The Bayesian formalism does not distinguish between A and ~A. You could rename all the negated variables to positive variables and vice versa and nothing would change. What is up here?

First of all, Billie would have been better off if he had read my post yesterday, because he does, in fact, have a ton of evidence that Alex can’t fly; it just isn’t all that easy to package into a single observational report. He’s spent his whole life around humans and never known any to fly. Alex’s arms just look like normal human arms, and Billie is pretty sure that creatures that can fly generally have something that looks like wings instead. And so on and so forth.

But suppose that Alex were instead making a claim that was more deliberately designed to be independent of observation, for example, “I have the invisible dragon from Carl Sagan’s fable in my garage”. Billie would at least have a harder time producing evidence against that claim, and, in any case, it seems like there’s something right about Alex’s insistence that Billie is the one who should have to produce evidence.

Luckily for us, the Bayesian formalism gives us an exact way to measure amounts of evidence, as well as the amount of evidence required to get from, say, 1% probability assigned to a hypothesis to 99% probability assigned to that same hypothesis. Briefly, to get from a probability of 1% to 99% assigned to a hypothesis, you must observe something that is 9,801 times more likely to happen if the hypothesis is true than if it is false.

The question then of who has to provide how much evidence to who to has an exact answer. If Alex assigns a probability of 99% to the claim that he can fly and thinks that Billie should assign at least the same probability, and Billie says that he currently assigns 1% probability, then Alex just has to show Billie something that is at least 9,801 times more likely if he can fly than if he can’t. And of course, the same goes for Billie. If Billie assigns 1% probability that Alex can fly and thinks that Billie should assign no more than that to the claim, then he only has to show Alex something that is 9,801 times more likely if Alex cannot fly than if he can. Problem solved, no?

You might here want to complain that this retains none of the intuition we sought to recover at the outset. It does not make sense that the amount of evidence that Billie should have to provide to convince Alex is the same as the amount of evidence that Alex has to provide in order to convince Billie. Are they really on equal footing?

Part of the math that underlies the reasonableness of this intuition is that negations are generally logically equivalent to very large disjunctions. Take the claim: “The cat is on the mat.” Compare this to “The cat is not on the mat.” The second is equivalent to the very long (possibly infinite) disjunction that reads something like: “The cat is on another mat, or the cat is on the grass outside our house, or the cat is in orbit around Jupiter, or the cat is on the floor next to the mat, etc.”

If you start out with very little information about where the cat is, a reasonable strategy is to assign approximately the same probability mass to every place that the cat might be (or a nearly flat pdf over them if there are an infinite number of places the cat might be). You would then find that the negation of any specific claim about where the cat is has much more probability than the specific claim, since its probability is the sum of the probabilities of every other possibility. So, unless you have a reason for believing one of the specific claims about where the cat is, you should generally assign more probability to the negation of any specific claim about the cat’s location than to the specific claim.

But what if Alex was just born with a very high probability assigned to some positive claim that is compatible with all of the evidence we will ever see, and what if this claim by its nature does not make any predictions that we could falsify or verify through observation or experiment? For instance, perhaps Alex believes in a deist god who created the world but who only picked the laws of physics and the starting condition of the universe and does not interfere with the universe in any way. This does not actually violate any theorem of probability theory, but it sure seems like we should be able to levy some kind of epistemic criticism against Alex’s views here.

Ultimately, what we want here is some way to be able to criticize an agent’s prior. I hinted before at one way that we might do this. It would be strange for Alex to start out with such a low entropy distribution, but of course, the entropy of a distribution depends on your partition over possibility space, so unless we have some theory of how one should partition possibility spaces in general, we are going to end up in a similar spot arguing with Alex about what partition he should be using.

What we want is a full theory of what kinds of priors are rational. There has been some progress towards a right answer according to me, but it gets fairly complicated fairly quickly. I would rate the philosophical progress on the question of what the right prior to start off with is as a solid C-. There has been some progress but ultimately some of the results are unsatisfying and there are many things we might want out of such a theory that the best theories available do not get us.

I will leave you with this. In order to vindicate the intuition that positive claims at least generally require more evidence than their negations to be justifiably asserted, we must somehow figure out some way of criticizing some priors, e.g., priors that start out assigning very high probabilities to a deist god. We want to say: “Alex should only assign that high of a probability if he observed some evidence in favor of his claim” and we can only say that if we can criticize his prior in some way. In most practical cases, this does not come up, because people actually seem to have fairly similar intuitions about what priors they should have, but there are real world cases where it does come up, e.g., the debate between people who are into the MWI interpretation vs the Bohemian interpretation of quantum mechanics. That C- philosophical progress on the question of which priors are rational does actually help some on the debate between Bohm and MWI fans, I will make sure to write about how sometime.

Subscribe now

There’s a bunch of common constructions that go something like “Do you have any evidence for that claim?” or “What’s your evidence?” I mostly think this construction sucks. I’ll explain why.

I have only one half-decent theory of what it means for one thing to be evidence of another. Roughly: A is evidence for B just in case P(B | A) > P(B).

That is, A is evidence for B just in case learning that A is true would make you assign a higher probability to B. To have evidence for B, then, is just for there to be some proposition A, such that you know A, and P(B | A) > P(B).

There’s a famous problem with that phrase “the probability of” which is that it makes it sound like probabilities are somehow quantities in the world. Maybe there are some cases where it makes sense to talk about objective probabilities that are really out there and independent of anyone’s beliefs, but in this case that would be a pretty wild interpretation. We’re talking about what it means for one thing to make another thing seem more likely if it turns out true, so we’re definitely talking about stuff going on in the mind of some creature.

So the question “Is there evidence for B?” is actually a weird question. Evidence relative to who? Whose credences are we talking about? Who has to possess this evidence, or be in a position to possess it, for it to count as “existing”?

Is it evidence that the world does not spin that we are not flung off of it? Well, it was for Archimedes, but probably not for you.1 Why? Because you know a bunch of things Archimedes didn’t. What exactly do you know? Maybe not that much about gravity in detail, actually. But you know something about it. You know that people who do know a lot about gravity don’t find this argument persuasive, and they’re probably not all conspiring. You know that flat-earthers have a certain epistemically vicious vibe. You know a ton of similar, hard-to-articulate things about how modern physics and modern institutions work.

Imagine having to explain to Archimedes why the fact that we stay on the ground and do not fly up into the sky is not evidence in favor of the claim that the earth is stationary. How would you do? You’d probably end up needing to port over a whole worldview: physics, astronomy, institutional trust, past successful predictions, etc. It’s not just one clean little “piece of evidence.”

