Henry Stanley

The median take is taken

Sat, 11 Apr 2026 00:00:00 +0000

Vermeer, The Astronomer (1668)

Occasionally a friend will say something really stupid to me. (Obviously if you’re reading this and you are my friend, I’m talking about someone else.)

If it’s something that makes me raise an eyebrow, I’ll usually ask Claude about it, especially if it’s trivially checkable.

USER: A friend of mine says that “UK economic indicators are looking broadly healthy.” Is that true?
CLAUDE: I’d push back on that. The picture is more accurately described as weak with mounting risks.
Growth is anaemic — the economy grew just 0.1% in Q4 2025. Unemployment hit 5.2%, up from 4.4% a year ago, with youth unemployment at 16%. Inflation has come down but remains at 3.0%, and a Middle East energy shock is pushing petrol prices up and making further rate cuts unlikely. <more stuff excised but you get the idea>
The honest read is: an economy muddling through with real vulnerabilities, not a healthy one.

My friend and I had a brief back-and-forth, with his smoothbrained hot takes crumbling in the face of my machine-enhanced gigafacts, until I got bored and dropped it.

There used to be a website called Let Me Google That For You, which let you send a passive-aggressive link to someone that showed the Google homepage and a cursor wandering to the search box and typing in their query – which they could have searched for themselves.

It would be pretty fun if there was a version of that for Claude.

Well now there is! Try Let Me Claude That For You, the fastest way to annoy and alienate your closest friends.

This does present an uncomfortable question: if Claude can distil any legible question into a competent 250-word response in ten seconds, then what is non-fiction blogging actually for?

Specifically: what novel insight are we producing? Why should anyone write a lit-review-as-effortpost, or “here’s what the data says about X”, or a 3000 word synthesis of a subject?

The most obvious case where humans can still add value:

Posing questions that no-one had thought to ask. Why is my MacBook less responsive than my Apple II? What is the mechanism behind civilisation getting so many things predictably wrong? Is most of what we’re told about sleep bullshit?
Naming a pattern that people have only half-noticed. Reality has a surprising amount of detail. Meditations on Moloch. The Gervais Principle.
Not flinching from a premise. We should take wild animal suffering seriously. We should use biotechnology to destroy suffering. Factory farming is the greatest evil on earth.

LLMs are not good at doing these. They probably can’t discover a pattern that people have only half-noticed. While they aren’t trained to be opinionless, they’ll often come up with pretty milquetoast takes, and won’t want to push an argument to its logical conclusion (especially if that conclusion is controversial). They can’t make a boring-sounding question seem urgent, or make salient a pattern you’re not noticing.

That might change as models improve, but this isn’t a capability limit – these traits are baked in during post-training. Models are trained not to take a hard stance on things which are out of distribution, or on which there isn’t broad agreement. You might be able to push LLMs to produce novel framings of existing topics, but it’s still difficult to discover genuinely new questions that people aren’t asking.

Models trained to be agreeable will not e.g. come up with an ontology of corporate drones as sociopaths/clueless/losers. If anything, their post-training cuts the other way. Model makers want their models to be helpful assistants and sand off some of the rougher, weirder edges of their responses.

Let’s have a look at Reality has a surprising amount of detail by John Salvatier.

The setup: he starts writing about building a set of stairs with his dad. He discovered that it’s unexpectedly complicated in a way which you only see if you actually try to do this – the floor isn’t level, the stairs need a particular rise so you don’t fall down them, the screws need to be a particular length so they don’t stab you in the foot.

The pivot: reality is full of complexity which only becomes apparent when you slam your plan into the world. It’s hard to model ahead of time. This is the reason why projects often overrun.

It sticks in your head because the title is the whole thing. Once you’ve read it, you can simply use that sequence of words to retrieve the idea, and to cause the person to whom you’re speaking to retrieve it. You could ask Claude about the topic if you thought of it, but you probably didn’t, and its reply wouldn’t be nearly so salient. It helps that Salvatier actually had to build the stairs.

We could distil the above categories into something like:

noticing a question
naming a pattern
refusing to flinch from a premise

There are perfectly good reasons to want to write posts that aren’t just “they will provide insights to people for many years to come”. But if you do want to write posts which are lindy – which will outlast the latest fads or blogging cycles and produce novel and interesting insights which can’t just be gotten by asking Claude – you should aim at one of the above.

Even better – and this is where the idea of craft comes in, another thing that LLMs struggle with – they should be written to really stick in someone’s craw. In particular, the most important thing to aim for is compressibility. It should be easy for your reader to compress the central premise into a load-bearing phrase. Moloch. Civilisational inadequacy. Sociopaths, clueless, losers. Reality has a surprising amount of detail.

These become things by which your readers can refer to your idea – both a conceptual handle they can use to retrieve it in their minds, and for them to induce others who have read your work to retrieve it too. And for people who haven’t read your work, it serves as a way for them to make their way back to it.

So don’t write Claude-shaped posts. Give your readers a load-bearing phrase, ideally in the title. And do the work Claude can’t. Claude aims for the middle of the distribution, for centre mass – your job is to go for headshots.¹1. Thanks to Alexander Wales for this excellent phrase.

1. Thanks to Alexander Wales for this excellent phrase.

Blogging – a colossal waste of time?

Fri, 10 Apr 2026 00:00:00 +0000

Circle of Joseph Wright of Derby, Portrait of a scholar in his study (1839)

I was struck by a comment another Inkhaven participant made during one of our first days here: that perhaps blogging is just a waste of time, especially if you want your work to be enduring.

It feels hard to disagree with the notion that most blog posts are just noise, and that even exceptionally good blog posts mostly just get thrown onto the pile of other such blog posts and will be lost to time.

There are of course some exceptional bloggers, whose work will stand the test of time. Scott Alexander is the canonical example given here; Slate Star Codex articles are lore and their ideas have spread far and wide across the memeplex. His current blog has more than 100,000 subscribers, his posts are taught at Harvard, and his readers sent him enough crypto during the boom that he can fund a $200,000-a-year grants programme.

But by mainstream standards he’s still just a very online psychiatrist whose most famous moment was getting doxxed by the NYT. So what does it actually mean to “stand the test of time” as a blogger, compared to say an author of classical literature or even someone who writes textbooks?

