iiinova — where your digital self keeps evolving

Andrew Perfors @perfors.net · 12/03/2026

13yo just did it and got this insane result. I suppose this is what happens when your eyes are young and fresh

Andrew Perfors @perfors.net · 12/03/2026

This is mine. Better than I thought I'd get (I'm not a vision scientist!) but still far from you. Still, all that time spent squinting at figures may have paid off :)

Mine is 0.0054. I would make a excellent paint-chip inspector.

130

Andrew Perfors @perfors.net · 11/03/2026

This is such a good piece. I feel so many of the same sentiments.

I am sick to death of this — of the tech industry deciding it can do what it wants without consequences. Everything is a subscription, everything is harder to use and enshittified. Everything is built on something that was stolen or degraded. So-called AI – which is not intelligent at all – is being shoved into everything, and makes nothing better (quite the reverse). “Move fast and break things” now extends to the reputations and identities of respected professionals and creative people. And you have the nerve to presume we should all put up with every outrage, with every act of destruction, with every theft. I don’t. I won’t.

0203

Andrew Perfors @perfors.net · 16/02/2026

I realise my Bluesky audience is speaking to the choir here, but the linked essay does such a wonderful job at explaining why I'm so passionate about *every* trans person receiving gender-affirming care. It is, quite simply, miraculous.

The ability to relegate one’s sex assigned at birth to an inconsequential, incidental characteristic feels to older trans people like alchemy. Trans children deserve to live those lives just as much as anyone else, and no one who attempts to justify withholding transition care appears to have an answer for why this should be any different.

111322

Andrew Perfors @perfors.net · 06/02/2026

This is an especially good article about why the Epstein files matter so much.

Being able to see their emails makes something very clear: The “anti-woke” movement was not a genteel intellectual inquiry, made by disinterested parties who cared deeply about free speech. It is a social circle of powerful people who feel threatened by #MeToo. Reading through the emails, it’s possible to see that Epstein himself coordinated pushback against #MeToo. Looking back, it seems obvious that sinking #MeToo also led to where we are now: a place where laws simply don’t apply if you have enough money and power. The same players are involved.

1249108

Andrew Perfors @perfors.net · 05/02/2026

A friend of mine shared this meme with me and I just had to post it here because it sums up so well where many of us are right now Hang in there! Like the lemon 😂

Top says "the lemons life gave me" and the picture is of a decaying lemon that looks like it is grinning hideously at you

1151

Andrew Perfors @perfors.net · 31/01/2026

I love Nevada! Took this picture off US 50 when I visited in July. Such an evocative landscape.

Picture of a lonely highway with scrub brush in the foreground and rolling desert hills in the background, and a completely clear crystalline blue sky.

1202

Andrew Perfors @perfors.net · 23/01/2026

Yeah, me too. I have a whole talk about this with this theme. Gratifyingly, it seems to resonate -- the difficulty is figuring out what to do about it (i.e., as a researcher in cognitive science, how we can best study the real problem in such a way that results in real impact).

One of the slides from the talk

Disinfo is systemic in its goals and methods

It is designed to capitalise on the psychological and structural weaknesses in our institutions and systems

Another slide

The solution must be systemic

We can no more individually solve our way out of this than we could expect people to eat healthily without food safety laws

130

Andrew Perfors @perfors.net · 22/01/2026

Part 7: My recommendations The previous parts have all been the meaty ones; in this I try to give some general suggestions for you – where “you” means everyone: students, parents, teachers, employees, scientists, policy-makers, people in charge, etc.

If you want to use LLMs:

1. Say no to FOMO. If you’re going to use AI, decide to do so based on
demonstrated need and defined use case, not fear of missing out.

2. Avoid magical thinking. The details matter. Make sure you’re clear
about the mechanism by which you expect AI to help, and consider the
impacts (both positive and negative). Ask yourself: what will you need to give up to make space for AI in your workflow?

3. Treat LLMs like interns. Don’t use LLMs for something you don’t
understand or need to learn. Thoroughly build checking into the process.

4. Evaluate objective impact. Decide ahead of time how you will test
impact in ways that measure the right thing and aren’t gameable.

5. Beware the sunk cost fallacy. Let yourself walk away or change your
mind, and try to plan in advance so you have a way to do that.