So if you challenge me to produce evidence of some claim for you, I will happily tell you how I came to believe the claim. But that story might not sound like it involves much “evidence” at all, depending on what you count as evidence and what your background is. Part of what annoys me about constructions like “Do you have any evidence for that?” is that they make it sound like communicating evidence of a claim should be easy, like: “Oh yeah, I have some evidence for it right here in my pocket. Now that I’ve shown it to you, and you can immediately tell that it’s evidence, you will surely update your beliefs.”

There is a good instinct behind asking someone whether they have evidence. It’s useful to find out why someone believes something. But I find those conversations far too often go something like this:

Alex: Do you have any evidence for P?
Billie: Uhh, well, let me think…
Alex: Why do you have to think before you remember your evidence for P?

Or:

Alex: Do you have any evidence for P?
Billie: Well, here’s my argument. Premise one…
Alex: Arguments aren’t evidence! What observation did you make?

Or:

Alex: Do you have any evidence for P?
Billie: Yeah, I ran a survey online and…
Alex: Was it a random sample? Was it published in a journal? If not, it’s not really evidence, is it?

Etc.

What I recommend instead, if you want to know why someone believes something so that you can figure out whether you should believe it too, is asking: “Why do you believe that?” or better: “How did you come to believe that?” This question makes more room for the possibility that reasonably coming to believe something might be a complicated process that draws on a lot of different observations, arguments, inferences, deferences, background models, vibes, etc.

Of course, I mostly doubt that anyone who is tempted to demand “evidence” in the specific way I’m criticizing will be very interested in this advice. I think a lot of the time what they’re really doing is making isolated demands for rigor: “Show me a gold-plated, journal-shaped justification for this particular claim I don’t like.” Prove me wrong.

In general, if your standard for believing something someone tells you is “Could they publish this in a top-tier journal?”, you’re going to have a very hard time in life. Not because you should believe things easily, but because the kind of work minds do to figure stuff out does not usually look like an RCT or an ANOVA.

I was tempted to quote a bunch of real “no evidence!” accusations I’ve seen in the wild and go after them one by one, but that seems too rude. Instead, here are some things I’m very confident of, that one might accuse me of having nO eViDEnCe FoR:

• There is no god.

• There is no afterlife.

• There are no ghosts.

• Santa Claus does not exist.

• There are no ontologically fundamental minds of any kind.

• Aliens have not visited Earth.

• Nobody has ever penetrated my anus while I was sleeping and then drugged me so that I didn’t wake up or feel it in the morning.

• The Copenhagen interpretation of quantum mechanics is not true.

• There is no invisible dragon in my garage.

• If you shine a laser pointer into the sky, that light will keep traveling for at least 10^20 years unless it interacts with something else. Notice that you cannot observe this happening.2

I do actually have evidence for all of these claims, but none of that evidence takes the form of a simple, standalone observation. In the case of aliens having never visited Earth, the main piece of evidence I would appeal to is that there is no nearby region of the galaxy where we fail to observe stars, and that even galactic voids still contain some galaxies. If that doesn’t immediately seem like evidence to you, that is exactly my point. If you ask me “do you have any evidence that aliens have never visited earth?” I might in some cases say something like “yeah, we basically see stars in every direction within the galaxy and galactic voids still tend to have galaxies”. I am much more likely to explain what that has to do with whether aliens have visited earth if you ask me something like “how did you form that belief?” Try it out.

Subscribe now

AI rendition 
of an oil painting 
of the public school I attended
burning down.
Based on a real photo.
Fuck that place 
forever.

My parents made the major mistake of teaching me that my body was my property.1 They taught me that independent-mindedness, questioning authority, and spirited debate are virtues. This may sound (to those readers old enough to remember it) like the banal liberal values of 1990s suburban America, but my parents were actually anticommunist political refugees from Cuba; they meant it more than most.

I was very young when I first started making them regret it, although I think they were mostly wrong to blame themselves, except perhaps to some extent genetically. I had diagnosed ODD, which contrary to popular belief I don’t think is a fake disorder. My guess is that it’s at least 50% heritable and also relatively predictable from a genome sequence, but I don’t have the time to look into that right now and I’m not confident.

My mom tells a story about me when I was five that I don’t think I really remember, even though it seems like I do. We were going to the beach, and she wanted me to wear sandals. I didn’t want to wear sandals because sand and the water felt weird on my feet. I asked why it had to be sandals; she said something like “because we are going to the beach.” I didn’t find that a satisfactory answer, and I believe we had some more back and forth. Eventually, exasperated, she said: “Look, Brangus. I’m wearing sandals to the beach, your dad is wearing sandals to the beach, your brother is wearing sandals to the beach.” I paused briefly and asked: “So, just because everyone else is doing something, that means I should do it too?”2 Mortified, she decided to put my sneakers on me instead.

One thing I really appreciated about my mom is that she was someone I could reason with—perhaps the only adult in my life that I fully trusted. That story does a good job exemplifying why: when she realized that she was using arguments that didn’t hew with the rest of her values, she not only stopped arguing, she conceded the conclusion and put the other pair of shoes on me.

I am tempted here to start going into my early childhood drug use, but that’s mostly best left as a story for another time. I do want to mention, however, to fully cement the impression that I really was just kind of a bad apple, that I learned to steal booze from my parents’ liquor cabinet around the age of eight. I knew that I was not supposed to; I suspect it wouldn’t have been as delightful if I didn’t know.

My trouble with school started as early as kindergarten. I had two complaints: the tasks we were asked to do were boring, and I felt like the other kids didn’t like me. I remember them picking on me for being physically slow. A strange thing to notice, but I think it was true. I was often the last to finish any activity—or I wouldn’t start the activity at all. I was the last to grab my backpack and be ready to be picked up by my parents. I took the longest in the bathroom when I had to go. And so on and so forth.

I was diagnosed with ADHD the next year.3 This led to my first conflict with adult authorities (besides my parents) that I really did not want to back down from. My doctors prescribed Ritalin (I don’t remember the dose) and told my mother and my school that I should take it every day. Unfortunately, I hated the way this particular drug made me feel. It gave me headaches, made me nauseous, and made it so that I stopped finding shapes in the texture of the ceiling interesting. I also didn’t like taking pills in general.

They tried various things. At first my mom tried opening the capsules and putting the powder inside of an Oreo cookie. This made the cookie taste extremely bitter. I eventually refused to take it at home, so they tried getting the school nurse to get me to take it. At first I tried not to swallow it by keeping it under my tongue and then spitting it out in the bathroom later. This worked for a while; I don’t remember how I was discovered, but eventually she started making sure to check under my tongue as well as the tongues of all the other children who were having a drug administered by the school. It was shortly thereafter that I decided this would have to stop.

