I finished the second editing, and I’m now into copy editing. That’s also almost done, with just the last two chapters to go. Which means that, sometime in the next month, the book will finally, after a bit over 4 years, be complete!

I’ve decided to go the independent publishing route. The book market has changed a lot, even just since I started writing, and my book is at high risk of being too niche to support a run with a big, commercial publisher. Thankfully, in that same time, independent publishing has lost a lot of its stink as the book market has transformed away from mass distribution, so it’s no longer a low-status move to go independent. Thus, I’ll be retaining the copyright, publishing the book free online for anyone to read, and also have print and audiobook versions for those who would like that available for sale with most major book retailers.

Very excited to get the final version of the book into everyone’s hands, and then to get onto my next big project. Expect a full announcement on it soon, but to tease you now, it’s a conference, it’s about epistemics, and vibes, and it’ll be happening a little later this year. More details soon!

I’m a week in now, resting and recovering, but I sadly had to admit that I wasn’t going to hit my new, lowered target of just one blog post a month while I finish the book. Then it hit me, I could use Claude to do some research into concussions and write something short about that!

So I asked Claude to do some research into concussion recovery. Specifically, whether there’s anything useful I can do beyond the standard advice of “rest”. I already sleep and meditate, and I’ve been hearing more about psychedelics as treatments for brain injuries, so I had Claude to do a deep literature review on all three as concussion interventions. We focused on psilocybin because it’s the psychedelic with the most research about concussions available. The full report is here. Here’s what came out of it.

The three interventions are complementary, not redundant. All three reduce brain inflammation after injury, but through different biological mechanisms. Hitting the same problem from three independent angles is a well-established principle in pharmacology, and it tends to work better than hitting it from one angle three times as hard.

Each one is best at a different thing. Sleep drives the brain’s waste clearance system. During sleep the spaces between brain cells expand by about 60%, allowing fluid to flush out damaged proteins and metabolic debris, and nothing else does this. Psilocybin provides the most potent signal for growing new neural connections, with a single dose producing structural changes lasting over a month in mice. Meditation offers the best-evidenced stress and immune regulation, creating the upstream conditions that let the other repair processes work.

Psilocybin works around a problem the other two can’t. After brain injury, inflammation hijacks the raw materials your brain uses to make serotonin and diverts them toward toxic byproducts instead. This depletes serotonin while simultaneously causing further damage. Because psilocybin is chemically similar to serotonin, it can activate serotonin receptors directly, bypassing the broken supply chain entirely. No endogenous process can do this.

But the evidence is profoundly asymmetric. Sleep has robust clinical data and an irreplaceable biological role. Meditation has one concussion-specific meta-analysis showing moderate benefit and near-zero risk. And psilocybin has zero completed human trials in brain injury populations, with the strongest direct evidence being a single rat study.

So the actionable takeaways are, sadly, anticlimactic. Prioritize sleep above everything, and not just “sleep more” but actively protect it, because the injury itself disrupts the very sleep needed for repair. Meditate if you already do and consider starting if you don’t. And probably don’t take psilocybin for a concussion yet, because the biology is exciting but there’s no human data and there’re unknowns around safety in an injured brain.

The full research report with citations is here.

The first makes the claim that minimum viable AGI is already here. The second is my take on what to do about it.

I can’t understate how important it is to understand what’s going on right now with AI and the immediacy of the need for action.

That said, there’s only so much I can do about it, so I also have updates on the book.

Book Updates

I reached an important milestone yesterday: I finished revisions on Chapter 7, which is the last of the core chapters of the book. All that’s left is to do a revision pass on Chapters 8 and 9, which are “what’s next” chapters after having concluded the core arguments.

I have no idea how long they will take to revise. I hope not long, but they do contain a lot of words.

I really want to get the revisions finished and the book actively moving towards print publication soon, so I’m going to be taking a little break from blogging. Not entirely, but I’m going to switch to posting semi-monthly instead of weekly until I get all the revisions done.

Historically, I always write more than whatever target I set for myself, so I’ll probably write more. I just have to remember that the less I write here, the more progress I make on the book, and the sooner the book is done, the sooner I can focus on blog posts instead of book revisions.

I practice Zen in the lineage of Charlotte Joko Beck. Her teaching style was, for its time, radically non-traditional. In an era when talking too much about your inner thoughts and feelings was discouraged by first-generation Japanese-American Zen teachers, she believed Western students needed to practice a Zen that leaned on familiar psychological concepts to make sense. One of those concepts is what she called the “core belief”.

The core belief is a deeply held, usually unconscious belief about ourselves. It almost always feels like some flavor of “not enough”. It forms early, operates automatically, and powers the reactive, habitual, and often maladaptive patterns of behavior that make up most of what we call our personality.

The cruel trick of the core belief is that it reinforces itself. It tries to protect you from noticing anything that might confirm it, and by doing so actually generates more evidence in favor of it. For example, if you believe you’re unlovable, you might push people away so they can’t prove you aren’t worthy of love, or you might stay so anxiously close to them that no one has a chance to notice how they really feel about you. It’s a psychological trap that heaps suffering upon more suffering, and almost all of us have been ensnared in it since before we can remember.

Joko doesn’t advocate for getting rid of the core belief. In fact, she argues that would be impossible. Instead, it’s about becoming intimate with it, noticing it when it shows up, and learning to face reality rather than hiding from it. The more you do that, the less power the core belief has to control your life, and the more you are free of the suffering it causes.

But describing the core belief as a feeling of “not enough” is rather vague. You can sit for years, intellectually knowing you have a core belief, and never catch a glimpse of what your core belief really is. It’s possible to have so many layers of psychological barriers in place that you never allow yourself to see it.

In the Ordinary Mind Zen School that Joko founded, we practice getting past these barriers by paying attention to sensations in the body. Much like in Gendlin’s Focusing, we try to notice the physical sensations that arise when we react out of anger, fear, or desire. We become familiar with those physical feelings, then, let our minds name them. Sometimes the names we give provide surprising insights. Other times, nothing comes, and more noticing is needed. Over time, with the help of a skilled teacher, one can learn to work with their core belief and tease apart how it limits life.

Now, it’s pretty normal in Zen to do things like this from scratch with a minimum of conceptual frameworks. And I generally endorse this approach, but sometimes it’s helpful to get hints. Based on my understanding of the Enneagram, gleaned from Michael Valentine Smith’s series of posts on it last year (1, 2, 3, 4), I think its nine types provide a map to common patterns of core beliefs, and may help a person better practice with their core belief when noticing alone leaves them stuck.

The Enneagram is a Map of Suffering

Over the years, I’ve occasionally taken Enneagram tests, and every time I found the results unhelpful. I’d get categorized as some type, be offered some explanation of what it means, and while it seemed like it might be true, it all fell flat for me. Am I a 9? A 3? a 1? Who knows! The outcome seemed to change based on my mood. I had little reason to think that the Enneagram was useful.

Michael helped me see value in the Enneagram by reframing it, not as a personality classification system, but as a map to how and why we suffer.

In his framework, each person has what he calls “Essence”, which is something like your true nature, the awareness and aliveness you had before reactive personality took over. Essence naturally expresses certain qualities, like love, clarity, peace, power, and freedom. But when Essence gets overwhelmed in early life, it creates a mechanical personality to stand between itself and the world. That personality tries to mimic Essence’s qualities, but it can only produce toxic imitations, and those imitations create self-reinforcing downward spirals.

He tongue-in-cheek summarizes the Enneagram as asking: “In which of these nine ways are you most screwed up?”

Reading his posts, I couldn’t help but notice that Michael’s “downward spiral” was not too different from how Joko describes the workings of the core belief. In fact, I think they’re pointing at the same mechanism, but are coming at it from different angles.

The Enneagram says personality tries to replace an essential quality, and fails because the replacement is mechanical. Joko says that the core belief generates reactive patterns that try to protect us from acknowledging it. Both say that these behaviors create lock-in, double down on what’s not working, and create a self-reinforcing loop of suffering.

What’s neat about the Enneagram is that, unlike Joko’s intimately individual approach, it gives you a map to the essential qualities your personality is trying to mimic. If the parallel between the Enneagram and Joko’s teachings holds, then each Enneagram type corresponds to a class of core beliefs. I might phrase those as:

• Type 1: “I’m not good/right enough”

• Type 2: “I’m not lovable enough as I am”

• Type 3: “I’m not valuable enough without proof”

• Type 4: “My inherent worth is missing or damaged”

• Type 5: I’m not equipped enough to handle the world directly”

• Type 6: “Nothing is reliable enough to trust”

• Type 7: “What’s here isn’t enough”

• Type 8: “I’m not solid/real enough”

• Type 9: “Things aren’t okay enough to fully engage with”

Of interest to me is that this mapping can give greater specificity to Zen practice. It can be hard to simply sit with not-enoughness. You have only a vague idea what you’re looking for, and people are different enough that the way one person talks about their feeling of not enough may sound totally foreign to you. The Enneagram helps explain this, because different people really do have different styles of core belief that feel quite different from the inside.

That said, I see some danger in the Enneagram. It’s a system for putting names on things, and Zen is ultimately about seeing through how our mental constructs imprison us. The self is not a fixed thing. Our stories about ourselves are just more thoughts about whatever is really going on. The Enneagram risks becoming a new way of formulating a self to latch on to rather than a way to become free of it.

To be fair, Michael himself warns about exactly this in his series. People who get into the Enneagram often start trying to explain everything in terms of it, and then start contorting their behavior to fit their type. He recommends holding your type “extremely lightly” and measuring its value by a single criterion: does viewing yourself this way make your life more wholesome?

That’s a good test, and it’s the same test I think Joko would apply. Is your practice making you more open, more responsive, more alive to what’s actually happening? Or is it giving you a more sophisticated story about yourself? The Enneagram is useful exactly insofar as it helps you see through personality. It’s harmful exactly insofar as it helps you solidify it. If you find yourself using your type to explain your behavior rather than to notice and release it, set the Enneagram down.

Share

Finding My Type

After reading Michael’s series, I got interested in what type I might be, since if my theory was right, it will help me in my Zen practice. When I’ve taken Enneagram tests, I’d variously score as a 3, a 5, a 7, or a 9. And if I’m honest with myself, I see something of myself in all the types. Hard to do much with that!

But as Michael argues, the tests are just looking at surface-level traits and don’t do a very good job of detecting Enneagram type. What you actually have to do is figure out which type helps you unwind the reactive downward spiral. As I think of it, you need to ask yourself: which type’s need, if it were fully met, would make you truly and deeply happy, and not because your need was incidentally met, but because your need was met fundamentally?

This is easier explained with an example. As I said, I often test as various types. Sometimes I test as a 3, meaning I need to prove I can achieve greatness. Other times I test as a 5, meaning I need to show off how much I know. But notice how I phrased those. I didn’t say I desire achievement or knowledge, I said I need to prove/show off. And you know what type needs to demonstrate personal specialness? That’s right: type 4.

As I think of it to myself, I’m happiest when my inner nobility is allowed to shine. Everything else is incidental. I’m smart enough that I can let my nobility shine by showing what I know. I’m capable enough that I can let it shine through achievement. In fact, I can make any of the types fit so long as it’s a means to showing off my specialness. I feel like this explains a lot about me.

What’s interesting from a Zen perspective is how a type 4 core belief maps to the central misperception that practice addresses. The 4’s spiral is powered by a search for inherent worth that was never missing. That’s basically what “seeing your true nature” is all about in Zen: recognizing that what you’ve been searching for was here all along. All you have to do is stop searching, and you’ll find yourself!

Now of course, following Michael’s advice, I hold this all lightly. Maybe I’m wrong about being a 4. Maybe someday I’ll find it makes sense to think of myself as another type. The point is not to be identified with a type, it’s to use the type to make sense of myself and point the way to actions I might take that would make my life better.

And the same is true if you want to try using the Enneagram. I suggest reading Michael’s series, and if you’re interested in learning more about Joko’s idea of the core belief, I suggest picking up her most recently published book, Ordinary Wonder.

What’s the Great Matter? It’s the question you can’t answer. It’s the fear you feel when you contemplate your own death. It’s the void lurking at the center of your existence, and no matter how hard you try, you can’t see into it. You’ve lived with it for as long as you can remember, fighting against its friction with your every action. It can’t be pointed at directly, but you know it’s there because it’s in the place you can’t look.

That’s all rather mysterious, and thankfully some people have managed to grapple with the Great Matter enough to say a little more. They often talk about non-dual experience, of crossing the divide between the relative and the absolute. They say things like there’s no separate self and that all is impermanent. And if they’ve read the right websites, they might even describe the Great Matter as the contradiction inherent in the intuitive self-model, and say that to resolve the Great Matter is to have the insights that transition you into a persistent, non-symbolic experience of existence (PNSE).

Or, in short, to resolve the Great Matter is to become enlightened.

But the longer it’s been since the Great Matter was resolved, the less I like the word “enlightened”. I did a whole video essay about it, but in short, the main problem is that “enlightenment” implies a permanent state of attainment and carries misguided cultural connotations. Other words like “awakened” drop some of that cultural baggage, but unfortunately keep the attainment framing.

And you might be saying to yourself, why would an attainment framing be bad? Clearly something has been attained! But, the only thing attained is a resolution of the Great Matter itself, which is the act of having an insight and having that insight spread deeply through your entire being, nothing more. The rest has to come later. All that resolving the Great Matter does is remove the obstacle that was blocking the way on the path to freedom from suffering.

And if you want to be free of suffering, you have to put in the work. First, you have to put in the work to resolve the Great Matter, which often involves untraining many of your maladaptive habits. It requires grappling with your addictions, your traumas, your hangups, and your insecurities. And then, after the Great Matter is resolved, you have to continue that work, but now on a deeper, more subtle level than was possible before. It does feel easier, because you no longer have to deal with both the habituated mind and the Great Matter, but that feeling itself can become a trap that will ensnare you if you aren’t careful.

Now, some people who resolve the Great Matter do so while living in a monastery or hermitage, or move to one after such resolving. In such a place, one has the opportunity to minimize decision making and planning. And when there’s no decisions to make, there’s no need to model the self, and so the constructed self-model can be let go and forgotten, replaced by a mind that only concerns itself with this moment. That can be nice, especially if done for a limited time to strengthen one’s practice, but it comes at the cost of helping the world.

To help the world, you have to live in it, and that means constructing a self-model to make the predictions necessary to plan and take actions. The best one who has resolved the Great Matter can do, if they are committed to benefitting all beings, is to find a way to live holding that self-model lightly. To see, in each moment, that the self and the not-self views are equally real, for both are needed to tackle the great many challenges facing our world.