Screaming into the void

I’ve heard some people here say that they don’t read any modern works of literature at all, which is pretty fucking funny if you’ve decided to go on a writing retreat for bloggers. But if you care about reading things which are lindy, then it makes sense that you probably shouldn’t read blogs at all – they’re probably among the forms of media which are the least enduring (arguably podcasts are worse, and I do think a lot of people fill their brains with slop by listening to podcasts every minute of the day).

Substack has the same creator dynamics as X or OnlyFans – where a tiny sliver of creators monopolise almost all the attention. The median number of X followers is 1 and the median OnlyFans creator makes $100 a month (it has a Gini coefficient of 0.85, higher than South Africa’s). The top ten authors on Substack collectively make $40 million a year, but only about 5–10% of Substack readers are paid subscribers – the vast majority of writers are giving away content for free.

Although there’s a lot of amazing content on Substack, inevitably most of it will be lost to the void. There are a lot of writers who are literally world-class who get basically no attention at all.

Inkhaven theory of change

Ben Pace talked about how “we all produced exceptional writing and I want to see more of it”. Which was very flattering, and motivating. But it’s not clear that simply writing more, and becoming better as a writer, helps you produce better insights.

The theory of change isn’t that the daily 500 words of slop we each vomit out at 23:59 is going to change the world, but that the program will create new, exceptional bloggers – “the next Scott Alexander or Gwern”.

This is a problem because even Scott Alexander isn’t the next Scott Alexander. It’s possible that Scott’s best work is behind him at this point – that he’s captured all the lindiest, lowest-hanging fruit. And I’m sure while Gwern will produce more great stuff seems unlikely that any of it will be as influential as his earlier work on modafinil, darknet marketplaces, and nicotine.

I do buy the idea that blogging more often is a way to become a better blogger, both through raw honing of the skill, and also because taking more shots on goal increases the chances that you produce a post which is a banger. I also think that writing every day, or at least regularly and often, is likely needed for this to happen for a particular type of blog post. I think Bentham’s Bulldog is a great example of this: lots of shots on goal (like 1000+ at last count), many psychotic takes, some excellent ones.

This is far from being the rule though. Brian Tomasik has dozens of posts on his website, and a great number of them are excellent and convey genuinely novel insights. It was his work which convinced me and many others to go vegetarian and take animal suffering seriously (and I often cite them in conversation).

Likewise, David Pearce in his day was incredibly prolific, and his writing on abolishing suffering with biotechnology put suffering reduction on the map for a great many people, including Tomasik.

Producing better insights

There are other bloggers who have also produced a relatively small amount of extremely insightful writing. I’m not sure the method of posting often as a way to improve your skill as a blogger really generalises to them, except insofar as it makes you a better writer, which is of course not the same as having better insights.

The skills that make a blogger good aren’t really writing skills at all: noticing things other people don’t, synthesising insights from across domains, having priors which turn out to be unusually well-calibrated. Writing well is necessary but not sufficient.

It does seem the case that writing more improves your thinking. Paul Graham talks about putting ideas into words as a mechanism not just for writing things down, but for thinking them through, for making them better and more complete.

That said – it seems hard to believe that simply writing more can produce Scott Alexander levels of insight. He reports that he essentially sits down at the keyboard and has a post flow out of his fingers, no draft, no substantial edits. He can one-shot blog posts that get hundreds of thousands of views. I’m suspicious that any amount of writing can get you to this point – more likely it comes from thinking obsessively for years, gathering interesting experiences from the outside world, and (the real blackpill) having a very specific type of brain.

Eaten by the machine

There’s another way in which blogs can be lindy: by being incorporated into LLM training data.

LLM base models (i.e. the raw model before it’s transformed into a helpful AI assistant by post-training) use huge text corpora as input, including the bulk of information written online. That includes your sick-ass blog! So by “writing for the LLMs”, as some call it, you can be guaranteed immortality in the form of residual tiny changes across their outputs.

The algorithms by which the ‘importance’ of different sources are determined are pretty opaque, although interestingly LLMs themselves do a pretty good job of labelling the importance of different sources. But Wikipedia is intentionally weighted higher as a high-quality source, and gets another boost because it’s mirrored and repeated across the web in various guises.

If it seems odd that an encyclopedia which anyone can edit is weighted so highly, consider that if your edits persist for any appreciable amount of time without being reverted or overwritten, those edits themselves are lindy. Lindy reinforcing lindy!

So if your intention is to have your ideas persist in the substrate of future AI, you might be better off editing Wikipedia than writing a blog post.

Don’t optimise for longevity

Of course it’s possible that one shouldn’t optimise for longevity. I would personally rather write about some weird fascination of mine, or a piece of personal non-fiction that moves other people, than edit some Wikipedia articles on the margin just so they get included in some LLM corpus.

Plausibly there are lots of other reasons to write: to refine your own thinking, to give people useful advice that might not apply in many years’ time, to move the Overton window on a given topic. Many pieces won’t stand the test of deep time but can still provide a huge amount of value and fun in the near-term, and it would be a shame if they didn’t get written.

The connection gap

Thu, 09 Apr 2026 00:00:00 +0000

Hopper, Room in New York (1932)

Extroverts are the best connectors in the room – as long as the room is filled with other extroverts.

For an extrovert, conversation is the path to connection. Conversation allows two people who feel comfortable externalizing their internal world to create a shared world extremely quickly.

Extroverts are more likely to initiate conversations and talk more and faster than introverts. They also enjoy conversation more. There are a number of ways this has been tested asking people directly, asking their peers, and by observing their behaviour, for example by recording snippets of their conversations (which is neat because we don’t even need to ask). These recordings show that extroverts say more words in daily life. Their mood is also lifted during social interactions beyond their tendency to just be happier in general (which they are as well).

EVELYN: Conversation, my partner once observed to me, is my special interest. From a very young age it was clear that I really, really want to talk. My mom has many stories about what this was like when I was growing up. Sometimes my desire to talk extended beyond her capacity to listen, and she would tell me that I could keep talking but she wasn’t going to listen anymore, and I would just keep talking. Every elementary school report card had the same notes: “She’s very smart, but she can’t stop talking during class.”