I made all sorts of arguments to the nurse and to my mother—e.g., it makes me nauseous, it gives me a headache, I don’t like the way it makes me feel—but eventually I settled on: “It’s my body and I shouldn’t have to take a drug if I don’t want to.” Hearing this argument was quite flummoxing for many of the adults in my life. I didn’t understand why they reacted that way at the time, but I did have a sense that them reacting with that particular sort of stunned, confused expression was a good sign for me being able to eventually get what I want. That facial expression would soon become a guiding signal for my future crusades; cause the face and you have nearly won.

Similar strategies, however, worked far less well with my peers. I don’t want to be unfair to my bullies; I did in fact always delight in shocking people, and I was odd then in similar ways to how I am odd now, but also worse at masking and less charming. The intensity of my disgust reaction, then and now, is less than first percentile. I never really understood disgust and assumed that my peers and family were pretending or something. Bad smells never really smelled bad to me. It was only recently that I smelled something that actually bothered me, and it was the smell of long-decomposing, wet, dead rats. Actually, one of the other early conflicts that caused a lot of tension between me and my father was that I refused to shower. I found the sensation of getting into the shower and getting out of the shower very uncomfortable. I still do and have sometimes gone a month without showering.4 I’ve averaged about once every two weeks in my adult life. (It was here that I learned another strategy that will come up again later: going fully limp and becoming dead weight. It would often happen that my dad would yell at me to get in the shower, I would go dead weight, and then he would physically pick me up and put me in the shower.)

So, as I was an un-showered, low-disgust-response first grader with absolutely no instinct for conformity or punishing deviation from what seemed to me like totally alien norms, it is absolutely no wonder that the other kids picked on me. Early on, they bullied me because I wasn’t that great at playing with others. For instance, when, in approximately second grade, my peers started playing the game of “squank,” which is where all of the boys chase all of the girls and try to grab their butts and say “squank,” I refused to play. Another kid mentioned the game to me. I was confused and said something like “Why do you want to grab her butt? Do you like her?” He and the other kids did not take kindly to this. I remember another time that I got into an extended disagreement with a kid who would later be one of my biggest bullies (let’s call him Mitchel) where I insisted that girls did not have penises, and he insisted that they did have penises but they were on their butts and this was why they sat down to pee. (My parents were atheists and were very forthcoming when I asked them early on about how sex worked and such. The other kids’ parents were mostly not atheists or forthcoming about sex.)

This next incident perhaps deserves its own post. I taught my friends who would come over for playdates both how to access porn on our home internet dial-up connection5, as well as how to masturbate. We didn’t actually masturbate in front of each other, but I did explain it by showing them the mechanics on their thumb and explaining that if you did the same thing on your penis, it eventually feels really good. Friends were no longer allowed to come to my house for a while after that, and it took me a very long time to understand why. (I think this happened in about fourth grade.)

Earlier that same year, influenced by some of the anarchist teenagers I met at the hobbyist shop where I would play D&D and such, I had discovered dumpster diving. One day I noticed a giant garbage bag of about a hundred freshly made bagels had been put out in the dumpster behind one of my favorite delis. Excited, I brought the free bagels to the other kids who were playing in the park by our school. I expected them to rejoice. They reacted with some initial hesitation and confusion but did seem interested. A few of them took some bites, but then later found out that I had gotten these bagels from the dumpster and got extremely angry. This reaction seemed insane to me. You could tell by sight, smell, and taste that the bagels were perfectly fine. Child protective services would later pay a visit to my house, leaving my poor mother to have to explain that there was plenty of food in the house and that her son was insane.

In any case, I am trying to be detailed here because I want to be fair to my bullies. I was an irresistible target, but they were also quite brutal on a few occasions. The kid I got into a fight with about whether girls had penises at one point pushed my head into a hook intended for backpacks. I think he could have potentially permanently damaged me, but he only ended up hurting me really badly. This was in third grade. The next year, that same kid and his friends (who I would describe now as the white, popular, and Catholic crowd) jumped me in the basement of a church during a Cub Scout meeting and beat me pretty badly, badly enough that people wondered what had happened. I mostly refused to cooperate with authorities (I sort of thought we had an agreement to not rat to the adults, which I was clearly wrong about). Most of these folks have now at some point apologized to me, and I don’t hold any current grudges, but I did, at the time, want to smash their stupid fucking faces into asphalt with pieces of broken glass scattered on top.

Anyway, teachers and parents mostly didn’t believe that I was picked on, although they totally believed that something was up with me. Mitchel’s parents in particular claimed that actually I picked on him, which I could understand how they would think that, what with making him cry by telling him Santa doesn’t exist and that girls don’t have penises and tricking him into eating dumpster bagels and such. But I am fairly sure that that is not the description of the situation they would have settled on with full information.

Teachers were somewhat sympathetic to my bullies. They found me incredibly annoying as well, and also with good reason. I would refuse to do assignments, cheat on my homework or flat-out refuse to do it. I would simply refuse to attend gym class. In one particular incident (I think in early fifth grade) a teacher said that I was acting like a second grader. I said to her that actually I was being reasonable and she was acting like a second grader. This made her very mad, and was one of several dozen times that my parents had to come in for a meeting with me and the superintendent. The superintendent said to my mother that I can’t be disrespectful to teachers in that manner, and I replied before my mother could reply that teachers shouldn’t be disrespectful to me in that manner. I asked the superintendent why she held kids to a higher standard than her own teachers, which I could tell landed well because it produced that same shocked and confused expression on both her and my mother’s face. I was asked to apologize and I said I would apologize if she also apologized to me; neither of us ended up apologizing to the other.

The previous year, in fourth grade, I refused to stand for the Pledge of Allegiance. This was something we all did every morning. Interestingly, I don’t think anyone ever asked us to do it, and I was confused as to why we did it. To be clear, I was into America at the time (still am, relatively) but I didn’t love the whole “one country under god” bit. The first time I tried this, the teacher reacted quite badly, yelled at me, and sent me to the principal’s office. The principal decided to speak to me before calling my parents. He asked me why I didn’t want to stand for the Pledge of Allegiance. I said something about how I thought it was a better tribute to the values the pledge stands for to refuse to stand for it than to join everyone else in mindlessly reciting it. He later ended up calling my mother in for a meeting and tried to get us to agree that I had to stand for the pledge. I told him, in front of my mother, that I would contact a lawyer because I was pretty sure it was illegal for him to force me to stand. The delightful facial expression of hesitation and almost awe ran over both of their faces again, and that reward signal again let me know that I was doing great. I was in fact allowed not to stand for the pledge after that.

While this was going on, in about mid–fourth grade, I got into Marilyn Manson and Slipknot. I honestly do not remember how, other than through downloading their songs and music videos on Limewire (I also bought some of their CDs). I started dressing in all black (badly) and refusing to wear bright colors. This did not help with the bullying.