General recommendations

6. It’s okay not to use AI. Most of what we value is the meaning and
connection we get from other people. Don’t lose sight of that.

7. Value process more than results. LLMs make things easy, but this can rob us of growth. Difficult roads lead to beautiful destinations.

8. Use common sense. Ask yourself: does this make sense, based on my actual experience? Who benefits if I do this? Who gets hurt?

9. Remember your values. Figure out what matters to you and what kind of person you want to be. Don’t lose sight of that.

10. Don’t give up your power. As much as possible, avoid creating a
reliance on an external tool that you don’t understand and can’t control.

1324

Andrew Perfors @perfors.net · 22/01/2026

Part 6: How do LLMs impact society? This video takes a broader perspective to look at other impacts on our world: our workplace, the economy, and the environment. These are the topics that I have least expertise in so be more cautious but I couldn’t do an series without going over these issues.

LLM impacts on work

Two general types of claims, as far as I can tell

1) It helps me do something I couldn’t otherwise do

Do you have the expertise to catch errors?
Does this take jobs or steal from people who do it?
Does this hurt the development of future expertise?
Does this enable you to spend your time better?

2) It has the potential to revolutionise something
To what extent is this marketing: vague and always years away?
Who benefits from the revolution?
What societal groundwork is necessary to make this possible?
What will happen if this is achieved?

The benefits of
LLMs to the
workplace may
be overstated

We run the danger
of embedding
tools we don’t
control into our
processes, and
then becoming
dependent on
them

The
environmental
impact of LLMs
is terrible

1160

Andrew Perfors @perfors.net · 22/01/2026

Part 5: How do LLMs impact humans? We’ve seen how LLMs work and how they affect how we learn, but of course we are more than that! This video is about other ways LLMs might impact us: in particular, how they affect our information systems (via disinformation) and how they affect our well-being.

How does disinformation create distrust?

Modern disinformation warfare uses multiple techniques, many of which are supercharged with LLMs and AI:

‣ Identity and emotion exploitation
- Deepens group animosity

‣ Conversational control
- Manipulates how we make decisions

‣ Illusory consensus and repetition
- Exploits our biases to shape our beliefs

‣ Destruction of common ground
- Undermines our shared reality

Endgame of LLM-driven
disinformation: to undermine
our shared reality

Human connection matters

At best, LLMs are a kind of "junk food" version of social connection

They are a "superstimulus" for the part of our psychology that craves
connection — they feel good because they fit what we’ve evolved to see as indicative of a mind and connection, but they don’t have the same nutritional value

We evolved to trust and
connect with each other
and the world

LLMs undermine this process of
shared meaning-making

2150

Andrew Perfors @perfors.net · 22/01/2026

Part 4: How do LLMs impact learning? Now that we understand something about LLMs and human learning, let’s put them together and explore what we know of the impact of LLM usage on how we learn.

When are LLMs helpful?

Second language learning: Little to no research looking at LLMs per se vs bespoke platforms

Purpose-built systems: Potentially sometimes? but that’s not usually feasible or relevant; also, need more high-quality studies

Motivation: Heavily confounded and often poorly measured, very hard to tell

soo... Can’t rule these out but not much good evidence for it

Creativity: Helps individuals in the moment but collectively worse (makes everyone think of similar ideas to each other) -- so this is a bit of a monkey's paw situation

There is no strong evidence that generic LLMs are helpful for learning (other than brainstorming)

But even there, LLMs hurt us collectively (by diminishing our uniqueness)

When are LLMs not helpful?

Psychological outcomes: Less social support, more procrastination and memory loss, lower GPA - but hard to disentangle cause and effect

Poorer learning: In STEM and humanities, LLMs lead to worse performance on tasks where students can't use them -- but students often unaware and think they get it

Worse higher-level skills: LLMs are easier and faster, but this means less depth of thinking, engagement, analysis, and processing.

This is suggestive of offloading the thinking to the LLMs without realising. Need more research into the exact mechanisms of this, and how robust it is

LLM usage
undermines the
deep processing
and engagement
necessary to
develop real
expertise and
critical thinking
skills

1170

Andrew Perfors @perfors.net · 22/01/2026

Part 3: How do people learn? In order to understand how LLMs are different from humans and how they affect us, I first need to explain some of what we know about humans and how we learn.