HENRY: I went from an introverted teen to a very extroverted adult. Conversation, social groups, making new connections – I wanted it all and it never seemed to be enough. English social circles are often hard to break into and connections ending up being dyadic, atomised; this led to a lot of loneliness. I can demand a lot from people in terms of their attention and connection, often more than they want!

As 90th percentile extroverts, the authors both seek out and enjoy conversation, sometimes without awareness of others’ need for quieter forms of communication. Sitting quietly with others tends not to be as satisfying of our need to connect with others; conversation tends to be the desired mode. Sometimes this comes at the expense of introverts’ desire to be left alone – whether that’s wanting to have dinner when they want some time to themselves, or feeling disconnected from communities that aren’t focused on regularly scheduled, energetic social interactions.

This is perhaps one of the biggest differences between introverts and extroverts. Not only do introverts not enjoy conversation nearly as much, they also find it difficult to express themselves through conversation. Their inner experience is simply more complex or ambiguous to “find the right words”. You’ve probably come across introverted people who tend to process more before they speak – they’re less able to articulate their inner world in real time.

Introverts also feel prone to feeling misunderstood more generally, that their social interactions don’t reflect who they really are. This might in part be because they’re basically acting to be more extroverted, which can feel inauthentic.

So extroverts, who want to connect with others, have a tool that works beautifully for doing so – as long as the person you’re talking to also feels capable of expressing themselves in conversation. You can think of this strategy for connection as high-bandwidth but only with other extroverts. When two people like talking: boom, instant connection. But when this strategy throttles a sense of connection for one party – or even makes them feel alienated – the favored tool of the extrovert becomes an obstacle to connection.

So then what? There are lots of strategies besides conversation that help people feel connected. Working on something in parallel or together, watching something, shared physical activities, even sitting together in silence can produce this effect. But, like love languages, bridging the gap between connection languages can be troublesome.

This leads to an asymmetry for extroverts: to the extent they are overly reliant on conversation to connect with others, they’re limited in their ability to connect with others who don’t get the same level of connection through conversation. Despite being empathetic and attuned to others they may be defaulting to a mode of interaction that promises connection but actually pushes people away. Their over-reliance on conversation to connect with others might mean they never fully develop the ability to connect in ways in which introverts prefer, accentuating the connection gap.

If your only tool for connection is conversation, you’re going to feel disconnected from about half the population – and you won’t know why, because you’ll be too busy talking to notice.

We want to test the hypothesis that extroverts connect through conversation more than other, non-verbal modes of being together.

We also want to test what happens to extroverts’ felt sense of connection when they are connecting with people for whom conversation is not the preferred mode of connection. How flexible are they in this regard? When interacting with someone for whom silent physical presence is the optimal tool, can they find a way to feel connected as well – or do they mirror and empathize with the other person but privately feel disconnected in the absence of conversation?

We’re imagining two studies:

The first is a survey, targeting the first question above. We will measure how well extroverted traits correlate with connecting through more verbal ways of being together (small talk, joking and riffing, heartfelt conversation, deep intellectual conversation, etc). Our hypothesis is that higher extroversion will be predictive of more feelings of connection in verbal modes of interaction, and that lower extroversion will be predictive of more feelings of connection in non-verbal modes of interaction.

The second study will be an experience-sampling study at Inkhaven! We’ll build a program that pings people randomly throughout the day to ask questions about the activity they’re doing, who they’re doing it with, and how connected they feel. Additionally we’ll measure each participant’s degree of intro/extroversion, and their preferred modes of connection.

We’re hoping to find evidence that extroverts do prefer verbal modes of connection, and our hypothesis is that they’ll find other kinds of shared activity less connective.

We’re also hoping that we’ll get at least some data of two participants interacting that we can use to measure the partner-report gap (the gap between the feeling of connectedness between introverts and extroverts during nonverbal interactions). This would be good evidence in support of our theory that extroverts don’t feel as connected as introverts during non-verbal interactions.

Finally, we’re hoping to gather information from some of those same two-person interactions that might point to our theory that extroverts aren’t actually tracking how connected an introvert feels during non-verbal interaction.

Together these two reflections on the experience of the extrovert would point towards a flaw at the heart of the extrovert’s perception: the inability to feel a sense of connection through non-verbal interaction, and the consequential discounting of the possibility that these interactions could be connective for others.

Inkhavening

Wed, 08 Apr 2026 00:00:00 +0000

It’s 4am and I’m lying awake. Can’t sleep. Walk out into the cold, perfumed air. Sliders crunching on astroturf. In the kitchen I down a Soylent Mint Chocolate Meal Replacement Drink and a Pamplemousse La Croix. It tastes like TV static. One person is still up, surrounded by Diet Coke cans, working on tomorrow’s post.

I check online. Three people still haven’t posted. It’s 11:56pm; I posted at 2pm. That either means I’m very disciplined or that I’m not really trying, that I should be plumbing the depths of my soul instead of writing about land value tax. I hope tonight’s the night someone gets kicked out.

Ben Pace announces that our kind sponsors wordpress dot com have provided swag – Moleskine notebooks and baseball caps, one per person. I take two of each and feel surprisingly little shame. Petty crimes in an age of abundance.

The heating here doesn’t use convection, it’s radiative. Click on the thermostat and a panel above your head beams infrared at you. So very clever, like so many other things here.

I was proud of my first post, an unusually squishy piece which appeared at my fingertips at 6 in the morning and was done by 7. I learn this is called “personal non-fiction”. It gets a lot of praise, and I’m a sucker for flattery, but it alienates two of my friends.

Everyone here is one of: smarter, more insightful, funnier, higher status, a better writer than me. Sometimes all of the above. We swim in the same water, I am familiar with the Sequences (haven’t read them though, that would be cringe) – but I am still an interloper from a poor European island which has seen zero wage growth in a decade. Meanwhile their Microsoft RSUs are worth a million dollars.

It feels very validating to have every need taken care of: food, snacks, a fridge full of fancy beverages, towel service, a cabinet full of toiletries. It makes me feel important. I try not to think about going back to Peckham.

People ask me about my timelines. I talk a good game but it feels like kayfabe – I’m not sure I really believe AI is going to kill us all. Or maybe I do believe it, but I just don’t really care?