I’ve left out many details you’d probably like to know. That’s because I write this post with some hesitation. Although this is not the first time I’ve publicly admitted to being “enlightened”, it is the least obscure and most easily found. But I decided to take the risk, because I feel that it’s been long enough that I can speak confidently of my own experiences, however they are labeled, and because by speaking about those experiences, I may be able to help others.

When I had not yet resolved the Great Matter, I was very confused about it. Some people wrote some straightforward things that were confusing. Other people wrote confusing things that were straightforward. I needed all the help I could get from others pointing the way for me. Now it’s my turn to help others by pointing. This is one attempt, and I hope it proves useful.

But I think the word has legs. I just have to figure out what it actually means!

On the surface, it’s obvious. It’s the combination of “vibes” and “epistemics”, so more or less naming the core idea of the post/meta-rationalist project. But again, what does it actually mean? It’s easy to point at a large body of work and say “I don’t know, whatever the thing going on over there is”, but much harder to say what the thing actually is.

So to start, let’s talk about epistemics. What is it? I see people using the word two ways. One is to mean the way we know things in general. The other is to mean the way we know things via episteme, that is knowledge that’s reasoned from evidence, as opposed to doxa and techne and many other ways of knowing (if those Greek words mean nothing to you, I highly recommend reading the post at the link before continuing). Unfortunately, some people equivocate between epistemics-as-knowing and epistemics-as-knowing-via-episteme to give the impression that episteme is the only good way to know anything. That, to me, is a problem.

I think it’s a problem because such equivocation discounts valuable sources of knowledge that aren’t easily made legible. Now, to be fair, there’s some reason to do this, because the pre-rationalist epistemic stance says legibility doesn’t matter and logic is just a means to justify one’s preferred ends. The rationalist stance is largely that everything that can be made legible should be, and that which cannot be made legible needs to be treated with great caution because that’s how we slip back into pre-rationality. So I understand the desire to equate epistemics with episteme (and, etymologically, the English language tries very hard to do this), but I also find it frustrating because it encourages excessive devaluing of other ways of knowing, especially metis, techne, and other forms of knowledge that are less legible.

That’s where the vibes come in. They can rescue us from an excessive focus on episteme and temper the excesses of legibility. But what are vibes and how can they help?

Vibes are the embodiment of what we care about. The stoner, for example, has stoner vibes because they care about chilling and feeling good. The Christian has Christian vibes because they want to do what Jesus would do. And the rationalist has rationalist vibes because they care about knowing the truth with high predictive accuracy. For any vibe, there is always something the person expressing it cares about deeply that causes them to have that vibe.

This matters in epistemics because knowing is contingent on care. I make this argument in detail in Fundamental Uncertainty (currently in revision ahead of publication), but the short version is that we have a mental model of the world, truth is the degree to which our mental model is accurate, we want an accurate mental model because it’s useful, and usefulness is a function of what we care about, thus truth is grounded by and contingent on care. And since vibes are the embodiment of care, vibes have an influence on the act of knowing, hence, vibestemics.

(If this argument seems handwavy to you, it is. You’ll have to read the book to get the full argument because it takes about 10k words in the middle of it to lay it all out. If you want to read the first draft for that argument, it’s in Chapter 5, 6, and 7 which start here. Alternatively, although I think “Something to Protect“ does a poor job of emphasizing the epistemic relevance of care in favor of explaining a particular way of caring, I read it as ultimately claiming something similar.)

Share

Okay, but that’s the theoretical argument for what vibestemics is. What does it mean in practice? Let’s dive into that question by first considering a few examples of different epistemic vibes.

Woo: The epistemic vibe of woo is that whatever’s intuitive is true. Woo is grounded in gnosis and largely eschews doxastic logic and careful epistemic reasoning. That said, it’s not completely devoid of epistemics. It’s definitionally true that whatever you experience is your experience. Unfortunately, that’s roughly where woo stops making sense. It interprets everything through a highly personal lens, so even when it leads to making accurate predictions, those predictions are hard to verify by anyone other than the person who made them, and woo-stemics easily falls prey to classic heuristic and bias mistakes. This severely restricts its usefulness unless you have reason to fully trust yourself (and you shouldn’t when it comes to making predictions).

Religion: The vibe of religion is that God or some other supernatural force knows what’s true. Knowledge of what God knows may require gnosis, or it may be revealed through mundane observations of miraculous events. Although not true of every religion, religious epistemics can be a friend of logic, and many religions demand internal logical consistency based on the assumptions they make. Sometimes these theological arguments manage to produce accurate world models, but often they have to be rationalized because the interpretation of the supernatural is fraught and we mere mortals may misunderstand God.

Science: Science as actually practiced by scientists involves empirically testing beliefs and updating them based on evidence. The vibe is pragmatic—build hypotheses, test them, see what happens, and revise accordingly. The only problem is that science requires the ability to replicate observations to determine if they’re true, and that’s where it hits its limits. When events can’t be observed or can’t be replicated, science is forced to say “don’t know”. Thus, science is fine as far as it goes, but its vibe forces it to leave large swaths of the world unmodeled.

Rationality: The vibe of rationality is to be obsessed with verifying that one really knows the truth. This has driven rationalists to adopt methods like Bayesian reasoning to make ever more accurate predictions. Alas, much as is the case for science, rationality struggles to deal with beliefs where predictions are hard to check. It also tends to smuggle in positivist beliefs for historical reasons, and these frequently result in an excess concern for belief consistency at the cost of belief completeness.

Post-rationality: The post-rationality vibe is that rationality is great but completeness matters more than consistency. Thus it attempts to integrate other ways of knowing when episteme reaches its limits. Unfortunately, how to do this well is more art than science, and there’s a real risk of getting things so wrong that a post-rationalist wraps back around into pre-rationality. Arguably this is what happened to the first post-rationalists (the postmodernists), and it continues to be a threat today.

What I hope you pick up from these examples is that different epistemic vibes are optimizing for different things and making different tradeoffs. Although it may seem strange, especially if you’re a rationalist, that someone could have a good reason to ignore predictive accuracy in favor of intuition or dogma, for those with woo and religious vibes that choice is locally adaptive for them. They similarly look back at you and think you are deeply confused about what matters, and this is a place where arguments about who’s right will fail, because they’re ultimately arguments about what each person values.

All that said, it’s clear that some vibes are more epistemically adaptive than others. Accurate world models convey real benefits, so adopting a vibe that leads you to develop better world models is usually a good move. This, incidentally, is what I would argue is the pragmatic case for post-rationality over rationality: it’s rationality plus you can break out of the rationalist ontology when it’s adaptive to do so (though admittedly at the risk of it becoming rationality minus the guardrails that were keeping you sane).

And this ability to shift between vibes is why I think having a word like “vibestemics” is valuable. When we can only speak of epistemics, we risk losing sight of the larger goal of living what we value. We can become narrowly focused on a single value like accurate model prediction, Goodhart on it, and forget to actually win. We can forget that knowledge and truth exist to serve us and our needs, not the other way around. Vibestemics invites us to know more and better than we can with episteme alone, if only we have the courage to let our grip on a single vibe go.

Chapter 7, as you may recall if you read the first draft, is both the “cybernetics chapter” and the “tie everything together” chapter. Originally it was largely based on the two posts where I first worked out these ideas, but as I’ve been revising, I discovered that it contained both a lot of extraneous material and didn’t have quite the right focus for where it sits in the book. These were both problems I knew about when I wrote the first draft, and now in the revisions I have to solve them.

As a result, it’s been a slog to find the right way to present these ideas. I’ve tried maybe 5 different approaches. It takes time to develop them out enough to see if they work. I’m hopeful that the 6th approach will be the final one, but it’s not done yet, so no promises.

Medium

Hey, did you know I used to run a blog on Medium called Map and Territory? It originally started as a group blog for some folks in the LessWrong 1.0 diaspora, but the group aspect quickly collapsed after LessWrong 2.0 launched, so then it was just me. (All my posts from it are now mirrored on LessWrong since I trust it more than Medium in the long run.)

Anyway, every few months somebody, usually this guy, references my most popular post from the Map and Territory days. It’s titled “Doxa, Episteme, and Gnosis”, and it still gets about 100 new reads a week all these years later. I’ve tried a couple times to write new versions of it, but they never do as well.

The “Many Ways of Knowing” post from two weeks ago was the most recent evolution of this post, though this time excerpted from the book. I like it, and I think it fits well in the book, but it still doesn’t quite capture the magic of the original.

The original succeeds in part, I think, because I was naive. I presented a simple—and in fact over-simplified—model of knowledge. It’s accessible in a way that later revisions aren’t because it’s “worse”, and I suspect it’s helped by putting three Greek words in the title, which I am pretty sure helps with SEO from students trying to find out what these words mean.

Anyway, this is all to say I got some more posts lined up, and hopefully I’ll at some point be naive enough to write another banger.

Back in October I shared how I write code with Claude. A month later, how I write blog posts (though ironically not this one). But I wrote those before Opus 4.5 came out, and boy has that model changed things.

Opus 4.5 is much better than what came before in several ways. It can think longer without getting lost. It’s much better able to follow instructions (e.g. putting things in CLAUDE.md now gets respected more than 90% of the time). And as a result, I can trust it to operate on its own for much longer.

I’m no longer really pair programming with Claude. I’m now more like a code reviewer for its work. The shift has been more subtle than that statement might imply, though. The reality is that Claude still isn’t great at making technical decisions. It’s still worse than random chance at picking the solutions I want it to pick. And so I still have to work with it quite closely to get it to do what I want.

But the big change has been that before I would have the terminal open, claude in one split in tmux, nvim in another, and we’d iterate in a tight loop, with claude serving as something like very advanced autocomplete. Now, I use Claude in the desktop app, get it to concurrently work on multiple branches using worktrees, and I have given it instructions on how to manage my Graphite stacks so that even for complex, multi-PR workflows I usually can just interact through chat rather than having to open up the console and do things myself.

Some tooling was needed to make this work. I had to update CLAUDE.md and write some skills so that Claude could better do what I wanted without intervention. I also had to start using worktrees, and then in the main repo directory I just check stuff out to test it (the local dev environment is a singleton) and do occasional manual operations I can’t hand to Claude (like running terraform apply, since I can’t trust it not to randomly destroy infrastructure on accident).

Still, this is not quite the workflow I want. Worktrees are annoying to use. I’d prefer to run Claude in cloud sandboxes. But the offering from Anthropic here is rather limited in how it can interact with git, and as a result not useful for me because it can’t use Graphite effectively. Graphite has their own background agents, but they’re still in beta and not yet reliable enough to use (plus they still have restrictions, like one chat per branch, rather than being able to have a chat that manages an entire stack).

But as I hope this makes clear, I now use Claude Code in a more hands-off way. My interactions with the code are less “sit in an editor with Claude and work in a tight, pair-programming-like loop”, and more “hand tasks to Claude, go do other things, then come back and review its work via diffs”. I expect this trend to continue, and I also hope to see new tooling that makes this workflow easier later in the year.

Share

That’s coding, but what about writing?

Well, Claude still isn’t fantastic here. It’s gotten much better at mimicking my style, but what it produces still has slop in its bones. It’s also gotten better at thinking things through on its own, but I still have to work to focus it and keep it on task. It will miss things, same as a human would, that I want it to look at.

For example, I was editing a paragraph recently. I made a change to some wording that I was worried might convey the right sense but be technically wrong. I handed it to Claude. Its response was along the lines of “yes, this looks great, you made this much more readable”. But when I pressed it on my factual concerns, it noticed and agreed there was a problem more strongly than I did! These kinds of oversights mean I can’t trust Claude to help me write words the same way I trust it to help me write code.

So I’m still doing something that looks much more like pairing with an editor when I write with Claude. This is good news in some sense, because it means I’m still needed to think in order to produce good writing, but bad news if you were hoping to automate more thinking with Claude.

This past week there came news of some novel mathematical breakthroughs using LLMs. The technology is clearly making progress towards critical thinking in a wider set of domains. And yet, writing remains a nebulous enough task that doing it well continues to evade Claude and the other models. That’s not to say they aren’t getting better at producing higher quality slop, but they still aren’t really up to completing a task like finishing revisions on my book the way I would want them done for me.

Subscribe now

Where does this leave me feeling about LLMs right now?

We made a lot of progress on utility in the last 12 months. Last January I was still copy-pasting code into Claude to get its help and using Copilot for autocomplete. It was almost useless for writing tasks at that point, and I often found myself wasting time chatting with it trying to get things done when it would have been faster to sit and think and do it myself. That’s just no longer true.

As always, though, I don’t know where we are on the S-curve. In some ways it feels like progress has slowed down, but in others it feels like it’s sped up. The models aren’t getting smarter faster in the same way they were in 2024, but they’re becoming more useful for a wider set of tasks at a rapid rate. Even if we don’t get LLMs that exceed what, say, a human ranked in the 70th percentile on a task could do, that’s already good enough to continue to transform work.

2026 is going to be an interesting year.

What does it mean to say “I know”?

This might seem like a strange question to ask since knowing is such a fundamental and intuitive activity. It’s hard to imagine being a person and not knowing things. In fact, the only time in our lives when we aren’t swimming in a sea of knowledge is when we’re newborns, and we quickly wade in by learning the voice, face, and touch of our caregivers. Within days we’re able to tell them apart from strangers, and our long relationship with knowledge begins. So if we want to understand what it means to “know”, we’re going to have to spend some time exploring how we use this almost invisible word.

When we say to each other that we “know” something, we generally mean that we’re able to recall an idea, hold it in our mind, reason about it, and say things based on understanding it. Consider these examples of things I claim to know:

• I know my friend Eric.

• I know how to tie my shoes.

• I know how to speak English.

• I know that Paris is the capital of France.

• I know what it’s like to ride a rollercoaster.

• I know that if I eat too much food I’ll get a stomach ache.

Although all of these sentences begin “I know,” the knowledge expressed in each is not the same. Knowing that Paris is the capital of France is knowing a propositional fact. When I say that I know my friend Eric, though, I’m not claiming to state a proposition, but rather that I can recognize him by sight and sound and am familiar with his patterns of behavior. There’s also no propositional fact about what it’s like to experience the thrill of riding a roller coaster: it’s a lived experience that simply is. Rather than using “know” to mean many things, perhaps it would be useful to have different words for these different forms of knowledge.