In late fourth grade, we were given an assignment to write a poem. I wrote what seems to me now like an incredibly embarrassing poem about killing myself. In it the knife was a key and other people were a prison or something, blah blah. (If my mom still has it somehow, I’ll make sure to add it in the footnotes someday.) I asked to read it out in front of the class, I think because I wanted the other kids to know that they were hurting me. I did, and this led to an interesting expression on my teacher’s face, but in this case, things did not end up going well for me. (It’s perhaps relevant that my parents got divorced around this time, fairly amicably.)

Needless to say, this poem landed me in the school therapist’s office. They tested my IQ to check if I was an idiot. It turns out that I was not, but I do believe that this was when I ended up with my ODD diagnosis. They classified me as “emotionally disturbed” and put me in special ed classes. These classes were small, and the teachers were actually in some ways better, more engaging, perhaps only because the classes were smaller, but my peers here were of course even more ruthless. From this point on it felt like I was “in the institution,” as it were.

The following year, in about mid–fifth grade, I started refusing to go to school. This was quite simple in the beginning. I simply stopped walking to school. At one point my mom and dad found out; my mom pleaded with me to no avail. My dad eventually did hit me for not going to school. I told him that meant I had to not go to school for another week, and that if he hit me again it would be longer. This enraged him and he started making threats, and so I made threats in turn. Eventually I was committed to simply never going to school again. I’m not sure how this information got to the police, but they had something called a truancy officer, who I believe was part of the normal police department. He started picking me up at my house and then dropping me off at school in a police car. I would deadweight when he arrived, he would try to intimidate me, notice it was totally pointless, and then physically pick me up and put me in his car while being noticeably upset. I would then immediately leave the school and walk back home. This happened for about a week.

Soon thereafter, a guidance counselor who had been assigned to me made a visit to my house accompanied by two agents from child protective services. They made it clear to me in front of my mom that if my parents could not manage to force me to go to school, that would be considered neglect, and that they would have to take me away from my parents and put me in foster care. I actually respected that they seemed to be negotiating with me directly rather than trying to negotiate with me through my mom, but this did also absolutely enrage me. I felt at the time that there was no way other people had the right to force me to be somewhere against my will for a third of every day. If it were true that they did, then I was not a free person. It was also around this time that I came to fancy myself something of a civic activist on behalf of kids.

I didn’t want to be taken away from my mom, so I started walking to school and then, once in school and accounted for in attendance, refusing to attend classes. I didn’t enter a class for most of four months. There were one or two teachers I liked who were able to get me to come into their class by finding me in the hallway, but that was about it. I made sure to take standardized tests because it had been explained to me that I would have to repeat fifth grade otherwise. They again tried to use the police to intimidate me, but I asked point blank in front of some other teachers and such if they would also take me away from my parents if I refused to attend classes, and they didn’t answer.

So things went on like that for a while. In fifth grade, a particular teacher who was somewhat sympathetic to me talked to me and suggested that I should try attending an alternative school. I also asked my mom to homeschool me and let me design my own curriculum. She looked into it and found out that I would still have to take standardized tests and pass them and such, and she didn’t trust me or herself to make sure that happened (she had a full-time job) and so she made it clear to me that that was off the table. Teachers, my dad, and other adults also advised her against this. My mom also considered putting me in a Montessori school and other alternative schools.

I was not quite sure why, but I didn’t want this to happen yet. I hated that fucking school. The idea of rerolling on another one seemed to make sense, but it felt like giving up. Eventually I figured out that what I really wanted was to finally gain the approval of my peers. My mom wanted to put me in another school because she could tell that I hated it and that the other kids picked on me. I pleaded with her to give me one more year to try to win the other kids over and then I would go to another school. She eventually agreed, although she thought it was probably a bad idea.

To my surprise now, this basically worked. In sixth grade6, those same other kids who had picked on me for being mean to teachers or skipping class or refusing to finish assignments now seemed to look up to me for it. In particular, some of the girls had now started also listening to Marilyn Manson and wearing all black, and they came to think I was cool. This included some boys and girls who were in the grade above me. That may be most of what made this plan work out, I’m not sure. It was also coincidentally, this same year that I met teenagers who did drugs at the mall, but I’ll cover that another time.

Having successfully won some degree of acceptance and popularity among my peers, I agreed to try an alternative school. This one went from seventh grade all the way to the end of high school, but only had forty or so kids in attendance total. It is hard to describe how hilarious that place was. It was fairly wholesome in some ways. Many of the teachers seemed like they themselves had maybe been fuckups when they were growing up, and some of them were incredible teachers. They had a particularly good music teacher who got me into jazz and 20th-century art music and stuff.

One kid at my school got arrested, and presumably imprisoned because I never saw him again, for giving a teacher a stick of gum laced with LSD. Another kid got drunk in the school and then died in a driving accident right outside of the school. A vigil was held for him the next day.

They also had smoke breaks twice a day, which some of the teachers would join. This was a great way for me to bond with the teachers, and most of them didn’t seem to mind very much if I skipped classes. Some of the students there would do things like smoke crack in the bathroom or have sex in the bathroom or get drunk, etc., so if I was just not attending class but mostly entertaining myself and staying out of trouble, they mostly didn’t bother me. Sometimes they would let me spend all day in the music classroom playing with all of the different instruments, and that was great.

Unfortunately, however, the commute to this place was quite far, about fifty minutes in a bus every morning, picking me up at about 7:30 AM, and I still felt like it was a waste of my time. It didn’t seem to me like graduating from this place would be much better for me than dropping out and getting a GED. I brought this up to my mom; she initially thought it was a bad idea, but I asked her to get me a tutor for the GED specifically, and to let me spend some time studying on my own. I had always been a bit of an autodidact, but from the period of sixteen to eighteen (when I got my GED) I learned perhaps the most that I have ever learned in a two-year period in my life. That was when I got into philosophy, that was when I read Gödel, Escher, Bach and The Selfish Gene and taught myself first-order logic and some basic programming. It was an awesome time.

I did get my GED when I turned 18, and I then went to community college. I was fairly hesitant about this, but it turns out that at community college, people treat you like an adult, and that was an incredible surprise for me. I ended up getting an associate’s with a 4.0 GPA. I later transferred to a decent school, twice actually, and ended up getting a bachelor’s in philosophy with a 3.85 GPA or so, mostly because some of the required classes, like “The Science of Water,” were total bullshit I hated having to attend. If anything, the fancier colleges had a lot more bullshit than the community college from my point of view, although I also had to admit that they had some much better and cooler teachers as well.

It turned out I was good at arguing and writing philosophy papers, and my professors thought it was reasonable for me to go to grad school while they advised most of their students not to try for grad school. I wrote a good writing sample on the sleeping beauty problem and functional decision theory and got a 97th percentile combined score on the GREs the second time I took them (I failed to study for the test the first time I took it). I got into a top philosophy PhD program straight out of undergrad. I won’t say which exactly, but I will mention that it is consistently rated as one of the top three programs in the English-speaking world. I didn’t finish the PhD because better opportunities came up, and I got sort of cynical about the discipline when I saw how the academic sausage was made.