Higher-level expertise requires mastery of the lower levels

Our brains have limited short-term capacity (attention, memory, etc)

Which means the more basic stuff that can be offloaded to long-term habit & chunking, the more is available for complex reasoning at more
abstract levels!

The kind of practice matters

It also needs to be at the zone of proximal development: focused on the things you’re almost on top of, but not yet

The purpose of assessments is to force learners to practice the right way

The process is the point; it can’t be skipped

2211

Andrew Perfors @perfors.net · 22/01/2026

Part 2: What can (and can't) LLMs do? Having seen the basics of how LLMs work, this video goes into more detail about what kinds of things they can do along with a little bit about why.

Why do LLMs bullshit? [hallucinate is crossed out]

Hallucinations are a mathematical inevitability that cannot
be eliminated even in principle — but there are practical
problems that may be nearly insurmountable as well

‣ Reinforcement learning penalises for guessing, and the
business case means this will probably not change

‣ Running out of training data, so training on own output
(which means that errors get exaggerated with more training,
somewhat like recessive genes)

‣ Performance on benchmarks may actually be going down

Training data matters a lot: the more unique or idiosyncratic something is, the more likely an LLM will be to hallucinate bullshit about it

What LLMs can(not) do

LLMs lack the underlying concepts and grounding
in the world that humans have

The more a task requires reference to the world
(beyond the statistical probabilities encoded in
the model), the more LLMs will struggle

LLMs mimic humans very well

They were trained on (and memorised parts of) the entire
internet, and our language use reflects our internal concepts

They were trained to produce text
that people find plausible, which
means errors tend to be subtle and
require expertise or existing domain
knowledge to easily identify

3272

Andrew Perfors @perfors.net · 22/01/2026

Part 1: How do LMMs work? Goes into the basics of LLMs in a non-technical fashion, mostly with pictures.

LLMs: Very powerful autocomplete

LLMs work by learning the statistical relationships between words
(i.e., which words tend to follow which other words) in a way that
is (loosely) modelled on neurons in the brain

LLMs: Very powerful autocomplete

With lots of cycles of training, the model is basically
approximating a function in very high-dimensional space

What changed recently?

Mainly massive amounts of computational power and training data

1180

Andrew Perfors @perfors.net · 22/01/2026

Just to whet people's appetites, I'll make a thread below which contains a bit about each video, including screenshots from some of the conclusion slides for each (you have to watch the video and look at the references to understand the reasoning, which doesn't really fit in a screenshot 😂) 🧵

Overall plan for the series

Part 1: How do LLMs work?

Part 2: What can (and can’t) LLMs do?

Part 3: How do people learn?

Part 4: How do LLMs impact learning?

Part 5: How do LLMs impact humans?

Part 6: How do LLMs impact society?

Part 7: My recommendations

2264

Andrew Perfors @perfors.net · 06/01/2026

So much is fractally, recursively wrong here. I actually laughed out loud at this:

“If the woman is about half a standard deviation more agreeable than the man, that’s the optimal point for relationship durability,” he said, to which I must have emitted an observable psychometric signal.

“Look at your face,” Mr. Zervas said. “You’re horrified.”

21784

Andrew Perfors @perfors.net · 31/12/2025

As 2025 ends, I say goodbye to the only perfect-square year I will ever experience. Personally it's been good, but I don't think I'm alone in feeling that 2025 was mostly a dumpster fire. Still, I always try to remember that no matter what, it remains a beautiful world. Here's to a better 2026!

Photo of a path in my garden. Rocky steps heading up towards fern trees in the distance.

0382

Andrew Perfors @perfors.net · 27/12/2025

Nothing like using your Christmas holiday to build a shed and remind yourself that there is a reason you are not a professional handyman 😂

PIcture of a large shed looking kind of rickety

5351

Andrew Perfors @perfors.net · 18/12/2025

My favourite parts are my word cloud and emoji usage. I kinda like that I apparently mostly talk about "people" and mostly use the 😊 emoji

Word cloud whose most used word is people, followed by "not" and "its" and "things" and "good." I am a simple person with simple vocabulary but those are some pretty excellent words!

My mood this year is mainly the smile emoji, followed by the sad one and a bunch of hearts and rainbows

030

Andrew Perfors @perfors.net · 18/12/2025

I have been emailing with a student offering advice about navigating the post-PhD job market and it struck me that the advice was general enough that more people might benefit. So here you go! tl;dr: Figure out what you want your contribution to be, and make choices that can make that happen.