The ancient Greeks did exactly that. They used multiple words to break down knowing into several categories, including:

• episteme: things you know because you reasoned them from evidence, like knowing that water boils at 100 degrees celsius because you boiled a pot of water using a thermometer to see when it started to boil

• doxa: things you know because others told you, like knowing what happened at a party you didn’t attend because your friend tells you

• mathema: things you know because you were educated in them, like knowing how to spell words because you were taught them in school

• gnosis: things you know through direct experience, like how it feels to jump in a lake

• metis: practical wisdom, usually collectively constructed from many people’s experiences, and shared with others, often starting at a young age, like how you know to look both ways before crossing the street

• techne: procedural knowledge that comes from doing, like the “muscle memory” of how to ride a bicycle

These categories aren’t perfectly distinct, though, because the same information can be known multiple ways. For example, Ada recently saw Ed wearing a traffic cone on his head. Later, she told Grace about it. As a result, Ada had gnosis of Ed’s appearance, while Grace had doxa of it. And because nothing is ever simple, Grace also eventually saw Ed wearing the traffic cone on his head and gained gnosis of her own in addition to her pre-existing doxa.

Does this mean Grace now knows more than Ada does? Maybe, if the combination of gnosis and doxa provides a deeper understanding than gnosis alone can. Or maybe gnosis trumps doxa and the value of doxastic knowledge is lost once gnosis is gained. Whatever the answer, that we can ask this question at all shows that the lines between different kinds of knowledge aren’t always clear. Perhaps that’s why English and many other languages have collapsed these distinctions into a single word for knowing.

Subscribe now

Yet, sometimes we need to make distinctions between different ways of knowing, as political science professor James C. Scott does in his book Seeing Like a State. In it, he examines how modern and historical governments have differed in their treatment of knowledge. He then goes on to explore how those differences have had significant impacts on people’s lives.

Prior to the modern era, states placed a high value on metis. They frequently saw themselves as the defenders of tradition, often to the point of killing those who dared to challenge the established order. Modern states, in contrast, often throw tradition out in favor of rational, scientifically-grounded episteme. By prioritizing episteme over metis, modern states have created a bounty of benefits for the people living in them, including more reliable food supplies, better medicine, and increased access to what were previously luxury goods. But, as Scott explains, these benefits aren’t guaranteed, and sometimes overturning tradition leads to disastrously worse outcomes.

In the middle half of the 20th century, there were numerous attempts to modernize agriculture. Many of them ended in failure. Between the 1920s and the 1970s, Russia, China, and other communist countries saw catastrophic famines resulting from forced collectivization, unrealistic production targets, and bad agricultural science. On this last point, it was the misguided theories of Soviet scientist Trofim Lysenko that magnified what might have been survivable food shortages into widespread, mass starvation.

Lysenko trusted his own episteme over everything else. He claimed that his methods were more rigorous than established science, rejected the accumulated practical knowledge of farmers, and dismissed all critics as ideologically impure. He treated his own reasoning as the only valid path to truth. When communist leaders implemented his ideas alongside their political programs, the resulting crop failures led directly to the deaths of tens of millions of people.

Over-reliance on episteme was not a problem unique to communist countries, though. In the United States, the Dust Bowl and associated food scarcity of the 1930s was the direct result of careless industrialization of farms in the 1920s. Early agricultural science thought it could achieve higher crop yields simply by working the land harder, and for about a decade this was true. Using mechanized plows and harvesting equipment, farmers converted a hundred million acres of Great Plains prairie from low productivity cattle pasture to bountiful fields of wheat and corn.

Scientific farming was a boon to the people of the United States, right up until drought hit in the 1930s. The intensive agriculture of the previous decade had damaged the natural ability of the land to protect itself, and overworked fields crumbled to dust as they baked under cloudless skies. The situation only began to turn around when rain finally returned in 1939, but it took decades for the land to fully recover, and even longer for farmers and scientists to develop and widely adopt sustainable farming practices that work in the Great Plains.

Do we know better now? We like to think so. Certainly we’re less naive because we tell each other about the failures of the past (metis) and learn about them in school (mathema). Politicians no longer propose that we solve all our problems with science, and we know that episteme doesn’t have a monopoly on the truth. Yet even as we’ve learned from our mistakes, we risk overcorrecting and forgetting the power of episteme. In fact, many would say we still don’t value episteme highly enough and too often ignore scientific results that offer us clear ways to improve the world.

But it’s not that we need to spend more effort finding the right balance between different forms of knowledge. Rather, we need to integrate all our knowledge together, whatever the source. When we privilege one way of knowing over another, we risk incomplete understanding that leaves out things we could have known. We must instead strive to update our beliefs in proportion to all the evidence we have available, and by so doing build the most accurate models of the world that we can.

Share

So I was naturally excited when Apple TV teased a new Fraggle series and delivered it to us in early 2022.

Unfortunately, I was deeply disappointed by what we got. I’ve made it through every episode of Back to the Rock, but it was a painful exercise in patience. After watching the Christmas special they released this year, I’m starting to get a sense for what feels so wrong about the new series to me. I’ll see if I can work it out below, but apologies in advance if this descends into ranting about how Apple TV is ruining my childhood.

NB: Obviously, there will be spoilers, though I’d be surprised if anyone cared. Also, I’m going to assume a fair degree of familiarity with the source material for the rest of this post. Sorry if that means it doesn’t make much sense to you, but if that’s the case then you probably didn’t care that much about what happens to Fraggle Rock anyway.

Setting

Back to the Rock (BttR hereafter) is nominally set in the same universe as the original series and is a continuation of it. The reality is more complicated.

There is some continuity with the events of the original series, and those events are occasionally referenced. We also have a bunch of changes that break continuity, which is terrible and confusing. Thus, BttR is neither a reboot nor a sequel, but something far worse: it’s fan fiction.

For example, Sprocket is somehow both the same character as before and not, so he chases Gobo but also knows him and seems to remember having lived with Doc in the workshop, which means he should already be friends with Gobo. Similarly, Junior Gorg is both back to having an interest in catching Fraggles, but also is now much friendlier to Fraggles, and quickly switches to being friends with them instead of needing multiple seasons of stories to change his heart. Every character, even minor ones, has problems like this.

The timeline is also unclear. In some sense, the show takes place approximately 30 years after the end of the original series, reflecting the real-world timeline. But also somehow Sprocket is still alive, as are the individual Fraggles from the original series, and none of them have really aged (though maybe this is fine for Fraggles since we have little information on exactly how long they live).

Now to be fair not all the changes are actual consistency issues. The caves in the Rock look different and have a different layout connecting them. This is fine because the original series already established that sometimes the caves in the Rock move and change on their own. The Gorg’s castle looks different, which is a little more concerning because the Gorgs are established to live for thousands of years and are resistant to change, but perhaps we can give them the benefit of the doubt here given how the original series ends.

There are also some general changes to the show’s color palette and the characters’ designs, but they reflect general trends with the Muppets and are not specific to Fraggles or BttR, so I’ll ignore them here.

Characters

I could probably ignore the setting and continuity issues if those were BttR‘s only problems. Alas, the changes to how the characters are written is where things really go off the rails.

Let’s take the example of Gobo. His defining characteristic in the original series is bravery. He often ends up as the main character because he’s the only one courageous enough to take on some responsibility. Of course, he also suffers the follies of excess bravery, which makes for interesting stories, but the Fraggles depend on him to be the humble hero.

In BttR, he’s still sometimes put into this role, but I get the sense that the writers want us to feel bad about it. His dialogue is a bit different. The fear and uncertainty and awkwardness is turned up. He’s more brash chungus than brave hero. He’s often not allowed to solve his own problems. He doesn’t rely on his friends so much as get saved by luck channeled through them.

Or take Cotterpin. She was originally a fish-out-of-water Doozer who ends up apprenticed to the Architect and is in line to replace him. We can reason that since BttR is 30 years since the original series she’d be further progressed in her career, and she is, but in ways that don’t feel like real development. She’s often presented as being the only Doozer with any real smarts. The Architect is still her boss, but now no longer a source of wisdom to learn from. He’s almost always played as an old fool who is somehow still in charge.

But this doesn’t seem right given what we know of Doozer society. It’s very competence focused. If the Architect can’t do the job, I’d expect him to give it up to Cotterpin when that became clear. Instead, he’s just an empty boss character who wields nominal power that Cotterpin gets to react against. She gets to be a girlboss with no actual purpose to it. In the end, we never feel like Cotterpin is actually competent, just the least dumb of the now much dumber Doozers. I’d much rather have seen a BttR where Cotterpin is now the Architect and must deal with the responsibility and accountability that comes with leadership.

And then we have the new characters. I’m just gonna pick on the most poorly written one, Pogey. Pogey is something like a token diversity character. I think they’re supposed to have autism and ADHD or be otherwise neurodivergent, they come out as non-binary in one episode and then this fact is never referenced again, and they also get increasingly inserted into stories for no purpose other than being there. Pogey is also annoying—not just to me but to the Fraggles—and we see the other Fraggles taking pains to tolerate them.

I think there’s a version of Pogey that could have been good. Maybe we could have had a version that just has autism or ADHD or is non-binary. Great! Then we get to see stories about how sensory issues or trouble focusing or being outside the gender binary affects their life. But instead I get the feeling that someone in the writers’ room really wanted to make a character that demonstrates intersectionality and intersected everything they could into a single character, ruining what makes Fraggles Fraggles in the process.

See, the thing about Fraggles is they all have a thing. Gobo is the brave Fraggle; Red is the competitive one; Wembley is indecisive, and so on. The intersectionality happens on the group level, because individual Fraggles aren’t, in some important sense, whole people. Instead, the wholeness comes from their interactions. Boober, for example, is not so much a full character as he is the archetype of anxiety and fear, and he serves the same role in Fraggle society that anxiety serves in our own minds. Yes, he is presented as a full character, but it’s clear that he, and all the other Fraggles, would live a kind of half life if left on their own. They only get to be whole by being part of a whole. That’s their beauty.

But the writers of the new series destroy this by misunderstanding what Fraggles are. They try to write them like they’d write humans instead. The result is, to me, disappointing.

Songs

Most of the new songs suck. This is not worth getting much into, but they mostly suck because their lyrics need to reflect how the characters are written (already discussed) and what the characters are doing (we’ll get to that next). Also, I think too many of the new songs aim to sound contemporary rather than fitting the folksy, out-of-time style common in the original series. Luckily for me, the new writers seem to be adding fewer songs as the show goes on, so maybe this problem will go away on its own.

Plot

Given how the characters are written, it’s not surprising that the episode plots in BttR also have issues.

In the original series, an episode generally has a plot built around some kind of universally applicable lesson, like about sharing or apologizing or dealing with difficulty. Some of these were quite intense, like in one early episode where Red and Boober almost die in a rock slide and grapple with the potential of their own death before they are saved. The stated goals of the original series were to promote world peace and to teach kids from any nation or culture about empathy, interconnectedness, and tolerance. In BttR, this seems to be forgotten.

The episodes in the new series are generally part of larger story arcs. The individual episodes and the arcs have lessons to teach, which is fine and good for this kind of show. Unfortunately, usually those lessons are progressive or otherwise politically leftist. I’m sure I will take some flak for this because in the episodes the lessons are presented as if they are universal, and in some sense they are, but the details make it clear that the writers have a political agenda that is not just empathy, tolerance, and peace, but empathy, tolerance, and peace on their terms and their terms alone.

Consider the main arc of Season 1. The Doozers discover a new building material that lets them be more productive, but it turns out to pollute the water supply. This destroys Craggle Lagoon, displacing the Craggles who must come live as refugees with the Fraggles. The pollution also harms the Merggles downstream. Resolving the issue requires everyone to work together and convince the Gorgs to destroy their fountain so water can flow fast enough to clear the pollution. Meanwhile, in the “Outer Space” segments, the new Doc is a PhD student researching how to remove microplastics from the ocean.

Now, none of this is objectionable on its face. The original series was always about interconnection, and “pollution is bad” isn’t a partisan position. But notice the specific shape of the story: industrial productivity causes environmental harm, which creates displaced populations, which requires collective sacrifice from those with more resources (the Gorgs and their fountain). The Doc segments make the real-world referent explicit. It’s not teaching kids that actions have consequences or that we depend on each other—lessons that would resonate across political lines. It’s teaching a particular story about how environmental problems arise and how they must be solved, one that maps cleanly onto progressive climate politics. The universality has been lost.

There’s a world where this story was written differently. If the writers had Abundance politics, rather than the Gorgs sacrificing their fountain, they enlist the Doozers to help them find a technical solution that makes everyone better off. Or maybe there’s a conservative version where everyone RETVRNs to the way things were before the Doozers started building with their new materials. There’s any number of ways to tell the story differently, and those differences would mostly reveal different political biases.

There’s also a world where BttR tells a story about pollution and interconnection that’s more abstract. Most likely this wouldn’t be a long, multi-episode arc, which is perhaps why those were rare in the original series. Instead we’d have something like a one to two episode arc about Doozers polluting the water and then finding some clever compromise solution to deal with it. It doesn’t push the same kind of political message, which I think the writers likely explicitly wanted to do, but it does better serve the original spirit of what the Fraggles are.

Conclusion

I really wanted to like BttR. I’ll honestly probably keep watching whatever new Fraggle Rock content Apple produces even if I don’t like it. But I wanted to write this anyway because I wanted to at least lament what we lost.

Fraggle Rock was, like many of Jim Henson’s projects, meant to be for everyone, no matter their birthplace, language, or culture. The show was even localized, and if you look you can find alternative versions of Doc and Sprocket, just like you can find alternative versions of Big Bird on Sesame Street. What I love about it is that it’s a show that’s ultimately trying to teach compassion, and that’s something I think we desperately need more of in the world.

Unfortunately, BttR turns Fraggle Rock into yet another front in the political battle over everything. It’s now a Blue Tribe show for Blue Tribe kids, because don’t you know we can tolerate anyone except the Red Tribe, exactly what Fraggle Rock was meant to work against. I’m saddened that we got BttR instead of something that honors and extends Fraggle Rock’s legacy.

Luckily I, and you, can keep watching the original series whenever we want.

But it was also a good year. I curated my mom’s art. I built an AI oracle. I wrote 37 blog posts, gained 500 MMR in DoTA2, lost 10 pounds, volunteered at 2 conferences, and revised 5 book chapters to make them much, much better. And none of that is to mention all the quality time I got to spend with friends and family and all the cool places I got to visit.

Year boundaries are a good time for setting goals. Here are mine for 2026:

• finish revisions on Fundamental Uncertainty and get it into print

• run a conference at Lighthaven (details still in the works, more to come)

• continue to do whatever I usefully can to prevent existential catastrophes

• live my life well and love all the people in it

Although I have plenty of reason to worry for the future, I’m generally hopeful, and I look forward to seeing how things unfold in the year to come!

We had to let Sammie go last Wednesday.

The lump first appeared on her nose sometime in the Summer of 2024. It was small. You could miss it if you didn’t know her face.

The vet did a biopsy. Melanoma. We were told the median survival time was 6 months.

Over the next 18, we did everything we could to make her life comfortable. When the tumor grew too large, we’d have it debulked. When she couldn’t clean herself anymore, we brushed her and gave her baths. In the end, we had to hand feed her when the tumor, now inoperable, got in the way of eating on her own.