I now have a cool job, which is going fairly well all things considered, and have been part of several impressive projects which I am proud of. I also contributed to those projects centrally. I am well paid and have decent savings for retirement, especially for the field I work in, and I do work that I think is important and feels meaningful to me. So things basically turned out fine from my perspective, if not surprisingly well.

I will perhaps one day write more reflections on all of this—how it affected my life, how I feel about it, what I learned, etc.—but for now I will leave you with this: The basic lesson I took from this is that I was right most of the time and the adults around me were wrong most of the time, at least in my case. It was actually fine for me to try drugs. It would have been far better if I had been allowed to teach myself and study for standardized tests on my own like I had asked. I should not have been forced to spend all day surrounded by idiots doing trivial tasks. Fuck that school forever. The only adult who I felt like was on my side—who seemed to believe at all that I was a thing that has a mind inside of it that works in reasonable ways—was my mom. So if you are perhaps dealing with an ODD kid yourself (a fate I would wish on almost no one) my recommendation is that you try to talk to them as if they are reasonable. They maybe are not all that reasonable, but they do have a mind, and that mind works in some way, and their complaints about what they are being forced to do might be more reasonable than you are giving them credit for (especially given how convenient it would be for you if they weren’t reasonable). They are probably worse at arguing than you. Maybe you should see what happens if you try to help them construct the best version of their argument that you can muster, and then force yourself to have to argue with that. They might well still be wrong, or be missing that there are constraints and costs and benefits that you can see clearly because of your experience, but at the very least, make sure that you hear what their complaints are, and take them seriously.

Subscribe now

There’s a broad category of cognitive technologies1 that seem to serve a similar function to deontological imperatives, e.g., “thou shall not steal.” That function, I claim, is something like stabilizing preferences or policies over time as different urges get activated by the environment. Having such stabilization can be extremely practically valuable for agents such as ourselves, who have to deal with all sorts of urges coming up that are counter to our long-term projects. Interestingly, for many—perhaps for most—of these injunction-like cog-techs, having clear insight into their function or purpose makes it much harder to continue to employ them effectively. Insight into their workings undermines their working. This suggests a puzzle: can you effectively employ injunctions in your own life without having to avoid looking clearly at how those injunctions function? I claim that you can, but that it requires a deep appreciation of your own stupidity coupled with something like stubbornness.2

In this post I will give a few examples of injunction-like cog-techs, explain how insight into those cog-techs undermines their ability to function, and say something about how you might rescue some of their usefulness while retaining clear insight into how they function. I will also provide a generalized just-so story for how we ended up with these injunction-like cog-techs.

I felt I needed to invent this clunky “injunction-like cog-tech” phrase because I wanted to make sure that you did not understand me to be talking only about normative imperatives of the sort that philosophers most often talk about. I mean to include any cognitive technology that helps you stabilize your policies or preferences over time. That said, I do think that standard moral imperatives make a decent paradigm example of injunction-like cog-techs.

An interesting aspect of moral-speak as most often practiced by everyday people is that the imperative mood is actually quite rare. It is rare for folks to say “you shall not steal” or “you must not steal,” and much more common for folks to say “stealing is wrong.” The fact that much moral-speak happens in the indicative mood is something a lot of philosophers have commented on and appropriately treated as a puzzle. On a super-naive semantics of a phrase like “stealing is wrong,” the meaning of the phrase is something like “there are some events that count as instances of stealing, and those events have the property of being wrong,” where “being wrong” is some weird property that makes events inherently to-be-avoided regardless of what your preferences or beliefs are. It has been noticed by many that this is a very weird property to try to make sense of on any broadly natural account of the world.

I won’t go into the various ways one might try to give an alternative account of the semantics of normative indicatives here. I only bring this up to point out that it is a central property of the way that people do moral talk that they express their moral judgments as statements, and that the truth or falsity of these statements does not depend on the attitudes of any agent, including the one who is assessing that judgment. Moral relativism and moral antirealism are popular even among folks these days, and I don’t want to claim that this is the only way that people use moral-speak, but it is a prototypical use of moral-speak that all native speakers are fluent in.

Why then would we end up with this weird feature of moral-speak? What purpose would this serve? One purpose that it does seem to serve quite often is to stabilize our policies or preferences across time. Consider someone who believes that it is morally wrong to steal—that is, they believe that stealing is to be avoided no matter what their preferences are, and perhaps also that if you do something that is morally wrong, that makes you a bad person. It seems like such a person would have an easier time not stealing, even if they were very tempted to in that moment. Compare this to someone who merely thinks that stealing is a bad idea most of the time, what with the people who get mad at you and maybe punish you if you get caught. Most of the time getting a free fish isn’t worth the risk. But of course sometimes the risk is low or the benefit is quite high, and in those cases, there is no reason not to steal.

So, sure, if you believe it is morally wrong to steal, you will have an easier time not stealing even when the risk is very low and the benefit is very great. But how could taking something like that on possibly be practically beneficial? If the risk is low and the benefit is great, that means you should do it! Duh! That’s like rule number one of being an agent.

To answer that question, let’s turn to a fairly different example of an injunction-like cog-tech: belief in a God who is everywhere and can see everything. Actually, any being or beings who are invisible and could be watching at any time would work about as well—spirits, ancestors, the animals, etc. How does such a belief help you stabilize your preferences or policies across time?

Well, people are much more likely to fuck over other people when they think they can get away with it (especially if the only reason that they have to not fuck over other people is a practical fear of punishment). Often (especially in the ancestral environment) much of someone’s judgment of whether they can get away with something depends on whether they think someone is watching them. So the extremely practical cognitive technology of believing that important judgmental beings with great powers are watching you at all times helps you not believe that you can get away with things that no other person will ever find out about. Again you might ask, how the hell is this practically beneficial to the agent with this belief? It makes sense why other people in the tribe might want to trick him into thinking that the ancestors are always watching, but how in the hell does that help him?

There are a few good answers here. One is that people discount future costs too much compared to immediate benefits.3 Imagine that you are a caveman who has the opportunity to fuck one of the alpha’s women. It seems like nobody is around. Well, the benefit of getting to fuck this very attractive woman is going to happen right now. The possibility that you get caught and punished is going to happen later.

A related reason is that parts of your environment trigger urges and general cognitive modes—e.g., horniness, hunger, etc.—which make certain kinds of costs and benefits more or less salient. When you see the attractive alpha female, and you’re really turned on by her, getting punished by the alpha later doesn’t weigh that much on your mind. You’re preoccupied by the hot babe being all seductive in front of you. I mean, there’s nobody around, and I’m pretty strong, how sure am I really that I couldn’t win in a fight with the alpha? I have some friends; I’m sure they’d back me up.