Something about America’s bigness makes this country feel like the centre of the world. A kind of wastefulness that signals abundance. Wash your clothes in the massive oversized machine and then toss them in the gas-powered dryer. Back home we hang up our smalls one by one and the windows fog up.

I ask people what they were doing before this. Oh, I was getting my PhD in Legal History from Cambridge while being a champion boxer. Oh, not much, just being paid half a million dollars a year at a FAANG which I quit to pursue being a dating coach. Oh I was blogging a bit (I check and their Substack says “over 10,000 subscribers”). Why, what were you doing?

I take a break from campus life and swipe a bit on the apps. I’m happy and busy but maybe I could get a cute date, have some physical touch. Wrong. I remember the Bay Area’s brutal demographics; they do not work in my favour. I turn over and go back to sleep.

Should anyone be vegan?

Mon, 06 Apr 2026 00:00:00 +0000

Epistemic status: I might disagree with this in a month, not sure

As someone who is vegan, “should you be vegan” seems like a stupid question. Why the hell would you be vegan if you didn’t think it was a good idea?

Animal welfare charity evaluators initially started by pushing make-people-go-vegan interventions like vegan leafletting. That is, until the rather obvious fact that almost no-one goes vegan by reading a leaflet was discovered.

The movement then broadly pivoted away from direct interventions to campaigning to improve the lives of farmed animals on the margin – things like corporate campaigning to get egg producers to commit to not caging their hens, or lobbying governments to ban or phase out the cruelest practices that farmers used when raising animals for food.

On the surface, being vegan seems obviously not worth it. There’s a huge personal cost, and for pennies you can massively improve the lives of farmed animals through personal donations.¹1. It’s beyond the scope of this post, but one could argue that reducing one’s own meat intake isn’t fungible with donating to animal charities. Some of those charities will do things that are positive in expectation but don’t pan out – bringing a legal case that they lose, for example. You might prefer the certainty of not eating animals, although of course it’s possible that the difference in the number of animals killed for food is smaller than expected (or at least is again positive only in expectation), assuming that the supply of animal products isn’t perfectly elastic.

Going below zero

There’s a parallel with people who valiantly strive to reduce their carbon footprint by not flying or by leaving the central heating off: you can’t reduce your energy use below zero, whereas if you simply bought carbon offsets then the your bank account is the limit.

(This applies beyond the personal level – Britain is responsible for 0.8% of global CO2 emissions, so we can by definition only reduce global emissions by that much.)

The most beautiful case of this was the Guardian out-Guardianing itself by suggesting that, instead of flying places, you should go by container ship. It makes for a cute story and who knows, maybe it’s fun to spend two weeks on the Ever Given and hope it doesn’t get stuck in the Suez Canal. But here’s how it stacks up against flying:

The container ship + train option took 15 days and costs at minimum €1900 including offsets, and probably closer to €3500 if you use a realistic sleeper berth price
Flying takes 24 hours with one stopover and costs €330 (and emits about 0.4 tonnes of CO2)

So the grown up answer is: fly, and use the money left over to buy high-integrity carbon offsets and offset sixty tonnes of CO2 – enough for you and 150 other passengers on the plane.²2. I’ve found that some people immediately respond by saying “carbon offsetting doesn’t work/is a scam”, which is mostly motivated reasoning – if done well, carbon offsetting definitely works. And if for some reason you don’t want to believe that, you can instead donate to the Clean Air Task Force, which advances technologies and policies to reduce carbon emissions and is a top-rated charity in the space.

Offsetting your evil non-veganism

Going vegan doesn’t really make sense for the same reason: you can just not be vegan and instead spend your money on donations to effective animal charities. And even if you’re not willing to assign a dollar value to the personal cost of being vegan, you could take the extra money you’re spending on vegan food and then donate that to animal charities, and you’re almost certainly still ahead.

Money is the unit of caring and if you really want to make a difference, you give your money to people who can do that most effectively rather than socially signalling you are a good person by e.g. not flying.

Charity FarmKind offers ‘moral offsetting’ (although they don’t use the term): through their website you can donate to effective animal charities and reduce the suffering of farmed animals to offset the harm you cause by eating meat.

They suggest that for $26 a month you can offset the average American’s animal product consumption, plus buy offsets for the CO2 emitted from farming those animals. This is pretty cool: it helps animals, and it puts the idea of offsetting into the memeplex, which broadly seems a good thing (like carbon offsets, which are just robustly better than reducing your own carbon dioxide output). If people aren’t going to go vegan or vegetarian (they aren’t), this is a very good second best.

Given I’ve just written 500 words on why being vegan is stupid, you might think this is the post telling the world I’m no longer vegan. So why don’t I bite the bullet and take my own advice?

Spreading the meme is valuable

There’s some basic human trait that says: what you do is who you are, and changing your behaviour in a costly way tells others that you’re taking it seriously. People see you doing things that are clearly difficult and it makes them realise wow, this guy actually is putting his money where his mouth is.

Ultimately, the thing that made me go vegan was dating two vegan girls in quick succession.

Neither was critical of me being non-vegan. And neither attempted to make me go vegan, or even suggested it. They were just there, being angelically moral superhumans, cooking delicious vegan meals and ~~judging me silently~~ not mentioning it.

To be clear, I don’t think this is an effective strategy in the conventional sense. Them quietly hoping I would go vegan through the mere exposure effect was surely less effective than just donating a bunch of money to The Humane League. And them not caring about me going vegan makes even less sense. But it wasn’t nothing, and they meaningfully spread the meme, and so I spread the meme unto you, like some sort of tofu-loving apostle.

Internal inconsistency feels bad

There’s another benefit to living in accordance with your values:

You get to live in accordance with your values.

Doing the thing you know is bad and then paying money to offset it has a cost. Some will feel bad about it, dirty. You know you’re doing something wrong, and that you believe to be wrong, even if you’re buying yourself an affordance. Being a certain way propagates values (even if just to yourself) in a way that paying doesn’t replicate.

Living in accordance with your values is its own reward. Or at the very least, not living in accordance with your values exerts a constant, ongoing cost on your psyche.

That was the other part of my going vegan: the constant gnawing sense of doing something I know is wrong.