But we couldn’t feed her enough that way, and she was losing weight. The tumor was getting larger and soon would spread into her bones. Although she was always happy to cuddle, in the last week she spent most of her hours curled up in a heated bed, tolerating a pain she had no way to avoid.

She couldn’t use words to tell us her wishes. We had to guess. But if I were her, I’d want to be spared the pain and indignity that so often comes at the end. So we made the decision to let her go.

We made some art with her. We said our good-byes. And then, on a sunny afternoon, the vet let her slip off, sitting in my lap, purring to the last.

I adopted Sammie with my first wife, Sarah.

Sammie came to us from the SPCA. She picked us out. She jumped on Sarah and began snuggling immediately. I filled out the paperwork while she sat on Sarah’s shoulder.

Sammie was born on Halloween, so her full name is Samhain. But, as a joke, since many people mispronounce Samhain as /sam-hane/, we shortened it to Sammie.

As cat dynamics shifted in the house and Sammie grew from kitten to young cat, I became her favorite. I’d come home from work and she’d snuggle with me as I played on the computer or watched baseball. Many nights we’d fall asleep together on the couch with the TV still on.

After Sarah and I split, I moved to California, and Sammie came with me. The vet gave her Valium to fly. It made her extremely cute, extremely floppy, and extremely cuddly.

Sammie and I made a new life in Berkeley. We developed a routine. I’d wake up early and come home late from work. She’d climb in and snuggle with me in bed. If I played video games, she’d curl up on the computer, enjoying its warmth. If people came over, she’d wander out and demand their attention. I tried to take her outside a few times to enjoy the weather, but she wasn’t having it, so an indoor cat she stayed.

We moved several times to different houses with different people. Sammie loved everyone she met. And then she met Joy. The first night Joy slept over, Sammie snuggled her all night. Sammie was sure that Joy should be my wife.

In her last two years, Sammie got to know the fog of San Francisco after we moved to a house atop Twin Peaks. She looked forward to the afternoons when the clouds parted, the sun streamed in, and she could stretch out and bask in its warm, amber glow.

Now she’s cold. Very cold.

Like me and my wife, Sammie was signed up for cryonics. This is her new forever home:

I don’t know if we’ll ever get to see Sammie again. Maybe one day we’ll have the technology to bring her back. Or maybe we won’t. It’s something I can’t now know.

But she’s got a chance, which is more than most get. I’m glad she was able to take it.

Or at least, I did for one definition of “oracle”.

It’s called Kaku-Ora, and it’s an AI divination oracle inspired by the likes of the I, Ching, but trained on Zen koans. You ask it a question, it treats your question like the first half of a koan. It then gives a response and offers a capping verse to “explain” the response.

Why would I do such a thing? Several reasons:

• I find divination oracles like the I, Ching and Tarot interesting, not because I think they work by some supernatural means, but because injecting randomness into our thought processes can induce insight.

• I wanted to get some hands-on experience training AI models.

• Diffusion models are interesting!

• In particular, diffusion models seem like a good fit for mimicking how koan work is done.

• I have extremely vague ideas that it might be a good idea to teach AI dharma, as this might teach them compassion, and sufficiently compassionate AI might be aligned.

What follows is the story of how I built Kaku-Ora, and the many mis-steps I made along the way.

First, Data

To even get started I needed a corpus of training data.

Luckily, most core Zen texts are online. Unfortunately, there’s just not that much of it because Zen emphasizes transmission outside scripture. Also, a lot of the text is crammed inside low quality PDFs.

Thankfully, Claude turns out to be really good at creating training data (a fact that’s unsurprising once you think about it). I was able to hand it large chunks of text, give it a few examples of what I wanted, and in the end got nearly 20k examples in JSONL format—~5k “clean” examples from the koan literature, and another ~15k from non-koan Zen exchanges I found in texts.

Choosing Diffusion

I picked diffusion because it isn’t autoregressive. Most language models generate output one token at a time, in order. Diffusion models are different. They start with noise, then iteratively refine the whole output to remove noise until something coherent is revealed.

This kind of approach seemed like a better fit for koans. Koans don’t unfold through linear logic. They come from an expression of direct experience. When a response comes to you, it’s as if it appears, wholly formed, from the void. My theory is that diffusion better replicates that.

To be clear, I don’t think that diffusion can really do, on its own, what a human does when they work with a koan. But compared to other language generation approaches, it seems like a better approximation, so it’s the class of model I set out to train.

Continuous Diffusion from Scratch

My first attempt followed the Diffusion LM paper. The idea was to embed text as vectors in a continuous space, add noise, and train the model to denoise. So basically image diffusion, but for language.

I spent a week doing training runs. At one point I even tried fine tuning on an existing diffusion language model. In the end, I got models that mostly refused to generate anything, and when they did, produced long strings of punctuation.

The problem was that 20k examples was way too small to train continuous diffusion from scratch. Even with tricks I did to get the training set closer to 40k, it wasn’t enough. The original Diffusion LM work required a large data set and converged very slowly. I was doing training runs on my MacBook with a small corpus. It was a good idea, but I didn’t have the data to make it work.

Masked Diffusion

If I learned anything from my CS degree, it’s that discrete is better than continuous, so I tried discrete diffusion next, specifically masked discrete diffusion (MDLM). The way this works is to train the model on examples that have some of their tokens masked, like “The [mask] of [mask] is [mask] [mask]”. The model learns to predict what token is behind the mask, and then at generation time it iteratively unmasks a string of masks of the desired maximum length to reveal the output.

This worked better. The model produced actual words instead of just punctuation. But it fixated on common tokens. Lots of “the the the...” and similar repetitions.

I thought this approach had promise, so I next tried bootstrapping with GPT-2’s pretrained embeddings, hoping to give the model a head start. It helped a little. Now instead of repeating “the,” it would repeat meaningful words, like “the mind mind mind mind...”. Sadly, still not good enough.

At this point I decided to give up on diffusion. Maybe there’s a way to make it work, but the evidence was rolling in that my data set was too small and existing diffusion text models too primitive for me to fine tune on it. It’s possible that, given enough compute, I could have fine tuned a large diffusion language model, but that was going to require hundreds of hours to run on my laptop, so I started exploring other options.

Hi, BART

Masked diffusion was showing promise, even if it wasn’t producing the results I needed. Maybe the problem wasn’t masking, but diffusion. What if I tried a model that’s really good at masked prediction?

BART is a transformer model trained by corrupting text and having it learn to reconstruct it. And unlike GPT, which predicts the next token, BART predicts missing tokens anywhere in the sequence because it denoises, similar to what MDLM does, but with a different model architecture (encoder-decoder).

It was clear I didn’t have enough data to train a BART-style model from scratch, so I tried both fine-tuning and LoRA, a process similar to fine tuning that trains a small layer to transform model weights rather than updating the weights directly. The results were better than MDLM with GPT-2 embeddings, but still not useful. The model wasn’t quite as repetitive, but the responses still weren’t coherent, producing output like “the mind of Buddha is the mind of Buddha is the mind of Buddha is the mind of Buddha.”

Maybe the issue was that denoising, while it could work in theory, wasn’t the right approach for my open-ended, Q&A task. Maybe what I needed was to try a model specifically trained on Q&A and then fine tune it. So, that’s what I did next.

Koans, Decoded

Flan-T5 are a class of language models that put an encoder in front of a decoder, and they’re frequently used to train models that answer questions. Keeping in mind I’m doing all the training on my MacBook, I tried Flan-T5-small, which weighs in at just 77 million parameters.

It worked almost immediately.

It took about two hours of fine tuning against my full dataset to produce the Kaku-Ora model. No fancy tricks, just simple, straightforward fine tuning.

Could I get better output with a bigger model? Probably. Maybe at some point I’ll go back and try it. But the quality of the output was already sufficient that I was willing to call this experiment a success. You can see the kind of output it produces by trying Kaku-Ora for yourself.

The Capping Verse

If you’ve used the I, Ching or read a koan collection, you know that they contain more than hexagrams and koans. They also contain analysis. In koan collections, this typically takes the form of a capping verse—a short poem that elucidates the essential matter addressed by the koan.

I wanted that for my oracle. I was starting to envision deployment. A small website. You enter a question. The oracle contemplates. Then you get its response, and are also supplied a verse to help you make sense of it.

I initially used Claude Haiku for the verses. The quality was great. Haiku understood the genre and produced genuinely poetic commentary. But running Haiku costs money, even if not a lot of money, and I didn’t want to worry about abuse if I made the oracle publicly available.

So I switched to Qwen2.5-1.5B-Instruct, which runs free on HuggingFace. The verses are worse. If I cared to invest more time to make sure nobody can spam the oracle and run up my Anthropic bill, I’d have used Haiku.

Deployment

Model trained, how do I get it out in the world? The easy option was to run on HuggingFace since I have no need to keep the model weights secret. And running there on CPU is free. Good enough.

I then worked with Claude to vibe code a site. Much of the value of an oracle is in your interactions with it, so it was worth putting some effort into presentation.

What I Learned

Much of the point of training this model was to get more hands-on experience training AI. The trouble has been that, at work, RAG + hosted LLMs are more than enough for most tasks, and while it’s great to get the job done, it doesn’t help me learn the fundamentals.

Some things I really know now, having trained Kaku-Ora:

• Data matters a lot.

• The right model architecture also matters a lot.

• Insufficient scale is a problem, but also a chance to get creative.

• Getting a model to work at all is hard.

• Refining it is equally hard (but in a different way).

• Diffusion is hard because it’s slow to converge, even with good data.

I also came away with a sense that, if I wanted to make Kaku-Ora better, I could probably do it by investing more effort in a few areas:

• data cleaning

• more data

• training on larger models

What’s Next?

I don’t know! I’m thinking about working on evals to see if I can assess how well a model understands koans. I did a small experiment in this direction once, and the results were underwhelming but promising. Seems worth trying again, and in a more formal way.

Until then, Kaku-Ora is live! Ask it a question. See what comes back.

The code is on GitHub, and the model runs on HuggingFace Spaces.

I’m writing from my grandparents-in-law’s living room. We just finished watching the Ohio-Michigan game, now we’re watching the Minnesota-Wisconsin game, and my wife and I will soon be driving to the airport to fly home. It was a cozy Thanksgiving, even if about half the family has come down with a cold, and I got to see my first snowfall in several years.

I initially made good book progress this month, finishing up revisions on Chapter 6. Then I got sick and did nothing for a week. The following week I was busy catching up at work, and then I got obsessed with a special project that I’ll be posting about soon, so needless to say I didn’t make as much progress on Chapter 7 as I would have liked.

And December is shaping up to be a busy month! There’s a sesshin. There’s holiday plans. And there’s the aforementioned special project (and another special project you won’t be hearing about for a little longer). So, in the hopes that I can make any progress on book revisions at all, I’m not going to hold myself to posting weekly until after the new year.

In the meantime, in addition to, as always, plugging my Twitter, you might read some interesting posts from other folks, including:

• ’s series on the Enneagram (last post with links to them all)

  • (I still owe you all a post about Buddhism and the Enneagram inspired by Michael’s posts when I can finally get to it!)

• ’s series on consciousness (last post with links to them all)

• ’s excellent post detailing a personal story of a stage transition.

  • More of us should write more things like this. I think they create a record of incredibly useful evidence and supply examples that can help those struggling to break through from one stage to the next. I wrote a short comment on his post with my own Kegan 3→4 story, but I should put my money where my mouth is and write up all my phase transition stories.

• LessWrong ran an event called Inkhaven during November. If you don’t usually read LessWrong, there’s an unusually high volume of new, interesting posts from the past month that you might enjoy, and even more at the Inkhaven link.

Subscribe now

My brain is not naturally inclined towards language. I didn’t start speaking until I was 3. I’m more comfortable writing code than prose. And when I write for publication, I spend long hours editing to make my thoughts comprehensible.

Or at least, that’s how it was until about a year ago. Then, I started collaborating with Claude on most of my writing projects. At first I was just as slow as ever, but I eventually learned some useful patterns. Now, working together, we write about three times faster and hit a higher standard of quality than I did alone.

This post documents those patterns so that you might copy them to improve your writing.

Co-Writer, not Ghost Writer

Like I said in my post about writing code with Claude, I think of it as a collaborator and partner, not like a magic box that does things for me. When I sit down to write, it’s like I’m working with the most patient editor who’s ever existed. Claude is always willing to keep iterating on the same text until we get it right and is always available to work with me at any time of day. Some things Claude is especially good at include:

• Fact checking

• Generating variations on existing sentences

• Analyzing structural issues and suggesting alternatives

• Brainstorming

• Researching and synthesizing information

• Writing first drafts that roughly match my style

The workflow looks like this: I bring ideas, goals, and sometimes pre-existing writing. Then, I iterate with Claude in tight loops to organize, refine, and research. We keep going until the text meets or exceeds my standards, and when it seems like we’re done, I do a final readthrough. If everything still sounds good, then we’re ready for publication.

That’s the workflow at a high level, but as always, the devil is in the details, so let’s look at exactly how Claude and I write together, including some example prompts used to produce my recent essay about The Sort.

Subscribe now

The Workflow

1. Start with Brainstorming

Instead of starting with a first draft, like I would have done before working with Claude, I start with brainstorming. I’m not very good at brainstorming, so I get Claude to do the babble, and I do the prune.

Example:

Prompt: i’m writing an essay for my blog about fallout of the sort, which i’m calling living in the shadow of the sort. i want to spend a bit of time exploring some of the consequences of the sort, why it exists, and how we might get out of it (and if that’s even a good thing). don’t try to write the post for me right now, but give an outline of topics you think would make sense to include. right now we’re brainstorming, so err on coming up with too many ideas to include, and then we’ll use that to narrow down to figure out what makes sense as an outline later. the the idea of the sort comes from a patrick mckenzie tweet. here’s the whole thing so you have context: [copy/paste of the twitter thread]

Claude generated a couple dozen potential angles and over a hundred individual points. We then went back and forth to narrow these down and find the right essay structure. This took maybe 30 minutes, but it meant when we started actually drafting, we had a fleshed-out plan.

2. Draft in My Voice

Since I’ve written so much, I have a defined writing style. Whether or not my style is good is up for debate, but it definitely exists. When I draft a post, I instruct Claude to read several things I’ve written. It also helps that my posts are in the training data and it already knows who I am.

Example:

Prompt: okay, great, i think this is a good start. now, you have some samples of my writing and you can find more online. try to convert this into an essay. i think we should target something in the 3k word range, though we can go longer if that seems necessary to explore some particular idea

The output was generally in the direction of my style, but regressed in the direction of Claude’s voice. I needed to do extensive editing to produce something I was happy to publish, but for a first pass, the draft was a huge timesaver.