Q: When I think about what will set me up for a university job, what should I evaluate potential post-docs based on? Should I consider things like the institution and where it's based? Is it about the subject matter of the project and the kind of papers I'll publish and the skill set I'll use? Or is it more about the connections that I'll have the opportunity to make and going to conferences and that kind of thing?

A: I think the main things are: (a) the kinds of papers you'll end up being able to publish; (b) the skill set you'll acquire; and (c) the connections you can make. But all of these are contextually dependent -- really what matters is that they pave the path that you want to take and support your narrative of what you do and how you fit (because once a candidate passes a certain bar for quality, hiring decisions are usually made on the basis of skills and fit).

To illustrate: if you want to be the kind of researcher whose contribution to your school/field (and narrative) is that you have certain skills, or a certain area of expertise, then you want your postdoc to be one that develops those skills/expertise, where you write papers that showcases that you have those skills/expertise, and where you create connections with people who care about those skills/expertise. That's why it's so important to figure out your narrative and what you want to do: it all starts from there. So the question is not: how do I set myself up to get a university job? It is what kind of university job do I want, and how do I set myself up for that?

Some people are like, I don't care, I just want any job! But people don't want to hire people who have no "why" for what they do beyond just getting money -- people hire people who have a contribution to make. So figure out what contribution you want to make, and then choose a postdoc that will best set you up to be able to make it. (And you can always change direction, people do that often, but you want to start with a direction).

03010

Andrew Perfors @perfors.net · 11/12/2025

Thanks, this is helpful, although as far as I can tell (on a quick skim) it seems to be mainly about the "simpler" case of a single interaction term and I'm specifically interested in problems that arise from throwing multiple terms in there (either 3-way interactions or just multiple 2-way ones)

Consider the classical linear multiplicative interaction model that is often assumed in empirical work and is given by the following regression equation:

(1)
𝑌=𝜇+𝜂𝑋+𝛼𝐷+𝛽(𝐷⋅𝑋)+𝑍𝛾+𝜖.

In this model 𝑌
is the outcome variable, 𝐷
is the key independent variable of interest or “treatment,” 𝑋
is the moderator—a variable that affects the direction and/or strength of the treatment effect,Footnote 6 𝐷⋅𝑋
is the interaction term between 𝐷
and 𝑋
, 𝑍
is a vector of control variables, and 𝜇
and 𝜖
represent the constant and error terms, respectively

050

Andrew Perfors @perfors.net · 16/11/2025

Timeline cleanse, there is still beauty in the world. 💚 #tinyjoys

A sulfur-crested cockatoo visiting me on my deck. It's a little irritated at me for not feeding it but we had a nice conversation nonetheless.

31069

Andrew Perfors @perfors.net · 14/11/2025

Here's more of the list, again, with lots of scientists 😢

Lots of people including many repeats from above but also Gardner, Hoffman, Ramachandran, Christakis, Sheldrake, Gell-man, Seligman, Gould, Mandelbrot, etc

120

Andrew Perfors @perfors.net · 14/11/2025

As for why I think this is Bannon? In addition to the clues mentioned previously, it's because Bannon was actually filming a documentary about Epstein when he died. en.wikipedia.org/wiki/Steve_B...

Connection to Jeffrey Epstein
According to Michael Wolff, Bannon was introduced to Jeffrey Epstein in 2017 and Epstein introduced Bannon to some of his friends.[276] Bannon also worked with Ehud Barak and attorney Reid Weingarten to attempt to reform Epstein's public image.[277][276] Later reports showed Bannon reportedly prepped Epstein for an interview with 60 Minutes that ultimately never occurred. Bannon confirmed that he did tape 15 hours of interviews with Epstein but denied that he was coaching him for further interviews and that the footage was for an unannounced documentary on Epstein.[278][279][280]

130

Andrew Perfors @perfors.net · 14/11/2025

Given this context I think the infamous list for Steve Bannon is a list of people Epstein wanted to be in this documentary they were filming. scientists etc as alluded to in one of the screenshots above.