So the story here then is that through a mix of hyperbolic discounting, rationalizing, and stimuli-activated dispositions, we tend to underweight the chances and/or cost of being punished for a norm transgression, especially when we are in the moment of deciding whether to transgress. This is why injunction-like cog-techs are so mixed up with norms and norm enforcement: they are more practical in worlds where you have to deal with others establishing and especially enforcing norms.

It also goes further than this. It also helps you practically by helping you signal to others that you will follow a norm even when there is little chance of you being caught. Consider another very popular injunction-like cog-tech: feeling bad about things nobody will ever find out about.

This works quite well on our story. You do not need to worry about estimating the chances of getting caught or how bad it would be if you did get caught, or the fact that you might get caught much later than you would get the benefits. You already know that you will punish yourself quite harshly even if nobody else does. Heck, you are probably already punishing yourself now for even considering it.

This is also the sort of thing that other people can detect. If you know that someone is a guilty, scrupulous person, this makes them a far better candidate for guarding the food pile at night than someone who is merely smart and charismatic, especially if they seem kind of generally… loose, let’s say. It might sometimes work to merely lie about believing that it is wrong to steal food from the food pile, but it’s even better if you believe it yourself. Otherwise, when you are interviewed for the position, the interviewer might well be able to detect that you are lying, especially if they are an experienced interviewer.

So that then is our just-so story. Injunction-like cog-techs help us stabilize our preferences and policies over time. This is particularly helpful for beings that are bad at estimating risk, have steep discount rates, have different cognitive modes that get triggered by stimuli, and have to deal with norm enforcement from other beings in their environment.

The problem with this story is that applying it with clear eyes to the injunction-like cog-techs that we are currently implementing, and then investigating what those very cog-techs are good for, or how they got there, or how they work, tends to undermine our ability to continue using them. Let us go through a few examples.

This is easy enough to see in the case of believing in invisible beings who are watching you at all times. If the point of having that belief is to make it so that you do not ever feel like you aren’t being watched, and so it is as if the norm enforcers do not even have to bother catching you, it would require psychologically unrealistic amounts of compartmentalization to continue believing it. How did you actually end up believing the ancestors were always watching? Oh yeah, it was those norm enforcers who told you about it. They sure seemed real dead-set on making sure you ended up with that belief; that’s curious. Thinking about this clearly, it seems that the wool has been pulled over your eyes.

In the case of using the indicative to talk and reason about moral judgments, using our just-so story to analyze that cog-tech also seems to undermine it. Firstly, it is pretty mysterious what an utterance like “stealing is wrong” could possibly mean on any natural understanding of the world while not being either wildly improbable or devoid of most of its moral oomph. If it means “events that count as stealing are inherently to be avoided by all beings regardless of their attitudes,” that’s very difficult to make sense of in a naturalistic picture of the world. If it just means “stealing is punished around here” or “boo stealing” or “if you steal, we will fight you,” then it loses all of the practical import that I argued believing something like that has for an agent. So it seems like if you want to retain the practical value of believing that stealing is wrong, you are going to have to go with something like the first interpretation. But again, if you see clearly that the belief that stealing is wrong is some nearly nonsensical kludge of a belief that mostly serves you by making sure that you don’t steal when you shouldn’t because you are bad at estimating the costs and benefits of stealing, it sure seems hard to continue believing it in that case. And how was it again that you ended up with this belief? Oh yeah, it was those norm enforcers who were trying to get you to believe it. And oh, fuck, right, they sure did seem very motivated to make sure that you believed this, and they really didn’t like it if you asked them why this was true.

The case of punishing yourself internally turns out fairly similar. If the point of feeling guilty is to make sure that you yourself punish yourself ahead of time so you don’t risk misestimating the utility calculus on some norm violation, surely you would be better off just not punishing yourself internally and up-weighting your estimate of the risk of any given norm transgression you consider. If the point of it is to convince others that you will not violate a norm even when no one is looking, or that you won’t do it again, surely you could do this by committing harder, or by getting better at lying. Seems like that would be practically better. This particular injunction-like cog-tech, like all of them really, has a particular stickiness, but a clear view of it is still significantly undermining. Those of you who are unusually introspective might have at one point wondered whether you were feeling so unusually bad about some norm transgression because you were caught and wanted to signal remorse to those who were deciding your punishment. If you are even more introspective, you might have figured out that that was what you were doing, and decided to stop, perhaps out of a sense of duty or out of a sense of virtue.

So that’s that then. Injunction-like cog-techs don’t survive being seen clearly, and that which can be destroyed by the truth should be. Let’s just stop using them. I will follow the norms when they make sense and not when they do not. If I have to get better at estimating risk and benefit, so be it. That’s the way—not continuing to use twisty, non-sensical cognitive habits that were put there by forces actively trying to control me.

Well, as may be obvious by now, I think the problem with that is that these cog-techs serve an important purpose: stabilizing your preferences and policies across time. In the previous section, all of the examples of how an injunction-like cog-tech could be useful had something to do with avoiding transgressing a norm when it was practically unwise to. Those norms mostly served the practical aims of others (e.g., not getting stolen from, not having your harem sleep with other cavemen). Here let me give some rapid-fire examples of when injunction-like cog-techs are useful for projects that serve only or mostly the agent who has implemented them:

You want to work out because you want to be hotter or healthier. This requires going to the gym regularly, so you start feeling bad or doing negative self-talk when you miss a day at the gym. Alternatively, maybe you develop a belief that it is good to go to the gym. Or maybe you imagine that a hero is watching you. What would Arnold Schwarzenegger say about you skipping leg day?

Perhaps you want to quit smoking cigarettes. In order to do that you reconceptualize every individual decision to either smoke or not smoke a particular cigarette as the decision to either be a smoker or not be a smoker.

You are in a happy monogamous marriage and you want to make sure that your marriage does not fall apart. When you are in such a situation, you remind yourself of how much you care about your partner and think about what they would think of the situation you are in.

You have founded a startup and you want it to succeed. In order for it to succeed you have to work 10+ hours every day. In order to achieve this you build your identity as someone who ships, or someone who cares a lot about their career, enough that spending time away from the office makes you feel guilty.

You want to stop eating animals because you care about animals but you really love meat. In order to do this you cultivate a disgust response to meat in yourself. You keep watching videos of animals being slaughtered in gruesome ways and try to associate this with eating meat internally.

You want to write a blog post every day for an extended period…

It seems to me like a lot of the coolest things that humans have ever done could not have been accomplished by the humans involved unless they had some kind of injunction-like cognitive tech. Human decision making is too fickle on its own to accomplish that. You might think perhaps that a clear understanding of the injunction-like cog-techs mentioned above is less undermining. Maybe some, but not really enough. I’ll leave it as an exercise to the reader to think about how applying similar reasoning to our just-so story might undermine the adoption of the cog-techs above (in part because I sort of think you shouldn’t be exposed to that without explicit consent).