I don’t think non-vegans should beat themselves up, nor should vegans self-flagellate when they cave and buy a Twix (mmm). But there is something exhausting about constantly reminding yourself that you’re doing something you know to be wrong.

You get to model a better world

Ultimately, we strive for a world where animals are not bred, imprisoned and tortured because they and their secretions taste good.

(I also don’t think this applies to carbon offsetting. Moving around where carbon dioxide is emitted does not have the same moral cost as paying someone to improve animal welfare while still funding the industry that harms them. It’s possible you believe that getting to net zero will require a total phase-out of fossil fuels; I don’t personally believe this.)

It’s possible to donate to charities that have this as their specific aim, if that’s what you prefer – whether by doing vegan advocacy directly, like DxE or Animal Equality, or by working to create technologies to replace animal products altogether, like GFI. (I am not convinced this is the right approach, but that’s beyond the scope of this post.)

But being non-vegan is still in some sense being complicit in the system which you wish to destroy. I think this is at least somewhat… strange. It feels like an abolitionist keeping slaves, while donating to an anti-slavery charity because it’s more effective. If he’s only going to do one thing he should do the more effective one, I guess, but I still think he should do both.

Not everyone will feel bad about doing harm, even those who are animal welfare-pilled. Maybe you can be just fine eating pork and giving money to FarmKind. But I found that, after many years of taking psychic damage, I just didn’t want to do it any more.

1. It’s beyond the scope of this post, but one could argue that reducing one’s own meat intake isn’t fungible with donating to animal charities. Some of those charities will do things that are positive in expectation but don’t pan out – bringing a legal case that they lose, for example. You might prefer the certainty of not eating animals, although of course it’s possible that the difference in the number of animals killed for food is smaller than expected (or at least is again positive only in expectation), assuming that the supply of animal products isn’t perfectly elastic.

2. I’ve found that some people immediately respond by saying “carbon offsetting doesn’t work/is a scam”, which is mostly motivated reasoning – if done well, carbon offsetting definitely works. And if for some reason you don’t want to believe that, you can instead donate to the Clean Air Task Force, which advances technologies and policies to reduce carbon emissions and is a top-rated charity in the space.

Talking to strangers: an app adventure

Sun, 05 Apr 2026 00:00:00 +0000

WAIT! Want to talk to strangers more? You might want to take the talking to strangers challenge before you read on, otherwise your results will be biased!

Illustration by the extraordinarily talented Georgia Ray

Do you find it hard to talk to strangers? If you’re like most people, you probably do, at least a bit. This is sad. Talking to strangers is great! You can make new friends, meet a new partner, have a fling, or just enjoy a nice chat.

Most people think 1) people will not want to talk to them, 2) they will be bad at keeping up the conversation, 3) people will not like them.

They’re wrong on all three counts! Sandstrom (2022) did a study on this. People were given a treasure hunt app where they had to go and talk to strangers¹1. There were categories like “Al Fresco” (talk to someone outside), “Bossy Pants” (talk to someone who looks like a leader), and the excellent “On Top” (talk to someone with a hat… get your mind out of the gutter).. The control group just had to observe strangers.

The minimum dose was one conversation per day for five days. That’s nothing! You can totally do that even if you’re a massive strangerphobe! Participants averaged 6.7 interactions over the 5 days, so a little more than one per day. Presumably the more you do the better you get. Go team!

The paper finds that talking to strangers not only disproved the above beliefs, but also improved people's enjoyment and the impressions people thought they'd made on strangers. (However those last two also occurred in the control condition – it’s possible that simply observing strangers might do this.)

Importantly, the effects persisted when participants were surveyed a week later. So it might be a durable way to improve people’s beliefs around talking to strangers.

Crucial point: the paper notes that often people do have positive interactions with strangers, but that doesn’t seem to be enough to unlearn their wrongly negative beliefs about them. So participants had to do this every day for a week, not just once.

Do you want to love talking to strangers too?

Time to crack out Claude Code.

--dangerously-skip-permissions

I reproduced the app from the study, abridging the questionnaires as they’re a bit tedious. It also has an ‘express mode’ so you can do it just for a day – but remember that usually doesn’t work to actually get fix your limiting beliefs around talking to strangers!

I assume the study authors used the same design language

I assembled a small (N=3) study sample, drawn from an extremely unbiased population of nerdy rationalists. They’re a famously friendly bunch but also a bit weird, so this seemed good for testing the hypothesis. We wandered around Berkeley attempting to ruin people’s days with our bad chat.

Scores on the doors:

The results are good: a single conversation with a stranger obliterated nervousness, catapulted conversational confidence, and proved way less scary than predicted – exactly what the literature says will happen, every single time, and yet somehow it’s still a surprise.

We didn’t do the full five days, so we didn’t replicate the study. But we enjoyed it, and even in this single day we directionally confirmed the study’s result.

As the paper notes:

Despite the benefits of social interaction, people seldom strike up conversations with people they do not know. Instead, people wear headphones to avoid talking, stay glued to their smartphones in public places, or pretend not to notice a new coworker they still have not introduced themselves to.

I feel this. I’ve definitely worked at places for years where there were people I just NEVER TALKED TO. Which is insane if you think about it – you spend more time with these people than with your family! your friends! your polycule!

I want more people to challenge themselves and have an excuse to talk to strangers. Go forth and make new friends!²2. And don’t forget to email me the results!

1. There were categories like “Al Fresco” (talk to someone outside), “Bossy Pants” (talk to someone who looks like a leader), and the excellent “On Top” (talk to someone with a hat… get your mind out of the gutter).

2. And don’t forget to email me the results!

Cult of REWE

Sun, 05 Apr 2026 00:00:00 +0000

I don’t really remember my first trip to Germany, except that it was to Berlin. I imagine I did some clubbing and drinking and so on but something far more momentous happened: it was on that trip that I discovered REWE.

From the outside, REWE is just a big supermarket chain, like Sainsbury’s or Wal-Mart. But inside is a veritable cornucopia of stuff, somehow cramming in much more than you’d expect from a normal supermarket. Its name also cannot be correctly pronounced by non-Germans.

One area where Germany outdoes itself is vegan meat alternatives (and vegan stuff generally). The UK is decent for such stuff and most of Europe doesn’t even fucking try, but Germany is an exception. REWE sells at least a dozen kinds of fake meat made of lentils, soya and other non-sentient objects. Perfect for Abendbrot, that most German of meals (bread with stuff on it, aka girl dinner).