One of the lists of names that's already been shared in lots of places, I'm not going to type them all out, but a bunch of scientists including Minsky, Kosslyn, Randall, Damasio, Gould, Seligman, Chomsky, Mandelbrot, Greene, Ito, Bach, Edelman, and Krauss are on it

3211

Andrew Perfors @perfors.net · 14/11/2025

And it looks like less than a month before Epstein went to jail they were filming something that Chomsky and other scientists were in. My guess is that the person he's talking to is Bannon

Message from whoever Epstein is talking to (my guess Bannon) on 06/16/19 saying "Before shooting sked can I get subject matter experts in science , math, Wall Street -- I have overall smes need personal"

(Note this came before the previous one I just copy-pasted in the wrong order and am too lazy to redo)

06/17/19, message from Epstein with a list of names (my guess is people he thinks should be in the film). "melanie walker. Brain surgeon. Martin Novak. . joscha back . Seth Lloyd, Lawrence Krauss, chomsky , mark tramo. Music. Not sure re summers.. should we use Michael Wolff..? John Brockman. Edge."

06/17/19: Epstein: Most dead
Epstein: Minskys. Wife. gloria. Should be done asap
Other person: We can pull footage -- just get me names whether living or dead
Other person: Plenty of footage.

A few messages at first which is part of why I think this is Bannon, referencing "In the closet of the Vatican"

But then on 06/18/19 is a message from Epstein: "Chomsky asking when film"

1294

Andrew Perfors @perfors.net · 14/11/2025

They were actually friends who visited each other, as far as I can tell

Message from Epstein on 05/26/19 saying "Im going for lunch tomorrow at the Chomsky's'

3446

Andrew Perfors @perfors.net · 14/11/2025

Good article here trying to make sense of the senseless (the Epstein emails). So much about our current timeline makes sense when you realise that vastly more of the people with power are vastly more idiotic and more venal and more immoral than we thought.

That reading these emails feels surreal makes sense—they shatter the myth of genius and merit that the ruling class tries carefully and spends exorbitantly to cultivate, and they affirm the worst suspicions of the conspiracy-minded. As more revelations are made public, it may feel like the conspiracy theorists have won. But they’ve been wrong as well. Cabal is too flattering a word for this crowd of cosplaying, hunt-and-peck email addicts. Conspiracy theories are a flawed tool meant to help make sense of a nonsensical world. The truth is darker: You don’t need an elaborate master plan to dodge accountability when everyone’s all too willing to simply look the other way.

0181

Andrew Perfors @perfors.net · 13/11/2025

Hello fellow nerds! The upcoming Australasian Mathematical Psych Conference (AMPC) is going to be on Feb 23-25, and they've just put out the call for abstracts. Details below. This is one of my favourite conferences and this year it's in Singapore(!!) which should be awesome.

Call for Abstract (AMPC 2026): Applying Mathematical and Quantitative Psychology to Real World Complex Problems

Singapore, 23-25 February 2026. Abstract submission deadline 26 November 2024

Plus a bunch of other info. Go to https://ampc2026.com/ for details.

0198

Andrew Perfors @perfors.net · 06/11/2025

Getting nervous for the talk I'm about to give at a workshop about "using AI to drive impact" which features slides such as these.

Slide is titled: You don't need to use LLMs.

Science is a process of collaborative meaning making, by which we try to understand the world

Even if AI were perfect, we rely on it at our peril — it is not science if we (i.e., humanity as a whole) do not understand and cannot recapitulate all parts of it

I very, very rarely use LLMs myself. You can give yourself permission not to. Don’t FOMO yourself into it

Conclusion: Don't rely on something you don't understand and can't control

If you must use LLMS:

1. Treat them like you would an intern: only use them for things you can easily and thoroughly check

2. Make your process as robust as possible

3. Be aware of your own (human) cognitive biases

2537790

Andrew Perfors @perfors.net · 02/11/2025

Brilliant, lovely day here (except for the allergies). High of 26C. Tomorrow is going to be a shock!

Photo of a walk I took in the Dandenong ranges, with huge gum trees towering up into the bright blue sky.

Forecast for tomorrow (Monday) showing a high of 12C (low of 9C) and solid rain

171

Andrew Perfors @perfors.net · 29/10/2025

Some days are tougher than others, but always remember that it is still a beautiful world. #tinyjoys

Close up of pink and white flowers which I think are rhododendrons

Picture of large rhododendron bush, with fern trees in the background

0222

Andrew Perfors @perfors.net · 20/10/2025

10yo is writing a fantasy novel and it takes no prisoners

Excerpt of text saying "Prophecies, to me, honestly are just instructions written in the most annoying way possible."