Worse than that: plausibly, a lot of the coolest things that humans have ever done have been done using injunction-like cog-techs that really were of the particularly insane-seeming kind, e.g., believing that you are always being watched by a very powerful invisible guy. It would be sad to completely lose the ability to stabilize our preferences across time and scenarios that these techs afforded us.

So, what do?

I have two prescriptions: one a hack, and the other, the real deal.

First the hack.

You might have noticed that a lot of these injunction-like cog-techs function like commitment devices. They tie you to the raft, as it were. In particular, they mostly are cognitive commitment devices, and most of their practical import (at least to the agents that adopt them, rather than to the agents that install them) comes from their function as commitment devices.

But not all commitment devices are cognitive! You can, for instance, give a trusted friend access to your bank account and ask them to take money from you if you do not lose a certain amount of weight by some date. One of my partners and I once agreed that if we got into a fight over some period, we would both have to donate to the Communist Party of the US. (This did end up working.) External commitment devices have the advantage of not requiring you to mess with your cognition in order to implement them. They do, however, have the disadvantage of requiring some overhead relative to cognitive commitment devices. That said, they are relatively underutilized in my opinion.4

Now for the real deal.

The answer is to adopt injunction-like cog-techs on purpose because you correctly think you are kind of stupid and corrupted. And actually, that’s not quite right, because it doesn’t take seriously enough how difficult and extraordinary it would be to be able to just have stable preferences across time without using injunction-like cog-techs. It would actually be very hard. So perhaps you should also have an appreciation for that. You should both know that you are very stupid and that it is ridiculously difficult to have stable preferences across time. In order to deal with this challenge you on purpose design your own injunction-like cog-techs and adopt them.

Have you noticed that you are more likely to believe something when it is convenient? Perhaps you should believe things less when it is convenient to believe them—but this will of course be more tempting when the problem comes up than it is now. It’s easy to think that you should believe things less when it would be convenient if they were true. It’s extremely difficult, when you come across something that seems definitely true, to notice that it would also be convenient if it were true and then believe it less. So what should you do?

Perhaps you could take on an identity as someone who doesn’t do that sort of thing. Or you could publicly vow to not do that. Or you could develop a sense of disgust for certain kinds of cognitive moves that help you believe convenient things. There are a lot of options.

We might reasonably ask why a clear vision of how this cog-tech works wouldn’t undermine it in the same way that the other cog-techs are undermined by insight. The answer is that I identified a feature of the stability of my policy across time that I did not like, and I decided to implement this tech for the sake of my own aims. It involves no false or non-sensical beliefs, and the reason for it being there flows from my own judgments about what policy practically benefits me most. As such I do not need to worry that there is some other way to better achieve my aims. It may be a sort of kludge, but it is, according to me, the best option available.

There is a genuine cost to this strategy. Injunction-like cog-techs are harder to disavow than beliefs. There is a genuine reduction in your flexibility here. But that is the whole point. Regardless, the injunction-like cog-techs I design and adopt to serve my own aims are much easier to disavow when the time comes than the injunction-like cog-techs that I was taught to adopt as a child by parents and busybodies.

Subscribe now

You might reasonably be wondering how exactly a thirteen-year-old boy managed to fuck a pair of sixteen-year-olds. Well, I was the sort of thirteen-year-old who had already tried multiple hallucinogens and smoked about a gram of marijuana daily.

I met Melanie through drug use, and I met people who were doing drugs by going to the mall. Like my brother, the other kids at my public school would pick on me. They were more brutal than my brother, sometimes beating me up. It was left to me to find a friend group, and I did so by approaching interesting looking people at the mall. At the time I was into Hot Topic’s aesthetics (which were a fair bit different two decades ago). I had long hair and luckily a bit of a goatee. People basically never guessed my age correctly, consistently guessing that I was seventeen or so. I was also, in some ways, fairly mature for my age.

Part of what was going on here is that I grew up with oppositional defiant disorder and I still have some bits of it. Basically you can model me as not feeling much anxiety or discomfort when I would “get in trouble”. I didn’t believe adults when they said that something was unwise or dangerous, and I didn’t particularly mind running the experiment myself, eg, trying drugs, sex, skipping school, etc. I was also not terribly fond of the norms that my peers would try to impose on me and not shy about expressing my distaste. When they bullied me about something, I would normally say “fuck you” or something equivalent rather than cave.

I wish it were just that I ignored unjust commands, but there’s another component to it. To this day when a perceived authority commands me to do something, I feel an overwhelming impulse to show them directly that they have no power over me. I bring this up because it is relevant to the decisions I made which led to me fucking my brother’s wife.

I was visiting my brother on a break from school and outpatient rehab (both of which I was forced to attend). (I would later end up in an inpatient rehab, but that’s a different story.) He’s my half-brother actually and much older than me. I believe at the time he was in his mid thirties.

He was a very persuasive person with few moral or psychological hangups. He was always bringing home girls and such when I was younger and tended to be the center of his friend groups. He was an extrovert and ended up in a sales role even though he was originally an engineer.

My brother was proud of me for having lost my virginity. I remember him celebrating this with his wife one evening: “he popped his cherry already”.

He would pick on me quite a bit during this trip. It was always fairly silly stuff, like making fun of me for dating a girl with a nipple piercing. (Aria did indeed have a nipple piercing.) He forced me to mow the lawn and cut my hair. In general, it seemed like he wanted me to conform so that I could get along better in life, and more importantly he wanted me to want to conform. Now, mind you, most authorities weren’t able to convince me to do anything even using violence and threats of imprisonment (which they did try) but my brother was incredibly persuasive and dominant. He was totally willing to use guilt, threats, bullying, and kindness to get me to do what he wanted. He was skilled, and violence wasn’t necessary.

I fucking hated this. I would impotently fantasize about showing him the error of his ways and the various kinds of violent revenge I might take. (Instead I would have to settle for… well, you know.)

I suspect that he was always a bit resentful of me because I was the first (and only) kid our dad had in the US. My brother was born in Cuba, where my parents met. They fled to Venezuela and later to the US in the late 1980s, and I was born shortly thereafter. My brother naturally had a much harder time growing up than I did. He grew up in the intense poverty of 70s Cuban communism. My father also spent some time in jail in Cuba for “counterrevolutionary activities” while my brother was still growing up. I don’t know the details, but I’m sure that didn’t help.

By contrast, I actually grew up in a relatively affluent part of the east coast near a major city. It was a nice suburb with a low crime rate. We were poor for our town, but my parents felt extremely rich. They were anchored on the extreme deprivation they experienced in Cuba. Worse, they aimed to make sure that I felt rich.