REWE is also known for its distinct aesthetics. The REWE on Schönhauser Allee, Berlin is affectionately known as “party REWE” for its enormous flashing sign¹1. Thanks to Skaiste Anuzyte for recording a video of party REWE for me!, and recently received UNESCO World Heritage designation:

Indeed, the Party continues inside the store itself, with Partygarnelen (party shrimp) lit up with this cheerful display:

REWE is also the best place to buy Ritter Sport, the world’s squarest chocolate. You might be familiar with it from your local corner shop, perhaps the pedestrian Cornflake or White Chocolate Macadamia flavours.

But did you know there are forty five varieties, including vegan options, minis, and cubes? That there are a bunch of limited edition flavours that aren’t even on the website, which you have to literally go to REWE to discover like some sort of chocolate-themed treasure hunt? That the Ritter dynasty is so powerful that there is a fucking Ritter Museum featuring abstract geometric art?

From above it probably looks like a Ritter Sport

The supermarket shenanigans don’t end there. Germany has two discounter supermarkets, Netto and Netto. Or to be more specific, Netto mit Hund (Netto with dog) and Netto ohne Hund (Netto without dog).

This should never have been possible

The explanation is somewhat bland – Netto mit Hund is a Danish discounter (presumably they call it something else) and the other one is German. This is merely convergence, but Germany also has a notable Supermarkt-Schism: Aldi was split into Aldi Nord and Aldi Süd when the Aldi brothers couldn’t agree on whether to sell tobacco products. They’ve since gone on to become huge multinationals (and they now both sell tobacco) but they never merged back, a testament to the longevity of sibling rivalry and German stubbornness.

If you want to understand a country, skip the museums and visit its supermarkets. In the UK, we have over 1.8 quadrillion sad sandwich+drink+snack meal deal combinations (although I will concede they are a testament to the cultural institution that is the cheap British lunch).

Two packets of crisps and no drink. Right to jail

In Germany, supermarkets are a celebration. Flashing signs, garishly illuminated seafood, people happily milling about the aisles buying square chocolate (“square, practical, good”).

This is a country with its priorities straight.

(Disclaimer: this post is a pre-print, pending formal review from the Shrimp Welfare Project censorship committee.)

1. Thanks to Skaiste Anuzyte for recording a video of party REWE for me!

Gabapentinoids I have known and loved

Fri, 03 Apr 2026 00:00:00 +0000

Gabapentinoids are weird.

For a start, they don’t do what they say on the tin. It was named after the thing the inventors thought it would do, i.e. bind to and modulate GABA receptors, the ones which cause sedation and anxiolysis. But they have no activity at these receptors. Intuitively then they wouldn’t have an effect on sleep or anxiety.

They also don’t bind to dopamine receptors — you would think then that they wouldn’t be helpful for psychosis (most antipsychotics antagonise dopamine receptors).

And they don’t bind to opioid receptors, so they’re surely not useful for treating pain.

But they do! Gabapentinoids are prescribed for sleep, anxiety, bipolar disorder, and epilepsy, as well as neuropathic pain and restless legs syndrome.

Ok so what do they bind to then

Gabapentinoids bind to the α₂δ protein, a subunit of voltage-gated calcium channels (hence their alternative name of α₂δ ligands). Usually the concentration of calcium ions outside the cell is thousands of times higher than inside; these channels respond to a voltage by opening and allowing calcium to flood in. Depending on the cell they’re attached to, this can cause muscle contraction, neuronal signalling, and protein synthesis.

Specifically they bind α2δ-1 and α2δ-2, but only exert their effect through the former (as proven by trials on α2δ-2-knockout mice). There seems to be an as-yet undiscovered natural ligand for α2δ-1 and -2 which binds to the same site as gabapentinoids.

Importantly they don’t block calcium channels — instead they inhibit the release of monoamines (serotonin, norepinephrine, dopamine) and substance P¹1. Substance P is a neuropeptide responsible for transmitting pain signals. triggered by calcium influx. They also inhibit calcium channel-dependent release of glutamate and glycine in various brain tissues.

Sensitized calcium channels

There are states in which calcium channels become ‘sensitized’, such as in the case of neuronal injury, and gabapentinoids might selectively work in these conditions.

Activation of protein kinase C is required for gabapentinoids to reduce substance P released caused by capsaicin
Gabapentinoids reduce the size of postsynaptic currents in certain tissues in hyperalgesic rats (which have been bred to feel more pain), but not in normal rats
Glutamate release triggered by substance P is blocked by gabapentinoids

As they don’t simply block calcium channels, they have big advantages over drugs that do — they only minimally change synaptic function, unlike calcium channel blockers. They can essentially restore ‘normal’ functioning in overexcited calcium channels while leaving healthy ones alone.

Natural gabapentinoids in the body

Anticlockwise from top: gabapentin, leucine, isoleucine

Gabapentinoids have a suspicious structural similarity to leucine and isoleucine, two amino acids. Radiolabelling these amino acids shows they also bind the α₂δ protein, and L-isoleucine blocks certain effects of gabapentinoids, suggesting they compete for binding at the same site.

Some people have reported relief of their restless legs syndrome from acetylleucine, a leucine analog, which suggests it’s acting in a similar way to gabapentinoids (Fields 2021). Curiously this drug is very hard to find except in France, where it’s sold over-the-counter.

Gabapentin vs pregabalin

Unlike lots of drugs, gabapentinoids seem to be actively transported into the body by LAT1, the large neutral amino acid transporter.

This is a disadvantage over other drugs, because it limits how much and how quickly gabapentin can be absorbed. Gabapentin often has to be taken multiple times per day to avoid saturating these transporters. It also competes with other amino acids (the ones above) for these transporters.

Pregabalin, another gabapentinoid, is superior here because it is transported by other carriers, not just LAT1, so its uptake doesn’t saturate in the same way. It binds α2δ much more strongly than gabapentin, and in animals is more potent as an analgesic and anticonvulsant.

Can they block synapse formation?