Note: At this point I’ll put the draft in a Google doc and share it with Claude. This makes it easier when I need it to read the whole thing after I’ve made edits.

3. Structural Review

Next, I read the draft and work with Claude on sections that feel off. I’ll do some of the rewriting myself, but then come back when I notice that a section isn’t doing quite what it should. As is probably obvious, it’s a big time saver to get the structure of a piece right before trying to do stylistic edits on paragraphs that might never appear in the final version.

Example:

Prompt: hey the japan section feels underdeveloped. you set it up as an example of resisting the sort, but then say it’s a combination of factors and move on. i actually think the japan example is an important hinge in the essay. let’s figure out how to make it that.

This prompt saved me from polishing a section that was fundamentally failing to do what it needed to. By noticing early and working to fix it, we saved an hour or more of writing time.

4. Research

I write non-fiction, and that means doing research. I do plenty of this myself, but I need Claude to know about it, too, so I often do my research with Claude, asking it to do web searches, give me links, and synthesize what it finds.

Research can happen at any time. I’ll frequently find in the middle of a piece that I need to know more about something. If I think it’ll be quick, I’ll do the research in the chat I already have going. If it seems likely to be complex, I’ll start a new one, and when done, I’ll ask Claude to give me a summary to paste back into the original chat to hand over context.

Note that asking Claude to do research is importantly different from relying on what it thinks it knows! It’s often critical to include the words “do a web search” in a prompt to avoid hallucinations.

Example:

Prompt: do we have the details about japan right? do a web search to make sure the details about the lost decades are correct and that we aren’t missing anything important that would make this section wrong

This came back with some helpful details that confirmed we mostly had it right in the text but had a few minor details wrong that needed fixing.

5. Iterative Refinements

This is where I see big productivity gains. When I write solo, I’ll spend an hour struggling to fix a paragraph that feels wrong. With Claude, I flag what’s wrong and get several alternatives in seconds. None of these are usually the final version, but they help me make steady progress towards a version of the paragraph I like.

Example:

Me: somethings not working with this paragraph. i think it might sound off because the example is too on the nose. how might we fix it [paste paragraph]

Claude: [provides 4 initial options]

Me: okay, i ran with one of those options. here’s the new version but not quite there yet. suggestions? [paste paragraph]

Claude: [provides more variations]

Me: great, this is getting closer. i like the first two. here’s the new version i wrote, wdyt? [paste paragraph]

…

Eventually we got to something I was happy with. The trick is to never struggle alone. If I feel stuck and don’t know what to do, I talk to Claude about it to generate new ideas that get me unblocked.

6. Fact Checking

I regularly ask Claude to review drafts specifically for errors. This includes both factual errors, like when I get a claim wrong, and consistency issues, like when two parts of the text contradict each other. I ask it to fact check both before and after edits, since sometimes what seems like a simple stylistic change introduces language that makes the literal interpretation of words wrong.

Example:

Prompt: fact check this [paste paragraph]

I then iterate on the text until it passes the fact check. Sometimes this becomes quite involved if I’ve misunderstood something.

For example, while working on book revisions, I once had to spend several days with Claude to better understand the Löbian obstacle. In the end, I’m still not sure I understood all the nuances, but thanks to Claude, I’m pretty sure I didn’t say anything obviously wrong in the text, which is much better than I would have been able to do on my own. Perhaps a patient friend with a deep understanding of Löb would have been better, but I doubt they would have wanted to wake up and write with me at 6 am!

Taste and Thinking

This workflow works well for me. But, similar to what I said in the programming essay, it only works well because I bring good judgement and taste to the writing process. I wouldn’t be able to write the way I do with Claude if I didn’t know what good looks like.

I also need to be able to think clearly. Claude can easily become a mirror for my confusion, and not in a way that helps me become less confused. I regularly have to push back against Claude being too agreeable. It’s not a sycophant, but I can easily bias it in my direction, and sometimes I have to specifically ask it to be critical or to play devil’s advocate.

The good news is that writing is a path to clarifying thinking. As Paul Graham said, “If you’re thinking without writing, you only think you’re thinking.” The catch is, you can’t outsource your writing to Claude and expect to get better at thinking. So while Claude is a powerful tool for clearing writer’s block and providing detailed feedback, you must ultimately do some writing yourself to think and write well.

I grew up on the edge of The Sort. I probably would have been in the center of it, but I got caught in a gifted-kid trap, never learned how to work hard, and hit a wall in grad school. After spending a decade floundering before learning how to actually dedicate myself to difficult work, I got myself back into The Sort.

Now working alongside people who were more aggressively Sorted, I’ve been able to observe the culture like an outsider while still being part of it. And while The Sort offers you a lot in terms of money and respect, it also creates insularity, elitism, and steals away optionality.

I’ve chosen to stay on the edge of The Sort, getting some benefits without losing all my personal time and ability to pursue what matters to me. Everyone Sorted faces this choice, and it’s up to them to decide how to make it.

Transcript

This transcript was editing into essay form by Claude.

Last week, I wrote a post about The Sort—the idea that there’s a lot of systems in the world that incentivize people and enable them to maximize their productivity. This provides a lot of benefits to the world, but it also comes with some costs, especially costs for those who get Sorted.

There are a lot of good things that come out of it. You have access to money. You can go on nice vacations. You can buy cool things. But it also comes with a real psychological toll, or spiritual toll, on a person. So I wanted to talk a little bit about that today. I decided to do it as a video because I also wanted to talk a bit about my own experience being Sorted, and I thought this is maybe a nicer format for that.

For myself, I grew up on what I would say is the edge of The Sort. I didn’t go to the highest prestige university or work at the most prestigious companies. But thanks to a decent amount of smarts and interest, I’ve ended up working in software for a lot of years. Obviously, this puts me in quite close proximity to a lot of folks who have been more aggressively Sorted than I was.

Growing up, I had a lot of expectations put on me. Early on, they did some IQ testing, put me in gifted programs, and really ended up developing this idea of myself that I should always be able to do the best. I should be the smartest person in the room. And for a long time, that got reinforced because it was just true.

I think it wasn’t really until maybe high school that I started to meet other people that felt like they seriously rivaled me in some way. And it wasn’t really until I was in grad school that I even encountered any situations that I couldn’t just sort of float my way through. High school and undergrad, I never studied for anything. I didn’t develop any study skills at all. I would do the work and just be correct. Without really studying, I’d show up, pass the test, get good grades. It was all just fine.

But then eventually I sort of hit this point where I could no longer meet my own expectations anymore. These expectations that had been put on me and that I had been meeting and then had internalized—it just couldn’t happen automatically anymore. At least to continue to progress to the next level, to make the next thing happen, I could no longer be the best without putting in some real work. So I floundered for a bit, for about a decade. And it took me a while to figure out how to work—this is kind of a strange way to put it, but how to actually dedicate myself to something to make it happen, how to learn things when it doesn’t just come automatic.

Eventually I did sort this out. Maybe by the time of 2017 or so, I feel like I was now at least to a point where I can actually do things. I am not the washed out gifted kid anymore who is under-fulfilling his potential—which I should be clear was not actually that bad, right? I was working as a software engineer making plenty decent money. So don’t get me wrong, this was not failure in some absolute sense. It was only failure in this relative sense to which I had these expectations that I could do more.

And just to talk about that for a second, I think this is a problem that happens for a lot of people who get wrapped up in The Sort. A lot of people experience imposter syndrome. I know I certainly dealt with a little bit of that. What I dealt with more strongly though was just having these expectations for myself, expecting to be able to meet them and not being able to do it and finding this quite frustrating and not understanding why I couldn’t meet them. In the end, the answer had nothing to do with how smart I was. It had everything to do with my ability to actually sit down and do the work that is necessary to accomplish the things I’m trying to accomplish.

As I said, I eventually sorted this out and now found myself regularly in situations encountering more of the folks who have been Sorted, who have been caught up in The Sort and pulled along quickly towards the top to work at the most prestigious firms, to work at the most prestigious universities, and to feel like they have permission to work on the hardest problems. Although I should be clear that honestly, I think everyone has permission to work on these problems.

There’s something interesting, I think, that happens about the culture that develops around people who have been very aggressively Sorted in this way. And it has a few key features that I was able to see as a little bit of an anthropologist, looking from the outside in, because this had not been my life the entire time. I had washed out, you might say, for 10 years. Now I am back in. And I’m seeing this stuff a little bit as an outsider, not entirely. I’m good enough to be there perhaps, but have not been living this in the same way as some of my peers were.

There’s a certain level of insularity, I think, that develops. A kind of attitude of, well, look, you didn’t go to the best schools. You’re not working at the best firms. So how could you possibly be good enough? And I think there’s also this expectation that if you are good enough, then you will just appear and rise to the top. And this will happen very quickly. And as long as you have not yet been admitted into this inner circle, then clearly you must not be good enough in some way to get into it. It’s sort of this self-fulfilling thing.

And this is normal human stuff, right? Humans everywhere form cliques. They think that their own in-group is the best, and the out-group is obviously worse, and as you get further away from the in-group, people are just less good, less capable, less whatever it is. So it’s all just very normal human stuff.

The other thing that I think powers it is that it took me a little bit to realize this is what was going on, but a lot of people who have been successfully Sorted have never really experienced hard failures. They’ve spent their entire lives just being successful at the thing they’re doing. And so it’s not really because they’re not trying hard things necessarily. It’s just that they happen to be naturally gifted at those things, or they’re working within areas that they’ve been really well-prepared for. The consequence of this is that when you’ve only ever experienced success, you start to actually believe that you really are just special and able to accomplish things. And also if you haven’t experienced failure in this way, you’re much more likely to, when you see someone else has failed at something or is not performing well at something, to assume this is a deep fact about them rather than a fact about the situation or the type of problem that they’re facing, or a fact about how maybe you have succeeded, which is not because you were so special, but you also just had a lot of good luck and good preparation.

I think this really powers The Sort because the result is that it sort of creates an inner group that just sort of persistently reinforces itself that yes, we are special and we deserve to be here. And everyone else deserves not to be here because they have not demonstrated it.

Now, I want to be a little fair here to myself. I definitely feel some of these same impulses. Even though I didn’t get as aggressively Sorted as some of my peers, I have been Sorted more than many people. There are definitely times when I meet people and I’m like, oh, you clearly are not a real player in some sense, because you’re a person that read some things on the internet, you didn’t go to a good school, you couldn’t have gone to a good school.

And I try my best not to actually be like this, but sometimes it’s quite hard when you’re in a conversation with someone and they’re spouting what feels like a bunch of physics crank nonsense at you to just be like, okay, what’s really going on with this person? And why are they here? Why am I interacting with them?

So I admit that I think I have my own bit of elitism too. And I think that’s fundamentally what we can maybe call this. The Sort invites a kind of elitism because you spend your life among other people who are also in this category of elites. And it creates this bubble where you really expect that everyone else should be able to rise to this, and if they can’t, then they sort of become like NPCs, to borrow that metaphor. They’re not full agentic humans. They are people that just carry out roles and don’t have real stories of their own, which is of course not true. And it’s obviously not true. But it’s very easy to fall into these kinds of traps, especially if you surround yourself only with other people that are like yourself.

I also wanted to talk a little bit about optionality here. One of the other things that The Sort does is it tries very hard to take away your optionality. And this has a lot of impacts on how you end up living your life and maybe actually ties back into this NPC point.

The reality is that when you get very aggressively Sorted, you end up having high opportunity costs. I know I certainly have this. There’s a lot of things I would like to do in my life. For example, it actually sounds kind of great to me to maybe work or go live in a monastery for a while. Could I do that? Sure. Am I likely to do it? No, because I know how much money this would cost me. I don’t mean how much money I would have to pay the monastery. I mean, what is the opportunity cost of doing this? How many tens or hundreds of thousands of dollars of unearned income is the cost of jumping off the treadmill to go do this?

And I think that lots of other people see the same thing. Maybe I’m working—I’ve written a book, I’m working on revising it. It would be great to get this thing done faster, to spend more time on it, to spend more time as a public intellectual. That sounds fun and appealing if anyone’s interested in what I have to say. But it’s very hard to make it go when I need to spend a lot of time doing other things, spending time at work.

And this just becomes more aggressive the more you let yourself be Sorted. The work in particular really can take over someone’s life to the point that there are many types of activities and hobbies that I know people who are more aggressively Sorted than I am don’t let themselves do because they’re worried that maybe this will come back to reflect poorly on them, or alternatively, that they need to be engaged in the right hobbies to make the right connections with the right people in order to advance their career.

So The Sort really does a lot to constrain someone’s life, not in the same way that a lack of resources does. If you don’t have a lot of money but you have a lot of time, there’s a lot of things you can do with your life. But there’s also things that you can’t because you don’t have the money to afford them. The Sort works to create the opposite situation where you have plenty of money, but it does everything it can to eat all of your time and give you very little control over it.

So you have the money to go on a great vacation. But the only time you can go on that great long vacation is actually when you’re between jobs, because when you’re on the job, the idea of taking a month off is basically impossible. Even two weeks is kind of a lot, especially at some jobs.

What my point is here is that The Sort creates this situation where it offers real benefits to the people that are chosen by it and are sorted up. But it does come with some real significant cost, both to the individual and I think to society at large, which I talked a little bit more about in the essay.

I’m happy enough to be here on the edge of The Sort. And I frequently have to choose to resist becoming more Sorted because there are things I want to do in the world that if I let The Sort pull me up more aggressively, I can tell I will lose the ability to do. I won’t have the time or it will be in conflict where maybe, say, I take a job where the things that I blog about would become a liability and I can’t do that anymore.

So I’ve sort of made my choice here to try to strike a healthy balance of what I think is engaging with The Sort enough to get some benefits from it, but not so much that I lose all of my personal time, all of my ability to pursue the things that I’m interested in and I want to do in this life.

And everyone who’s Sorted has to make the same sort of decision for themselves. Are they going to try to strike a balance? Are they going to go all in? Maybe they’ll strike a balance—and a lot of people do this—by going very hard and then taking time off to do something else. And they go back and forth rather than trying to have it be persistently balanced over time. It’s just different choices that different folks choose to make. We all have to make our own.

Today we live in the shadow of The Sort. I first learned about The Sort from a Patrick McKenzie tweet. The Sort is the emergent system by which talented individuals are identified and relocated to maximize their productivity. It finds people, wherever they are, who have the skills and smarts to produce more of what markets and institutions demand, and it offers them money, prestige, and power in exchange for being Sorted. It’s highly efficient, and it’s operating at scale across the entire globe.