I don’t blame my brother for being resentful of me. I likely would have also felt resentful if my parents had devoted far more of their resources to my younger sibling. Nonetheless, I suspect that resentment made him somewhat more sadistic towards me when we were in conflict. He didn’t merely dislike my matted long hair, or want me to conform for my own sake. He also smiled while he forced me to lean into the tub and chopped off more than half of my hair. I remember him mentioning something about it being like washing a street dog and other creative shaming things. To his credit, the haircut came out looking pretty good. Ironically, his wife also seemed to think that it looked good. I don’t think she would have fucked me if he hadn’t cut my hair.

His wife was extremely hot. She was Argentinian I believe, and very fit. Dark hair, probably about 5’5”. Even at thirteen though I could tell that she was somewhat sad. She would at times talk to me about drugs, bragging about getting mescaline mailed to her by her friends. When my brother bragged on my behalf about losing my virginity, she gave me a high five. I will call her “Maris”.

I had been visiting them at this point for about a week, maybe a bit longer. It was the last night that I was staying there. I’d be returning home in the morning. I found a bottle of vodka and started drinking it. I knew I wouldn’t be tested for a few days, alcohol can’t be detected in your urine after a few days. His wife and I both smoked cigarettes. She found me smoking and noticed that I was drinking. She laughed and said I was such a character. She assured me she wouldn’t rat on me and joined me for a smoke.

I could tell that I was in, although I wasn’t sure exactly what I was in with. My brother, I presume, was already asleep. He would have to drive me to the airport early in the morning.

We talked and smoked. I don’t really remember what we talked about but I remember making her laugh a lot. She also started drinking wine coolers and drank maybe two or three before she brought up that she had some weed. She said she would share it with me but that she didn’t know how to roll it. I told her I couldn’t smoke since I would be tested in a few days, but I could roll it for her. I in fact didn’t really know how to roll joints. What I did was empty the tobacco out of a cigarette and use the antenna of my old Nokia phone to stuff the weed into the tube. Crude but effective.

At one point after she smoked weed she had me call someone that my brother was either flirting with or had hooked up with to sort of intimidate them or something? I wasn’t really sure what she wanted other than that she wanted me to say that I was “from New York”. I said something like “do not mess with my brother anymore, ok”. That lady’s husband later called back and was understandably upset.

In any case, at one point I remember making her laugh, again I don’t remember how, and she said: “we should step into the garage”. I didn’t really know what she meant at the time, and so I said “sure”. After we stepped in, she grabbed my wrist and then we had sex. I am hesitant to be too graphic for a variety of reasons, but it was not very good sex from my perspective. I didn’t orgasm, and she later had a bit of a freak out. She did some trembling and some crying. She couldn’t find her glasses, so I frantically helped her look for them.

I haven’t yet mentioned that my brother and his wife had a child. He was about maybe two at the time. I mention this now because Maris was freaking out in part because she was worried that this whole scenario would later lead to her being separated from her child—not entirely unfounded, it later turned out. She kept saying things like “they’re going to take my child away from me” as we looked frantically for her glasses.

After we found her glasses and she calmed down a bit, she did some rationalizing. She convinced herself to take a pill so that she wouldn’t remember anything. I believe it was Ambien, which I remember thinking was a crazy strategy at the time. Seems better to remember as much as you can so that you can reason about what to do in reaction to it.

She later invited me to have sex again, using the claim that she wouldn’t remember as part of her argument. She also mentioned that she would understand if I didn’t want to because I really loved my brother, but that otherwise we should do it. I did end up doing it again and this time turned out to be much better sex.

After the second time we had sex, she threatened me that if I ever told anyone about this… I don’t remember exactly what she said she’d do, but the vibe did suggest something violent. I considered this threat pretty impotent, but I had no intention of telling anyone what happened anyway and thought it was best to just let her go on believing that she had successfully threatened me.

The next morning I was back on the east coast and she called me. She apologized to me profusely and I replied with something like “don’t apologize, that was awesome, I don’t think you did me any harm”. I think that’s mostly correct. I don’t think that she did me any direct harm. It did strain my relationship with my brother for the coming years a fair amount, as I felt like I had this giant secret I had to keep from him, but our relationship wasn’t that great to begin with.

Things went on like this for about five years. The secret was well kept from my family, but I did tell some close friends about it. Maris never found this out. We would sometimes allude to it in a way that only me and Maris were aware of, eg, saying things like “all families have secrets” but that was it. Around when I turned 18, however, I got into a conflict with a close friend over some relationship drama, and he reacted quite badly to it. He threatened to tell my brother about what had happened. I really didn’t like him having this sort of control over me, so I decided to tell my brother about it myself.

I called him, and to his credit, he took it very well. I apologized. He said something like “don’t worry about that at all, you were a kid”. I sort of disagree with that perspective, but I do find the overall way he handled it very admirable. Unbeknownst to me, he was actually going through a divorce with her. He asked me if I would be willing to talk to his lawyers to help him get full custody over my nephew. I said yes.

I flew out to talk to his lawyers, it was an overall good trip. When I spoke to his lawyers, I was very straightforward. They asked me if I thought that she would do it again, and I said “I don’t know to be honest, my guess is no, but seems more likely than for a random person”. The lawyer said that this was a very honest answer. I never had to go to court, so I am assuming that the lawyers from both sides worked out some kind of deal. My brother did end up with custody, but would also let my nephew stay with his ex wife on weekends and some holidays.

In any case, I wanted to write about this for a few reasons. One is that I wanted to give information about an actual case of sex between a minor and an adult and how it went down. I think I have done that pretty well, working within various constraints, including respect for all parties involved.

The other reason I wanted to write about this is to say that I very much do not feel like I was a victim in this scenario, and it bothers me quite a lot when folks insist that I was. I claim this incident had very very little to do with any of my development. From my perspective, this was just a particularly messy and taboo thing that happened which had some unfortunate consequences, but not nearly as bad as my worst decision ever.1 The worst cost, which really was my cost to bear, was the guilt that I felt. It was of course an unwise and unethical decision on her part, and I’m not sure how to reason about that exactly, but I emphatically do not think that *I* was harmed.

I’m sure many folks will insist I was harmed by her, and that’s fine. I know I won’t convince most people. But do ask yourself, who do you think is in a better position to make that judgment, me or you? To be clear, this is not to make any claim about anyone else who is a survivor of child sexual abuse. I am very sure that harm can be done that way and most commonly is. The only claim I want to make is that I was not harmed. I understood what was happening. I didn’t understand all of the consequences all that well, but I understood them enough.

I could have decided not to have sex with my brother’s wife, and I didn’t. I decided to, in part because I was pissed at him. I should take responsibility for that, even if I was thirteen.

Subscribe now