0201

Andrew Perfors @perfors.net · 18/10/2025

It is still a beautiful world, always remember that. 💕

1290

Andrew Perfors @perfors.net · 06/10/2025

This is exactly right too. Oh heck just read the whole thing

And as for winning, I think it is an indictment even of Klein's apparent intelligence that he thinks that the way to persuade people to join the Democrats is to have Democrats abandon what are their alleged ideals, especially since they just lost so badly not by standing up for trans people, but by doing exactly what Ezra Klein is now proposing, because they listened to people like Ezra Klein.

You can't beat the mentality that leads to abuse by taking on the abuser's mentality. You can't defeat bad ideas by adopting them. You can't grow a coalition by selling out its members. And you can't persuade people who disagree with you that you can be trusted to fight for them when you won't even fight for those you say are with you.

Eventually you have to stand for something, or somebody else will tell you where to stand, and if you'd rather not, they'll force you to stand there, and stand there you will.

And there you'll be, standing in a place somebody else chose for you. You and the nothing you built inside.

150

Andrew Perfors @perfors.net · 06/10/2025

So many good parts of this, but I really liked this: when it comes to status quo centrist moderate people like Klein, it is very illuminating to realise who they instinctively empathise with and put themselves in the shoes of, and who they consider to be bargaining chips to appease the far right.

By his own admission, Kirk's murder affected Ezra Klein in a personal way that all the previous acts of supremacist violence committed against all its other targets across the span of history did not. So we see that Klein recognizes a "we" in Charlie Kirk that he does not recognize in trans people or in any of the other people I listed in the previous paragraph. In this, he reveals to us—perhaps without even knowing he is doing so—that for him, the disagreements he has with Charlie Kirk and Ben Shapiro are window dressing compared to the fundamental kinship he feels to them, three princes of public discourse, just trying their best to do what's best, whether "what's best" is feeding a hungry kid or blowing him up or zip tying his hands behind his back in the middle of the Chicago night.

When Klein scolds that "we have to live here with each other" he is making a statement about who it is he is getting ready to live with and who he is getting ready to live without, and most gallingly he is ignoring the fact that when it comes to supremacists all of us have been living with them already all along. Nobody is suggesting mass deportation of white supremacists, or the dissolution of straight marriages, or stripping away health care for conservatives. The troops and the cops weren't ever sent down primarily white streets, masked kidnappers aren't terrorizing white churches or corporate courtrooms. These supremacist hatemongers are being criticized, yes, and opposed, yes, and yes the culture of violence they have created sometimes ricochets back on them, but we are living in the country they demanded on having. We are all now experiencing the Republican proposition for humanity. Many of us are demanding they stop pursuing this proposition, and Republicans and other supremacist hatemongers insist this demand represents violence against them. Klein, by drawing the lines where he draws them, is agreeing with Republicans about this.

151

Andrew Perfors @perfors.net · 04/10/2025

Anyway, here's my favourite excerpt, but read the whole thing:

Author Czeslaw Milosz wrote that “A man may persuade himself, by the most logical reasoning, that he will greatly benefit his health by swallowing live frogs; and, thus rationally convinced, he may swallow a first frog, then the second; but at the third his stomach will revolt.”

How does a person convince themselves they need to swallow live frogs? They see themselves as being pragmatic in difficult times and willing to make difficult choices. They imagine that doing something unpleasant is the same as doing something effective. They think it will save them. But developing a rational theory of eating live frogs is not going to save anyone.

I’m reminded of the large news organization my spouse used to work for years ago. When a maelstrom hit—one that was affecting all of journalism—a top editor called together the whole newsroom to talk about the future. They would, he announced, have to decide which of their core values to keep.

The point is supposed to be that they’re your core values. The outlet lost its principles, and soon went into a tailspin that was in no way delayed and was perhaps accelerated by abandoning its values. If you’ve figured out your core principles, then they will guide you about what to do. Remember, you’re not just performing an idea, it’s your life. It’s who you are.