The Sort isn’t a conspiracy, though. Its mechanisms are transparent. It starts in school, where high performing students are Sorted into honors classes. Then, standardized tests, extracurriculars, and various status markers Sort everyone again to determine university admissions. Finally, elite university attendance opens the doors to prestigious careers in academics, politics, and business, and, once a person is hired, the Sort does its best to select among the Sorted for promotion.

But The Sort wasn’t always with us. It emerged gradually after the end of World War II. At first it grew slowly as access to higher education expanded and professional management took the reins of most economic activity. It grew faster in the 1980s when business regulations relaxed and tax rates reduced. And it exploded in the early 2000s when the internet tore down the barriers that kept the would-be Sorted out of The Sort. Today, with most of the old frictions removed, The Sort has reshaped the world and our communities.

Talented individuals are now concentrated in a few sectors like law, tech, medicine, and finance. Other sectors languish, struggling to attract even a small share of Sorted folks. The smartest person in a small town no longer stays to make it better. They can make more money and enjoy greater luxuries by moving to a bigger city. And in those bigger cities, The Sorted form their own communities, isolating themselves from the locals as if they were expats in a foreign country. They may live in New York or London or Shanghai, but they are only in those places, not of them.

If all this makes you afraid of The Sort, you needn’t worry. The Sort is already here, and you have already been sorted. In fact, if you’re reading this, there’s a good chance you got sorted near the top yourself, or at least onto the fringes of the Sorted. The only thing we really have left to worry about is how to live well under the watchful eye of The Sort.

Living with The Sort

To be Sorted is to live in fear of being Resorted.

From the outside, being Sorted sounds fantastic. You’re offered money, prestige, and a path to power. Who wouldn’t want that? But the Sorted know they are under Damocles’ sword. They deal with status anxiety, imposter syndrome, and the guilt of entitlement. They’re hyperaware that they could easily lose it all. The Sorted have internalized The Sort, and they’re constantly judging themselves and others by its measures.

They see the cool vacations their peers took on Snap and Instagram. They fret that their own holidays were squandered and that they took too many days off work. They stress over getting their kids into the best schools, not only for their children’s sake, but also because children are a reflection of their parents, and maybe the parents should be Resorted if the children aren’t performing well. And on the job there’s constant competition for limited spots at the top. There’s always someone younger, hungrier, and smarter waiting to advance their career given the slightest opportunity.

In these ways and more, The Sort keeps the Sorted trapped. They understand what the consequences will be if they stop playing the game. They know that if they “follow their passion” they’ll be giving up not only money, but all the markers of status The Sort has given them. It’s not uncommon to hear stories of people who burn out, go live on a farm or in a monastery or run an ice cream stand, and then come back a few years later when they can no longer resist the urge to be among the Sorted. Once The Sort captures your mind, it becomes almost impossible to do anything other than let it direct your actions.

The other “option” is to live the life Unsorted. But it’s not so much a choice as one that’s forced on you.

When The Sort doesn’t pick you, it leaves you behind. To the Sorted, you’re not a person. You’re a product, a consumer, or a source of cheap, temporary labor. At best you are the downtrodden worthy of pity. A fellow human, yes, but not a human capable of great things.

Unsurprisingly, the Unsorted don’t like this. They believe that the world should be fair, and they’re mad as hell that the Sorted think they’re better than them. They want to see the world returned to a time before The Sort, now almost forgotten. So they organize, both on the left and the right politically, to protest and try to fight back. Alas, when they do, the Sorted exploit their anger, orchestrate becoming the leaders of their movements, and use Unsorted unrest as a path to power.

But despite what I said earlier, The Sort is not inevitable. There are pockets of the world The Sort has largely left behind. There’s costs to living outside The Sort, but there are also benefits. And perhaps no place has better resisted The Sort for longer than Japan.

Reforming The Sort

Japan didn’t set out to avoid The Sort. In fact, up until the early 1990s, Japan was on track to become one of the most Sorted countries in the world. And then, the market crashed. They spent three Lost Decades with stagnant wages. It kept Japan the place we know it as today: a beautiful, high trust society where even the lowliest convenience store runs well. Now, changing demographics threaten to bring The Sort back to Japan, and Japan is going to have to change if it wants to thrive.

How did Japan stay out of The Sort? It’s a combination of factors. Lifetime employment norms and compressed wages meant talented people couldn’t be efficiently reallocated. Culture and language barriers made it easier for Japan to isolate itself from full integration into the world system. And economic stagnation took away the free energy The Sort needed to power itself. No one feature was enough to stop The Sort, but combined it made Japan into a bubble of pre-Sort life.

But staying out of The Sort came at a cost. The Sort, for all its downsides, brings real value to the world in the form of dramatically better goods and services available at lower prices than were ever previously possible. During the Lost Decades, productivity lagged, innovation stagnated, and young people had limited access to opportunities. It’s only now, with an aging population forcing the barriers down, that The Sort is beginning to seep back in.

The ideal would be to find a way to get the benefits of both The Sort and the Slack that develops when The Sort isn’t running. Unfortunately, I don’t know how to do that, and as far as I know, nobody does. The Sort knows no bounds on its hunger to optimize. Slack only exists in the negative space left behind by the places optimization missed, and it’s hard to design good policies that create just the right amount of Slack.

But maybe there’s something we can do. Maybe it’s a reformed tax code or stronger barriers to immigration or higher tariffs or universal basic income or something else. My hope is that The Sort will be the cause of its own moderation, because there are smart people working in policy who have been Sorted there, they also feel the pain of The Sort, and they, perhaps more than anyone else, have the power to nudge us towards a world that is less aggressively Sorted.

Share

This is the once-a-month status update, which is mostly about progress I’ve made prepping the book I wrote for publication. I post essays about various topics most every other Wednesday that I hope you’ll enjoy.

I’m happy to say that I wrapped revisions on Chapter 5 (basically a long and much improved version of my most popular post of all time from back when I was blogging on Medium), and now I’m on to Chapter 6. Progress on Chapter 6 revisions is going smoothly, which I’m glad of since this is the fulcrum chapter of the book. I’m aided by the fact that I’ve written the content of this chapter several times over already, so it’s mostly about doing a fresh editing pass with Claude’s help.

Speaking of Claude, earlier this month I wrote a piece about how I became a more productive programmer with Claude. I’m thinking of writing another, similar post, but this time about how I use Claude to write essays and revise book chapters more efficiently while producing higher quality writing in the same or less time. I know a lot of people struggle with writing and try to use various LLMs to write for them, often with limited success. I hope I’ll be able to offer some advice to help them produce less slop.

And speaking of slop, I’ve been thinking lately about the enshittification of Substack. I’ve noticed that there’s a lot more nudges to subscribe and full page overlays encouraging me to pledge support. At first Substack was great because it was still in the attract users phase, but it’s slowly creeping towards maximizing profits, and doing it at the expense of the user experience. It’s not quite as bad as Medium yet and nowhere near as bad as Facebook, but if this keeps going I guess I’ll be looking for another place to blog. Thankfully, I’ve got control of the uncertainupdates.com domain, and I can always jump ship if it gets bad enough. Hopefully, though, Substack reverses course and stays great for a while longer yet.

But you know what hasn’t been enshittified? My Twitter account! So point yourself to X the everything app, follow me there, and you can enjoy a steady stream of mildly interesting posts, like this one:

Subscribe now

In the talk I try to cover roughly how the postrationalist scene got started and how it later merged with other scenes thinking along similar lines but with different intellectual starting points. If it contains any errors (because you were there and remember it differently than me!), please let me know in the comments.

A Short History of Postrationality

Opening Presentation

Thanks for having me. This is exciting to get to talk about this stuff. I occasionally grab people at conferences and they’ll ask me questions about this or we’ll just get on this topic. It’s kind of fun because there’s a lot of interesting history here in the community and it’s not very well documented. Having opportunities to capture this stuff is helpful, especially now that we’re further on and there’s actually some history to tell, whereas for a long time there wasn’t because it was all just sort of new and fresh.

Jack said I want to talk about the history of the postrationalist movement as I think of it. Maybe before I jump into that, it’s worth saying a little bit about what I mean by this postrationalist movement, just so we’re defining terms a little bit here.

It’s obviously defined relative to the rationalist movement, by which I specifically really mean the movement that centered around the community that grew out of the Sequences that Eliezer Yudkowsky wrote, and then grew into this community around LessWrong and CFAR and other related organizations. It also developed in person in Berkeley, where we had an in-person rationalist community.

You’ll also notice that I’m always very careful to use the word “rationalist” as opposed to “rationalism.” This is mostly to help make a clear distinction between this movement and the only tangentially related philosophical position that happens to inconveniently share roughly the same name.

The postrationalist community really grows out of this. We have a rationalist community and then there’s a set of us that kind of want to, in integral theory talk, transcend and extend this thing that rationality is.

The Berkeley Origins (2014-2015)

Getting to the history of it, it really starts in my mind in Berkeley, around 2014-2015. I’m a little fuzzy on exactly when this was, but I remember it really all coalesced at this house party. The social scene in Berkeley really evolved around these house parties. One of the group houses where we all live decides to throw a party for whatever reason. There could be no reason, there could be a theme or not. Then 50, 100 people show up, cram into a single house, and we talk about stuff. There’s dancing, there’s drinking, there’s philosophizing. It’s generally a fun time.

At one of these parties there was this extended conversation that started between myself, Malcolm Ocean, and Ethan Ashkii. If you don’t know who the two of them are, that’s fine. Malcolm is a little more public online, so you may be more familiar with him. He and Ethan had been getting into Robert Kegan’s work and they had been reading it, so we were having a conversation about it. It went well into the night. I think that particular house party, Ethan and I stayed up till sunrise talking about this stuff after Malcolm disappeared at some point.

From there we formed something like a philosophical circle. We had nominally a book club—that was the official structure of it—but it was mostly just an excuse to get together every two to three weeks and talk about whatever we had been reading in this space of how do we be rationalist but actually win.

I think this is the seed of it. We had looked around and were frustrated with the fact that we had all these friends that were going through CFAR, which offers these rationality training seminars to try to teach people how to be better at applying the lessons of rationality. We had lots of friends that had done these things, but all of our friends still kind of sucked. Not that they were bad people, but they were not actually succeeding at making their lives better. They were not actually achieving their goals in the ways that we would hope that they would be able to.

So there was something around: what’s missing? What is the problem? I think we didn’t quite know how to have the whole answer at this point, but I think the direction we were going in was what’s missing is the embodied piece, the integration of this into life and how to actually do this. It’s not enough to be able to sit down carefully, reason this stuff out, and then have a theory. You have to actually live it. The actual process of living it is quite a bit more complex and nuanced than sitting down and doing some calculations on paper or writing a blog post about how you’re going to solve the world’s problems. Actually solving the world’s problems is quite a bit harder than writing about it.

Kind of obvious when I say it that way, but the rationalist community has this real bias towards the idea that hey, maybe if we apply intellect and smarts hard enough and better than people have done in the past we can achieve better. It’s not a terrible intuition. This does work sometimes. There are lots of problems that are lowhanging fruit that can basically be solved by just thinking about it harder because nobody had bothered to think about it very hard before. But I think they make the mistake of overgeneralizing from those examples and thinking that this is a general strategy that will work all the time, not a strategy that works in isolated cases where people have not tried hard enough.

The Reading Group and Foundational Texts

Back onto the story here. We did have this postrat circle, made up of myself, Malcolm, Ethan, and several other people that sort of came in and out. Two other people that I think are important are Mike Plotz and Michael Valentine Smith. You maybe know him, maybe you don’t. He was one of the instructors at CFAR, sort of part of the early in-person rationalist community as it was getting formed around 2011 or so. He would later link up with us more directly, but at that time he was still more heavily invested in the CFAR world and trying to make the rationalist project itself into something.

At this reading group that we had, there were a few books that were pretty important that we read. These are sort of the foundational texts of how we went from being steeped in this rationalist community to figuring out what was missing.

The important books were:

The Inner Game of Tennis - I think that’s the first book we read. If you’ve not read it, on the surface it is a sports book about how to get better at tennis, written in the ‘70s when tennis had skyrocketed in popularity. But it’s actually secretly a book about Zen. By the end of the book, the secret is mostly out. I can’t remember if he’s very explicit about bringing in all this Buddhist terminology stuff, but that’s essentially what the book is by the end of it.

Impro by Keith Johnstone - A book about improvisational theater, but it contains some really important concepts that turned out to be essential for what I think of as the core of postrationalist thought. Maybe one of the most important chapters in this book are about mask work and trance states. Mask work is essentially when you do improvisational theater but it’s based in shamanic rituals that people have been doing for thousands or tens of thousands of years, where people literally put on a mask and then become the character, inhabit the character that the mask represents. This points to some interesting aspects of human psychology that were really important for us to explore.

The other part that’s connected with mask work is trance states. There’s a lot of interesting stuff here around what we can think of as trance states. I think we can take a fairly expansive view of trance states—they can be these mystical things but they can also look like being hyperfocused on the work you’re doing in front of you. They don’t have to be this strange mystical thing, they could just be being hyperfocused on writing something or writing some code or drawing a picture. Even driving, you can get into trance state while driving. It offered a window into some interesting aspects of human psychology that turned out to be really important for us to get into.

Robert Kegan’s work - We read The Evolving Self and In Over Our Heads. I think maybe he had a couple of books out at that time that he had written with Lisa Lahey, and then I think a couple more books have come out since then that they’ve produced. We also spent quite a bit of time with the manual for doing what’s called the subject-object interview, which is a way of essentially assessing someone’s cognitive developmental level in Kegan’s system. We spent a bunch of time with that manual, interviewing each other and trying to figure this stuff out. That was kind of fun. Really getting inside this theory before, I think, the better explainers that exist online now. Those things didn’t exist—we had to read the source material to produce the explainers.

David Chapman’s Meaningness book - I’m sure many of you have engaged with that. It was in a much rougher state 15 years ago or whatever the timeline is here. David does a great job of pulling together a lot of disparate threads, so that was really useful for us to read and provided a lot of stuff for us to talk about and think about.

I also want to throw in here that we were also reading Ribbonfarm. If you’re not familiar with the Ribbonfarm blog, this was Venkatesh Rao on Twitter and Sarah Perry. There were a couple other people on Ribbonfarm, but they were the two key folks here. Their writing was definitely quite interesting to us. It turned out to be important because it actually linked us up with a larger movement of folks that were not in the space of the rationalist movement but were very interested in similar ideas and had come to them from a different direction—maybe not quite the reified rationality of the rationalist movement but very much from a modernist rationalist perspective and then looking to figure out what’s beyond that.

I will say that at some point further down the line someone read Ken Wilber and pointed us in that direction. I think Matt Goldstein, I don’t know if any of you are familiar with him, maybe introduced us to Ken Wilber and integral theory. That was kind of interesting. But by the time we read Wilber, this is maybe 2017, 2018, I think we had kind of already worked out all these ideas for ourselves. We were like, “Yeah, sure, this guy’s basically correct. He’s offering an alternative explanation of this stuff to the way that we are,” but definitely thinking along the same lines.