Even weirder: Eroglu 2009 found that α2δ-1 is a neuronal receptor for thrombospondin, a molecule which promotes synaptogenesis in astrocytes. Specifically, it forms part of a larger signalling system. It acts as the extracellular receptor for a “synaptogenic signalling complex”; when thrombospondin binds, it causes a cascade of events which switches on this complex and leads to the start of synapse development.

As gabapentinoids also bind to this protein… does that mean they reduce synaptogenesis? In vitro, yes: gabapentin powerfully blocks synapse formation. Though this sounds slightly terrifying it’s also probably an important mechanism for gabapentinoids’ effects in epilepsy and neuropathic pain — synapse formation can be triggered by neuronal injury in these conditions and might well contribute to the pathology of these conditions (although this is uncertain).

It’s worth noting that gabapentin and thrombospondin, while both binding to the same protein, don’t bind to the same part of that protein.

(It’s kind of nuts that it took decades for one of the key mechanisms of action for this class of drug to be discovered. Makes you wonder what else we don’t know, about gabapentinoids and other drugs.)

Memory, executive function, and dementia

Worryingly this suggests that gabapentinoids might affect the normal formation of synapses. Could this cause other deficits, such as in memory formation?

Behroozi 2023 attempted to test this and did not find an effect, although they were looking specifically at improvements in memory formation.

Gabapentinoids can certainly cause brain fog and slower processing. Eghrari 2025 also found an increased risk of cognitive impairment and dementia in patients with chronic low back pain prescribed gabapentin; when stratified by age, patients taking gabapentin had twice the risk of dementia and mild cognitive impairment. This risk was further increased in patients who had taken gabapentin more throughout their lives. Presumably this effect would also extend to pregabalin.

A billon-dollar scandal

Gabapentinoids are frequently prescribed off label (when a doctor prescribes a drug outside of the conditions for which it’s approved). Not necessarily a bad thing: doctors use their discretion to decide when to do this, and for a drug with as broad a therapeutic profile as gabapentinoids it doesn’t seem wholly surprising.

But there are strict rules around advertising a drug for this sort of thing, or pushing doctors to prescribe it off label. The drug is approved for specific conditions and drug companies (in countries where they’re allowed to advertise) can only push for it to be prescribed for these conditions.

Their subsidiary Parke-Davis promoted Neurontin (gabapentin) for at least eleven unapproved conditions, flying doctors to lavish retreats, paying kickbacks, and commissioning ghostwritten journal articles. Off-label prescribing accounted for 78% of Neurontin sales.

Pfizer pleaded guilty to criminal charges and paid $945 million in settlements. In a separate 2009 case, they paid a further $2.3 billion for off-label marketing of several drugs including Lyrica (pregabalin).

Separately, top pain researcher Scott Reuben admitted to fabricating data in at least 21 studies – including Pfizer-funded trials of Lyrica – without ever enrolling a single patient. He was jailed in 2010.

Can they make you suicidal?

More controversial. One epidemiological survey looked at a cohort of individuals before and after they were prescribed gabapentin, and found no increase in suicidality, as well as a reduction in suicide attempts in psychiatric patients (Gibbons 2011). A large Swedish cohort study found a significant increase in suicide – but only for pregabalin, and not gabapentin (Molero 2019).

It’s not clear why this would be the case, as the drugs work in exactly the same way (as far as we know). In fact, pregabalin was found to increase suicidal behaviour/deaths from suicide, unintentional overdoses, head and body injuries, road traffic accidents and offences, and arrests for violent crime, where gabapentin had no or almost no effect (and actually reduced road traffic incidents and arrests).

The obvious explanation is that pregabalin is simply more powerful, both due to the pharmacokinetic gap described above and because it binds α2δ much more strongly. The highest doses of gabapentin simply can’t compete with the highest doses of pregabalin.

Are they fun?

Certainly for some people they are. Gabapentinoids are notorious for diversion, where people score prescriptions and then sell the drugs on. The prescription rate for these drugs in prisons is double that of the general population. In some ways pregabalin is the drug of choice in UK prisons. In France, 81% of recreational teenage pregabalin users reported to poison control centres were homeless or living in migrant shelters (Dufayet 2021).

This should surprise us; they don’t have any dopaminergic or opioidergic activity, so they don’t tick the obvious addictive drug boxes. Nonetheless, some people clearly find them enjoyable, with effects somewhat similar to alcohol/benzodiazepines, and develop dependence on them.

This might explain why there are more than 50 million gabapentinoid prescriptions issued every year in the US alone.

In the hilarious Drug User’s Bible, in which the author takes basically every drug imaginable, there’s this snippet from taking 300mg pregabalin (which is a hefty dose, users are started on 75mg typically):

I totally underestimated this drug. I am basically zombified and largely mistuned to what is going on around me, which appears to be distant. My hands are numb and I am, essentially, stupefied, with head spinning.

This one was a shock. I clearly took far too much and paid a price in terms of a strong intoxication which at times was extremely uncomfortable.

Conclusion

Gabapentinoids are weirder than I had realised.

Of course, the conditions they are prescribed for are horrific – anxiety, chronic pain etc can be a living hell, and a drug which effectively treats them is miraculous. But it’s wild that it took us decades to actually understand the first thing about how these drugs work.

And the effects on synaptogenesis, unknown effects on memory, increased risk of various kinds of death and dangerous behaviour (in the case of pregabalin), huge abuse in prisons and migrant shelters, and increased risk of cognitive deficits and dementia should probably worry us given how widely they’re prescribed.

References

(with apologies to Sasha Shulgin)

1. Substance P is a neuropeptide responsible for transmitting pain signals.

Capture and coast

Thu, 02 Apr 2026 00:00:00 +0000

Have a look at two versions of the same post, one on Substack and one on LessWrong.

With even a cursory glance, it seems clear that the LessWrong interface is much better than Substack’s. Footnotes appear inline as Tufte-style sidenotes, meaning they can be seen alongside their context without having to scroll to the end of the page. (In a print book, having footnotes at the end of the page or endnotes at the end of the chapter makes sense – they’re a space-constrained medium – but on the web there’s no reason for this.)

Sidenote on LessWrong

Comments on LessWrong are much, much better than on Substack. (The whitepill here is that it’s possible for a small, mission-driven team to produce better software than a billion-dollar business.) WordPress has had nested comments since 2009 and Substack’s implementation is still worse somehow.