I think Klein, like many former liberals, is getting rolled by his his exposure to people who imagine themselves to be free-thinkers but are actually adjacent to hacks like Christopher Rufo and Richard Hanania. (If you don’t already know who they are, they’re not really worth looking up.) Because Klein doesn’t seem to currently locate himself in a tradition or perhaps has lost the ability to do so, he imagines that the pain of trading away core values is a positive example of compromise.

But those self-described freethinkers advocating surrendering whole communities will never become more like him. Klein will only become more like them.

191

Andrew Perfors @perfors.net · 04/10/2025

Nice essay! This bit in particular was really interesting to me (as someone who neither meditates nor knows much about Buddhism). It makes a lot of sense.

The Three-Legged Stool We Broke
Within Buddhism, mindfulness meditation is one of several integrated practices. For example there’s a concepual trilogy in Theravada Buddhism called dana sila bhavana—a three-legged stool where all legs are necessary for stability.

Dana means generosity and action directed toward others—social engagement, service, addressing collective suffering.

Sila means ethics—following precepts, moral guidelines for how to live in community.

Bhavana means mental development—meditation and mindfulness practice.

Popular mainstream mindfulness practice focuses almost exclusively on bhavana. It strips away the social obligation (dana) and ethical framework (sila), leaving only the technique for training your individual mind. When I attended a meditation retreat at a Buddhist monastery (more on that in a post later this fall), one of the senior monks warned explicitly that practicing meditation without the other two legs was “dangerous.”

But secular mindfulness doesn’t want the other two legs. It doesn’t want ethical precepts that might constrain individual freedom. It doesn’t want social obligations that might demand action beyond personal practice. It just wants the meditation technique—extracted, medicalized, and marketed as a tool for individual optimization.

The result? A technique for individual stress management and performance optimization, completely divorced from Buddhism’s social and ethical dimensions.

24114

Andrew Perfors @perfors.net · 25/09/2025

Chotiner interviews a Harper's Letter signatory

So you’re worried about cancel culture and free speech.

Yes, it’s a huge problem.

Can you give me an example?

A campus organisation sent a mass email telling us to be mindful of pronouns and cultural appropriation.

Oh. I see.

Also, a woke student group protested someone who they didn’t like! It was very stressful for that person to be called a bigot. Extremely stressful and unpleasant.

Are they a bigot?

Nobody should be allowed to say they are. Free speech requires it.

Well, the next time I hear someone complain about the government firing or deporting people for something they said, I will be sure to bring up the woke students to show that the issue is complicated.

0313

Andrew Perfors @perfors.net · 23/09/2025

Checking in from Australia: there has been no Rapture. Repeat: NO RAPTURE. It is, however, a lovely spring day 💚 #tinyjoys

Picture of some white flowers on a pear tree

59913

Andrew Perfors @perfors.net · 14/09/2025

Good day. Music festival with one kid, escape room with the other, gardening, reading... and most importantly, very little time spent on a screen or social media. Just trying to "be here, now" I need to do that a LOT more often. #tinyjoys

3310

Andrew Perfors @perfors.net · 11/09/2025

For those who, like me, could use the reminder today: it is still a beautiful world. #tinyjoys

Tree with big white flowers with yellow centres

1413

Andrew Perfors @perfors.net · 09/09/2025

I can't take credit for that part! It's from a lovely poem by Mary Oliver

Who made the world?
Who made the swan, and the black bear?
Who made the grasshopper?
This grasshopper, I mean —
the one who has flung herself out of the grass,
the one who is eating sugar out of my hand,
who is moving her jaws back and forth instead of up and down —
who is gazing around with her enormous and complicated eyes.
Now she lifts her pale forearms and thoroughly washes her face.
Now she snaps her wings open, and floats away.
I don't know exactly what a prayer is.
I do know how to pay attention, how to fall down
into the grass, how to kneel down in the grass,
how to be idle and blessed, how to stroll through the fields,
which is what I have been doing all day.
Tell me, what else should I have done?
Doesn't everything die at last, and too soon?
Tell me, what is it you plan to do
with your one wild and precious life?

Mary Oliver

093

Andrew Perfors @perfors.net · 09/09/2025

And the last one. I find myself judging him as an arrogant schmuck for thinking this was an appropriate thing for a birthday book (at least no evident pedophilia but that bar is so low as to be below ground)

Language is the one truly interesting thing that happened in the last 500 million years

180