I know I have said a lot of stuff, so this is maybe a good stopping point to open it up for questions if you have any before I continue the story.

Q: How do you think about the relationship between postrationality and David Chapman’s metarationality?

I think these are basically the same thing. We have different words for them basically for historical reasons. The reason I’ve talked about postrationality is that we weren’t actually originally thinking of ourselves as postrationalists. I think I’m the person that started using this terminology within the group and got us calling ourselves this.

It was largely because I had been thinking a lot about the modernist-postmodernist transition. I was reading a lot more philosophy than everyone else was and I was really looking at what happened in this process in the early 20th century where we go from what are the seeds of modernism in philosophical thought that developed over the 17th and 18th and through the 19th century, and then how we got to postmodernism, which really developed out of the seeds around 1900 or so, late 1880s, 1890s, in reaction to industrialization. It pushed philosophers to really grapple with the loss of humanity within the modernist project. Lots of people didn’t really catch up to this until later. I think as a whole, society didn’t fully grapple with this until maybe the post-World War II era. But philosophers were dealing with it.

Also there’s this strain where in the 19th century German philosophers were bringing in a lot of Buddhist thought into the European philosophical tradition. This also proved to be a really important mixing in order to create what would become the postmodernist philosophical project.

So I was getting very into this. I was reading Husserl, I was reading Heidegger and then Sartre, and seeing like, okay, there’s something that they got right that I felt like they had some understanding of things. But then also looking at how the postmodernist project fell apart and crumbled in on itself with a bunch of what I would now call grifters—people that weren’t really part of this.

I think this is maybe an interesting thing to bring up and gets me back onto the story: there’s this pattern where once you have some sort of way of understanding the world and then you transcend it, this can also look like the thing before you had that.

Maybe to be more concrete about this: the modernist project offered some improvements over what we were doing in the premodernist world. And then the postmodernist project tries to take all the good stuff out of modernism and then do something more, make it better. But it importantly is built on top of the scaffolding of modernism and can’t exist without this modernist project underneath it.

But if you are coming to this fresh and you are some sort of premodernist, then postmodernism suddenly gives you cover to do your stuff and maybe pass yourself off as a postmodernist even though in fact you don’t understand modernism well enough to be a postmodernist. And then this whole thing just sort of collapsed.

I think we see the same problem with rationality. We have this postrationalist or metarationalist movement, and it is really built on top of this foundation of rationality and says hey, these are a lot of really useful tools but they’re not enough on their own. We need to integrate them into this larger system—or in fact maybe meta-system, because it’s not really a single system, it’s much more nebulous than that.

But it unfortunately gives a lot of cover for if someone wants to be pre-rationalist to sneak their way in, because to the outside person that doesn’t have enough knowledge of all of this, it’s hard to distinguish between the postrationalist saying “Oh yeah, you got to integrate your understanding into your body” and a pre-rationalist perspective of being anti-intellectual.

I think this is a consistently tricky problem. And actually this gets me back onto the story. I almost very quickly regretted ever coining the word postrationality. I think maybe other people feel perhaps the same way about metarationality because it gives this cover.

The Movement Fragments

As things continued to evolve, the in-person scene that we had just kind of fell apart. Over time people just moved on with their lives. We had kind of explored these ideas and they were no longer the pressing things for us because we had figured it out.

I was writing some things about it. Malcolm did. This is where Val links up with this—Michael Valentine Smith writes a number of things about it. We’re also now linked up with this broader movement that’s happening around Ribbonfarm. It’s now really on Twitter. We see the seeds of what eventually become Metatopia start to develop really out of this Ribbonfarm-adjacent Twitter community.

Then it grows from there, but we’re now in a place today where the community is fragmented, splintered. It’s not really one thing anymore. It’s sort of too big to be one thing, so it can’t keep coherence. But I think also there’s this—what I want to say—I was sort of hinting at, there’s this grifter problem. Maybe that’s ungenerous. If we go to David’s model, there’s the geeks, sociopaths, and mops model, which I don’t know if you’re familiar with, but I will assume that you are for the sake of argument here. Essentially the mops and sociopaths showed up and kind of diluted the community in some ways.

However, because you are all willing to jump on a call here, I’m just going to assume that you’re all in the geek category and you’re all the true believers of metarationality/postrationality, so don’t worry, you’re all safe.

I think that’s kind of where we are today. It’s not really one thing. I think where I see the postrationalist movement, if it’s even a separate movement really, going from here is trying to thread some tricky path of how do we build a community that is supportive of helping people move through multiple stages of development.

I think this is quite interesting to people that are interested in integral theory and other things. There’s this challenge I’ve already hinted at where in some sense people need to move through a modernist rationalist phase in their life before they can properly get to a postrationalist, metarationality, postmodernist type phase and way of understanding. Without going through that, you can’t actually get to that place, or so I would claim.

I think this is quite tricky. It’s very easy to point people at one thing and say “Okay, here’s the example, just be like this.” It’s much harder to be like, “Actually there are multiple phases to this. You are going to have to spend time being in the not maximally correct phase of this thing because it is actually a developmental stage through which you have to move in order to get to the final one”—or if we even want to call it the final one. I think that’s maybe not even the right terminology here.

I think that is quite tricky to figure out and it is persistently the challenge for this type of community to figure out how to navigate that and support the development of folks through stuff.

And of course traditionally I think the way that we’ve done this is through things that look like organized religion. This is why I’m quite big on—as Jack mentioned, I practice Zen—and I think one of the things that’s really nice about Zen and some other forms of Buddhism is that they really create a space where this kind of development can happen and it is managed in a way through community structures and through your relationship with a teacher that help guide you through this without it becoming... yeah, without the risk to think that “Oh, I’ve totally got it, I’ve totally already figured it out” before maybe at least within the tradition we’re willing to say okay yeah, you figured it out.

So that’s I think maybe that’s all I really had to say. I’ll leave it there and we can open up for questions.

Q&A Section

Q: So I guess my first question for you Gordon is something about making sure I can distinguish postrationality from rationality. So in your view, what’s the defining feature of postrationality relative to traditional rationality? The second part is, one objection that I’ve heard about postrationality is, if rationality can be described like Eliezer does as the art of winning, then anything that works is rationality by definition. So is postrationality just rationality under a different label? And the last part here is just for you personally, what has postrationality actually changed about how you make decisions and take action? What do you do differently now than when you were just a rationalist?

Yeah, okay. What was the first part of the question again?

The defining feature of postrationality versus traditional rationality.

I would say the defining thing is maybe just this word integration. There’s a lot of complexity hidden behind saying that, but that’s fundamentally what I see as the difference.

Maybe this gets into the second part of the question. I think you’re right, this is a very common objection. I have responded to this objection so many times from people. They’re like, “Why do we even need this? Rationality is supposed to be this all-encompassing thing that if it’s the way that works then it is rationality.” And I’m like, okay, yes, I think there is this version of it that you can believe.

But the thing that is maybe written into the Sequences or the thing that people believe is the true rationality, that’s just not the on-the-ground reality of what people are doing. There is this vibe that rationalists have, and I think it is actually based in a bunch of metaphysical assumptions that rationalists tend to make. I’m not saying that these assumptions are not useful in some way and that they don’t produce good, accurate world models. I think that is definitely true—they do work for that.

But they have these assumptions that when taken to their extremes lead them to believe that essentially they can pull off something like the traditional rationalism philosophical project. I often describe it as: your typical rationalist is secretly a logical positivist. If you don’t know what that means, logical positivism was this movement in philosophy mostly in the first half of the 20th century, but it does actually stretch back a bit further. It really coalesced in the early 20th century where people really went hard on trying to make this work.

The idea is more or less that we can positively construct an accurate explanation of the world from first principles. So you could provide a logical accounting in mathematical terms of the entirety of the universe. The problem is that this just doesn’t work. Now, people figured this out about the middle of the 20th century, in the 50s and 60s. Within the tradition of analytical philosophy, logical positivism is effectively dead. There are definitely people that still harbor sympathies towards it, but there are clear definitive arguments that prevent this sort of line of reasoning from working.

So from a technical perspective it does not work. That does not mean that lots of people don’t have this intuition that it should be true—that if we just think hard enough and do enough math and have an accurate enough world model, then actually the map will fully describe the territory. In fact, I don’t even need the territory because I have the map.

As I say this, hopefully this all sounds like, okay yeah, obviously this is not right. But the problem is that the assumptions that power this worldview are often deeply baked in for people and it can be hard for them to notice it. So sometimes my arguments have to veer into the world of psychologizing people because I think that this is what’s going on.

That’s kind of why I think we need something like postrationality or metarationality, because people have these assumptions that something like logical positivism is possible. We need really some way to get people through that. Now not everyone has this. Some people just figure this out for themselves and never get caught in these traps. But lots of people will explicitly say, “Yeah, of course logical positivism is not correct, that’s not what I’m trying to do,” and then I think they immediately turn around and act as if it were true.

That’s where I really see the need here. And then what was the third part of your question?

What has postrationality actually changed about how you make decisions or take action? What do you do differently now than you did a decade or two ago when you were just a rationalist?

Sure. Yeah, that’s an interesting question. I’d have to go back quite a bit further. And this gets a little bit complicated, because for myself, I kind of can’t answer this question. That’s maybe what I’ll say, but I’ll give you an answer anyway.

The reason I sort of can’t answer this question is because postrationality didn’t exist for me—I helped co-create it. But the things that would become postrationality, or the cornerstones of it... for me it was some combination of, it’s inextricably tied up with the process of what in Buddhist terms we would maybe call enlightenment or awakening and the path to that process.

And maybe we don’t want to totally go down that route here, but it basically freed me from, I would say it was helpful for freeing me from my ontology, from the map of the world that I had and the narratives, the stories I was telling myself about how the world was, so I could actually engage directly with the world as it is.

This is kind of tricky because early on there’s a point where you actually can’t see the story. You can’t see that all of your knowledge is being interpreted through this layer—at least for myself, I couldn’t see that all this knowledge was being interpreted through this layer. I could not see the layer. It’s like the fish swimming in the ocean—I couldn’t see the water.

So it was very helpful in the sense of helping me through that process of breaking out of that, of being trapped by my own beliefs. And really living in a place today where I don’t know, I kind of just feel free to act and think and do things. I’m not so caught up in the idea of them.

I feel like that’s underselling it, but I feel like that’s the only way I can do it without giving a whole other talk about enlightenment and all that kind of stuff. So we’ll leave it there unless people have more follow-up questions about it.

And it’s funny that you were mentioning the logical positivism and the rationalism stuff because I was just reading recently Popper’s takedown of empiricism, rationalism, logic. That resonated quite strongly with me, your answer there. So thank you for that.

Q: Can I ask two things? First of all, you said the main motivation was to be more effective at winning. Did you actually try to measure how that was working out? And the other thing was, the Gemini language model told me postrationality has made some criticisms of the effective altruism movement. Maybe you could say a bit about that.

Sure. What was the first part of the question again?

Did you try to measure your winning?

Yeah, not explicitly. I haven’t quantified it. I haven’t put together a spreadsheet of how much am I winning at my life. However, what I would say is that having gone down this path, at least for myself I can say this is true—my life is dramatically better than it was 15 years ago. Was it improved in ways thanks to rationalist stuff? Yes, of course. Was it improved even more by what I am lumping into the postrationalist project? 100%.

There are a lot of things about life that I think aren’t really solved by doing math about it. There were things that were, you know, I had a lot of what I would call psychological shadow—things like codependent behavior, low self-esteem issues. I was just kind of depressed and anxious, just a lot of this stuff. And there was no amount of sitting down and thinking about it that could fix it. In fact, thinking about it arguably made it worse.

The things that I actually needed to do were much more like just getting out there and having experiences that would help me work through these issues and get experiences that would show me that, hey actually, there’s a different way to live and do things. But I attribute a lot of that to what I am considering this postrationalist strain of thought because it really says, you know, in modern terms we would say “go touch grass.” That’s fundamentally also one of the core ideas of this postrationalist project, whereas I think what I often see rationalists doing, and what certainly I was doing before moving in this direction, was: why do I need to touch grass? I can just think about grass real hard, and that’s basically just as good as touching it, so I don’t actually have to touch it.

So no real actual measurement here of stuff, but if I go back 15 years ago, average day of my life was maybe, if I ranked it on a scale out of 10, I would say that most days of my life were a three or a four out of 10. And now my life is nine out of 10 every day.

Now, I’ll be totally honest that a bunch of this is because of Buddhist-related things. But I wouldn’t have done any of that stuff if it wasn’t bundled into this postrationalist project. So this is as much a pitch for postrationality as it is for Buddhism. Hopefully that answers your question some, but I know it’s not a full accounting really.

The other part of your question in terms of criticisms of EA—I think I know that I myself have a lot of criticisms of EA, and I think maybe other postrationalists do, mostly because many EAs are utilitarians. And there’s a lot of stuff about utilitarianism that is I think just wrong. I think that’s as simple as that. I don’t even have to engage on a philosophical level. I think that is just wrong in terms of actually producing the best outcomes, and it’s because utilitarianism has this sort of baked-in idea of utility function but it doesn’t really address where utility comes from. This is not a fully answered question within the framework.

I think consequentialist thinking is quite important. It’s important to think through what will be the effects of my actions, how does cause and effect work. But I think many of the weird edge case things where EA goes off in strange directions are ignoring the reality of what humans care about.

The total scope insensitivity, for example, of many EA folks is I think kind of a mistake. It’s not totally unreasonable, but it is not practically reasonable from the way that we live our lives today. Maybe one day we will have enough abundance that being less scope sensitive will be possible. I think we see a trend over time that greater abundance allows lower scope sensitivity, greater expansion of the moral circle. I think that’s great, I think we should do more of that. But I often think it’s a mistake to immediately jump to “Okay, I have to cause zero harm today.” This causes people a lot of psychological distress and difficulty and I think pushes people in weird directions. I wish people wouldn’t do quite so much of it.

So that’s kind of my critique there. Hopefully that’s helpful. I’m sure other people have critiqued other things that I’m just not familiar with.

Q: You mentioned that the process of enlightenment for you was seeing the layer through which you interpreted reality. Do you think enlightenment is a one-time process where you see through the one layer, or is it a continual process where you discover layer after layer? I guess in traditional Buddhist terms, Gordon, obviously this is sort of the whole question of sudden versus gradual enlightenment.

Yeah, happy to talk about this. I think that it’s both. I don’t think there’s actually an either-or here. I think there is a critical insight that happens. I will assume that most of you are up to speed on what I’m about to talk about, but if you’re not, apologies.