There’s very little visual hierarchy, making the boundaries between comments hard to scan. Comments are strictly plain text, with no ability to format them or tag other commenters. Until recently you had to click through from an article page to see the comments section, and to view comments nested more than one layer deep you need to click through again. [EDIT: you actually still need to do this, but not on Astral Codex Ten, weirdly, which makes me think there’s some special mode for larger Substacks.] You can ‘like’ comments but you can’t give other reactions, you can’t downvote, you can’t tag other people, you can only sort by date, collapsing threads is done by clicking in an unintuitive place.

It’s dogshit.

(Substack has a bunch of other ridiculous limitations, like no support for tables.)

Platform ossification

Why am I bitching about comments so much? Because this sort of feature downgrade is endemic anywhere there’s platform lock-in. The product simply doesn’t need to be as good – the network is the product, and if you can’t get eyeballs on your blog unless it’s on Substack then Substack doesn’t need to have anything like feature parity with other platforms.

Platforms innovate, then once they have the network, they stop improving.

The same is true of the social media platform formerly known as Twitter. At its inception it was something genuinely new; at-handles, hashtags and retweets were all genuinely innovative. Then they spent a decade copying other platforms’ features (mostly badly), and latterly Musk fired 80% of the engineering staff, which left the platform functional but ossified. Longstanding bugs never get fixed. The network lock-in is now so strong there’s no reason to – it’s the internet’s de facto public space. Good luck getting your shitty Mastodon “toots” to have even a hundredth of the reach of your tweets.

A number of other examples of this kind of ossification/enshittification come to mind:

Slack search is still horribly broken
Reddit is actively hostile to its users – lots of gating that requires you either to sign in or download the mobile app, plus they killed popular third-party apps like Apollo overnight)
Google Maps’ UI seems to get progressively worse and harder to use with every passing year
Instagram fought chronological feeds for years after switching to an algorithmic feed (there was even a #RIPInstagram campaign to get them to switch back)

But platforms do add new features!

There is some amount of innovation going on – usually driven by a competitor implementing a new feature (think Stories being added literally everywhere, or Twitter Spaces as a response to Clubhouse). Instagram Reels and YouTube Shorts were a response to TikTok’s ascendency. (Maybe this post should have been called “the enTikTokification of everything”.)

And those features are horrible! I hate reels! I have to install a panoply of Chrome plugins to block Shorts! We ended up inventing a word to describe becoming horribly addicted to these things and collectively wasting billions of hours a day on them!

Why would you be proud of this

This is a different pathology from ossification, but it too comes from network lock-in. Shoving unwanted short-form slop into an app for sharing photos with your friends would be suicide for any company which didn’t have a total stranglehold over their audience.

When you scroll to the end of a Substack comments section, you just see a sad ‘Load More’ button. Infinite scroll never reached this part of the web, I guess.

Are we paying by the GET request or something?

A paean to bigness

Wed, 01 Apr 2026 00:00:00 +0000

On landing at SFO, my ridemate mused that the culture shock of coming to the US is greater than that of going to Europe despite their more different-seeming cultures (he is Scottish, I’m English¹1. Before my father disowns me, although I am half-Scottish by birth I am surely culturally English having grown up and lived there for almost my whole life.).

Everything is big. Our Uber takes us on the 101 north to I-80 East over the Bay Bridge. The road is huge, the cars considerably bigger than what we have in the UK, and driven much more aggressively (my pal is somewhat less stoic than I and flinches at a few last-minute merges). The billboards by the side of the road are without exception for AI startups, including one that helps detect who online is human and who is a robot.

We scythe through San Francisco while seeing essentially none of it, except the skyline and the large windows of a few converted lofts with tech workers tapping away at expensive keyboards.

“I always feel strange seeing people in office buildings,” he says. “Do they not know what’s about to happen?”

⁂

This is not a new observation of course. Big cars, big roads. But the bigness, and a sense thereof, pervades everything here.

On the Lighthaven campus, where I have to publish 500 words a day for a month on penalty of eviction, everything is big too. The bathrooms contain enormous pump bottles of shampoo, conditioner, and shower gel the likes of which are simply unavailable in the UK. A huge roll of paper towel is usually within reach wherever you are, and so I use it liberally (the idea of using a dishcloth to dry one’s hands feels like barbarism).

The kitchen is stocked with every conceivable snack, vegan or otherwise. A huge urn of coffee sits at the ready at any hour of the day (which I discovered when I tried to top up my tea with it by mistake). Packs of bagels are piled a foot high on the counter and I have to stop myself from eating four of them one after the other. The peanut butter is billed as ‘no stir’ somehow; god knows what industrial emulsifiers are used to achieve that. Even cans of Coke are bigger here – 355ml (12 fl oz in American) to our pathetic Europoor 330ml. The land of plenty.

I sit at a couch, of which there are very many. This one has a monitor on a floating arm. I go to balance my tea on the armrest, a precarious but necessary move in any British home, and find there is a wooden Thing on it to put it on. There’s a square holder for cups but my too big mug won’t fit in it until I realise there is a notch cut into it for the handle. It’s 6am and I can’t sleep and I want the comfort of a blanket, which when I go to pull it over me I find is strangely heavy because of course it’s weighted.

⁂

I got the address wrong and our Uber dropped us a few minutes’ walk away, so we rolled our suitcases along Telegraph Ave in the warmth. The air is sweet-smelling: eucalyptus (planted everywhere a century ago for timber and never cut down), Bay laurel, and jasmine.

A pretty girl walks past – young, I guess a student. UC Berkeley is just up the road. Its campus is also big in comparison to the inner-city campuses I’m used to. She glances at me; there’s a brief frisson. I feel something shift inside me, like I’m a teenager not sure where to look. London has plenty of beautiful people but there’s something new here too: American girls, even young ones, are confident.

Brits like to shit on Americans for their personality. We think that tall poppy syndrome and false modesty are less unbecoming, a sort of national suspiciousness of ambition. But what they see as loudness or brashness I see as self-assuredness about moving through the world. Something like poise.

1. Before my father disowns me, although I am half-Scottish by birth I am surely culturally English having grown up and lived there for almost my whole life.