So there is a critical insight that comes at the end of the path to what I would call enlightenment. And this is the insight that there is no separation between self and other, or at least that’s how I would phrase it. Other people might phrase it other ways. This is this final critical insight. Some people might choose to phrase it slightly differently or would say that there is a slightly different but related insight that you have to have after this. But we can elide those details and say that there’s that.

Okay, so you have this insight. Great. Your mental processes are totally reformed by this because your fundamental understanding of the world is now changed forever. And in fact, this doesn’t mean that you’re somehow permanently enlightened forever. You just crossed this first barrier of “Hey, you finally cleared out all of the croft that was separating you from being able to engage with reality directly.” But this doesn’t mean that you suddenly start engaging with reality directly all the time. There’s still plenty of opportunity to get lost in your own thoughts, to forget what’s going on.

I’m in fact doing some of it right now. I’m very focused on talking to all of you, which means that there’s a bunch of stuff that I am not paying attention to that I am just pushing out of my awareness, because I only have so much capacity to engage. I can only be so focused on so much stuff at once.

And it is really this continual process. There are all these, you know, I talk about them as habits of mind—I think other people talk about them other ways. But there’s just all this stuff lying around. So in some ways I kind of try to avoid even using the word enlightenment because, although it’s this thing that people know about, there’s really just this continual process of liberating yourself from your own confusions that you’re constantly heaping on yourself just through the process of living life.

The only way not to continue to keep heaping them on is to not live a lay life. You have to go live in the monastery and just sit all the time. That’s an option, you can do that. But there’s very few people that can live a continually liberated life and continue to free themselves in every moment from this stuff.

The reality for most of us is, okay, enlightenment is you cross this barrier, now you don’t have this fundamental confusion anymore, but you still got to wake up every day and commit yourself to the practice of staying awake.

Just a sort of brief follow-on question there just to make sure I’m understanding what you’re pointing out with the terminology that you’re using Gordon. Are you talking about sort of stream entry or like what Michael Valentine’s written about as kensho, like that initial glimpse? Are you talking about more of the final, you know, dying the great death maybe in your tradition?

Dying the great death. Yeah, I’m talking about dying the great death. Right, where the fundamental separate sense of separation never returns. Yeah. Okay.

Right, the process from stream entry until you die the great death, that is the path as we sometimes talk about it, and it is very important. But I’m in fact talking about what happens after that great death.

Q: I have a very practical question actually. If suppose I were to shop for a good Zen teacher, what can you recommend?

Yeah, totally. And you said you’re in Switzerland, right? Yeah, Zurich-based.

Yeah, okay, great. I know there’s a lot of folks in Europe, although I feel like there’s not as much... You know, I mostly know the Zen world. There is not zero Zen in Europe, but I know that there’s a lot more of other traditions in Europe than Zen.

My general advice is to find someone who you feel has their life together. That’s maybe the number one thing. There are a lot of meditation teachers out there and they maybe tell you that “Oh yeah, you do this stuff and you’re right,” but then their actual own life is a mess. This person definitely has some understanding, but I would be very nervous about taking them as my teacher and being in that kind of relationship with them.

This is perhaps very Zen of me, but I think that lineage is very important. It’s very important in the sense that their teacher has essentially certified this person understands and is emotionally mature enough to teach. I think that’s very important. I know that not every tradition has this or feels the same way about it, so take it with a grain of salt, but I think that’s quite important.

I’m not going to say that it’s perfect. I think no human institution is. And I think this is why it’s quite tricky to find a teacher, because if you work closely with a teacher on this process of waking up, you’re quite vulnerable. You have to be very vulnerable with them, very open with them in order to make progress. They have to be very vulnerable with you, and that is, as any time you’re in that type of relationship, it is possible to abuse it.

So I think it’s very important to be careful here. Pay attention to your intuitions. If you feel like, hey, this is not the right person, then they’re probably not the right person. If you get nervous about something, pay attention to that, notice that.

And otherwise it’s just like, go and try. That’s I think the best thing to do. If you haven’t found a teacher that you like yet or a sangha that you’re excited to be part of, just keep searching. There are tons of them out there. You will find the right thing for you, you just have to look.

Q: I have two questions. The first is, do you have a sense of how much your writings have helped people moving from a rationalist to a postrationalist stage? And the second question is, how would the world look different if the average person was a postrationalist?

Sure. On the first question, I don’t know. It’s kind of hard for me to figure this out. I know of a handful of people who have told me that they read things I wrote and it was very important for them. It was critical to their thinking and the development of their thinking. So I know that. But we’re talking like five, six, seven people here. So maybe there’s more people, I don’t know. I’d kind of like to think that there are, but although I’d also say that I try not to have too many delusions of grandeur here.

The reality is that if someone is, let’s say a rationalist today and they get interested in postrationality, it was actually going to be the culmination of reading things written by dozens of people that will be helpful to them along the way. No one person’s writing is really going to be sufficient, I think. I know we all want to be that person—or maybe not everybody wants to be this person, but as a writer it’s very easy to fall into this idea of “I will write the thing that wakes the world up” or something like that.

And occasionally there are people that pull this off. I think Eliezer Yudkowsky is maybe arguably a person like this. He has written a bunch of stuff that pretty consistently there are people that read it and their life is radically transformed by having read his writing. And I know because I meet them all the time—they’re many of my friends.

So this is, I won’t say it’s impossible, but I think the reality is that it’s a lot of stuff. And in many ways, even someone like Eliezer, it works because he’s building on a foundation that people are coming in with that actually comes from a bunch of different people. He is tying together a lot of sources, things that were written by other people, reusing examples, arguments that they had worked out, and then he’s taken them and made his own.

So maybe one day we will get the postrationalist version of Eliezer. Someone will do this. You could maybe argue that David Chapman is this, but I don’t think it’s quite there yet. I don’t think it’s 100% working. The machine isn’t working quite that way.

And then in terms of what was the second part of the question?

How would the world be different if the average person was a postrationalist?

Oh yeah. I like to think it’d be a lot nicer and kinder and more compassionate, and people would just be more reasonable about things. Yeah, I think maybe in my mind people would be a lot more humble.

I was talking to someone about this not long ago. They asked me a somewhat related question. They were like, “What’s the thing you want everyone to come away from reading your book? How do you want them to be different afterwards?” And I thought about it and I realized that the answer was I want them to have more epistemic humility coming out of it—deep epistemic humility, not the surface level epistemic humility that I think a lot of rationalists have, which is like “Oh yeah, I’m going to put probabilities on things and I’ll be nominally uncertain about stuff.”

I mean no, actually being very deeply uncertain that you even know what the hell is going on, that you are even correct, that you’re a person. Or that, you know, here I am drinking a cup of water out of my thermos. Is this a thermos? Am I even right about this? I’m not saying that you need to spend all your time navel-gazing and going down some very deep path of introspection here and being, you know, taking you in the direction of psychosis or something.

But that there’s a huge amount of scaffolding that the entire way we interact with the world in every moment is built on top of, and people just take that for granted. And I think that being able to see that gives you this kind of very deep humility that everything that I think I know is extremely contingent. And that in some fundamental sense, I don’t know. I don’t really know anything.

And I think this is sort of the end stage of the postrationalist project—something like “I don’t know anything, but I’m going to get on with my life anyway.” And I like to imagine that this would make the world a much more functional place.

But I don’t know. I don’t know if the average person being a postrationalist is in the cards, but it could be interesting to see if it ever played out.

Q: So you talked briefly about effective altruism and you talked about utilitarianism. I’m wondering, this kind of fundamental uncertainty about the world, if you have any takes on how that impacts the effective altruism project and in general our attempts at doing good in the world.

Yeah, totally. One of my favorite quotes that I want to say that maybe this is attributed to Anna Salamon, but I could be wrong about my memory here of who I originally heard this from, was that we’re great at figuring out the size of the impact of our work. We’re really bad at figuring out the sign though.

So it’s easy to figure out that yeah, we’re doing something, it’s having some sort of impact on people. The question is always, is this actually good or bad? And I think this is quite tricky.

Maybe just to pick on an example, cash transfers are maybe a great example here where it’s really hard to work out what are all the second and third and fourth degree order effects of cash transfers. On the surface it seems great—people are poor, let’s give them money so they can buy things and not be poor and bootstrap themselves out of poverty. Sounds great.

But there are a lot of interesting second, third order effects of implementing these sort of policies that are really hard for us to figure out. We might not know the answer for decades. If we run a cash transfer program for a long time and maybe the ultimate outcome is actually bad... I’m not saying that it is, I am saying that I hold a decent amount of uncertainty about this. And I feel this way about a lot of stuff.

There are some things that I think are a little bit more clear are straightforwardly good, although even there I’m like, I don’t know. For example, I think work on animal welfare mostly seems good in terms of making the conditions for animals in farming settings, especially factory farming settings, better seems straightforwardly good. But even there I hold some uncertainty about, well, I don’t know, what are the long-term consequences of this? Does it make farming more palatable to people and then actually we increase meat consumption because it’s less bad what we’re doing? I don’t know, it’s sort of hard to game this all out.

So that’s where I sort of say that we sort of just don’t know.

Q: So you’ve written about epistemic circularity and the problem of the criterion, right? So if we can’t escape circular justification, how does that inform how we approach these questions about what’s good and how to do good? Like how do we know at the end if we can’t escape that epistemic circularity?

Yeah, for sure. Well, I mean, I should be clear that the epistemic circularity that I’ve written about—and this is my book plug: you should definitely read my book. It’s on the web. It’s fundamentaluncertainty.com. You can find the book there. It’s currently in a sort of pre-print phase. There’s a draft of it that’s available and then I’m working on the final revision of it now for publication.

But yeah, essentially this is a problem that exists within our system of reasoning. It doesn’t actually exist in the world. It only exists because we’ve developed a system of reasoning that helps us model the world, but it is an artifact of that system of reasoning. And if we just don’t engage with that, this problem kind of disappears.

Now, this is not an argument for disengaging from all rational thought. And I think this is the thing that we have to be careful about to not lose sight of. It is instead saying that we can use rational thought to help us work out details of stuff. It is a very useful tool, but ultimately stuff is founded in our experience. And that is where all of our beliefs and everything come from.

So it means that if I’m looking for myself for what is important, it ultimately comes down to: what do I care about? What do other people care about?

I think that I’m very aligned with—maybe I’ll get in trouble for saying these words for reasons—but I feel like at least narrowly in this capacity I’m very aligned with Heidegger that care is the fundamental thing. This is the central thing at the heart of everything that is a part of our lives, of the way we understand it. Everything is built out of this thing that we want to, that it makes sense to call care. Of what is it that we’re motivated by? What is it that we are concerned with? What is it we are trying to achieve? And it all really flows downstream from there.

I don’t know how much of an answer that is, but that’s my sort of abstract answer to the question.

No, that makes sense. And it kind of reminds me of what you were saying earlier about within the utilitarian framework, the utility itself is not sort of examined or explained—it’s got to come from outside the frame in a certain sense.

Right, yeah.

Q: I do have one more question. I’ve read your sequence of posts on Zen and rationality, but I just kind of wanted to hear some more of your thoughts on how you see postrationality in Zen, or how is... I guess maybe I was thinking, how has your Zen practice informed your perspective on sort of questions of postrationality? Maybe that’s sort of backwards because from what the history you gave it seemed more like maybe your examinations of postrationality sort of led you into Zen practice to begin with. But I’m just curious to hear a little bit more about how you see Zen or maybe Buddhism in general or contemplative practice even more generally fitting in with these sorts of things.

Yeah, totally. Yeah, I would say that I came to formal Buddhist practice mostly because I was reading about all this stuff and then sort of a thread that I haven’t really brought up here is I was also interested in positive psychology stuff. You know, how can I make my life better by applying various sorts of positive psychology techniques? This is also very related to the CFAR project and what they’re doing. They’re fundamentally, I think—I don’t think they would frame themselves this way—but it’s a positive psychology project.

And yeah, sort of came to this place of like, okay, I’ve got all these ideas, I’ve tried all this stuff, and it’s still not... I can tell that there’s more and I don’t know how to do it.

And so over the course of about a year of searching around, I eventually convinced myself that I think Buddhism basically is the closest thing that we have to a living tradition of how to integrate this stuff. And I’m not going to say that this is true of every strain of Buddhism. I think some lineages are better at this than others. Some are very explicit about it, some are not.

I think for me I took it as: there is a set of stuff that Buddhists have figured out. They know how to teach it. That’s the part I’m missing, and I need to engage with it.

Now I would say that from having spent a lot of time in a lot of Buddhist spaces that Buddhism is not the same thing as metarationality or postrationality. There are in fact—the vast majority of Buddhists you will meet in the world are pre-rationalist, premodernist people and thinkers. And I don’t mean this in any way to be disparaging towards any of them. It is just that they have not yet fully engaged with rationalist thought or modernist thought in a way that would allow them to actually be sort of in this postrationalist, this metarationalist headspace.

So I really see there is sort of this synthesis here of taking all these parts together to build—I don’t want to say build a better Buddhism, I don’t know that that’s quite right—but I think that there is something here where we could imagine a Buddhist tradition that integrates more of these insights from the rationalist project.

And I think there are some teachers that do a bit of this. Shinzen Young is probably the first name that comes to mind in this space. But I think there are definitely other folks, and I think there’s opportunity here in terms of formulating a style of practice that is able to integrate both the anti-intellectual strain that exists within Zen—the sort of primitivist strain that is getting back to the fundamental reality of just being—and the benefits of being able to apply rationalist thought to actually live our lay lives, because we are not going to just sit on a cushion in a monastery. That’s not the life for most of us.

We have to live in the relative world, and living in the relative world means fully engaging with form. And that is sort of the culmination of what the rationalist project is—how do I fully engage with form? And Buddhism is how do I maybe fully engage with the absolute? And then integrate those two things together, the relative and the absolute together.

So yeah. I know we’re basically out of time, so maybe that’s a good point to stop.

[Closing remarks]

Yeah, I was just thinking the same thing. So I just wanted to say a big thank you one more time, Gordon, for joining us today. This I found super interesting and can’t wait to continue this conversation with you or these conversations with you. Thanks again, Gordon, for joining us today and this was really great. Looking forward to seeing you all again soon.

Great. Yeah, thanks so much to all of you for having me. Really appreciate it.

Thank you, Gordon. Thanks, everyone.

The natural questions to ask are “how do I do it?” and “can you do it, too?”

Let’s tackle these questions in order.

The Mechanics of 5x Engineering

Arguably, the way I achieve 5x productivity at programming is pretty simple: I use Claude Code to do almost everything!

But the reality is more nuanced, because lots of people are also using Claude Code to do almost everything and aren’t seeing similar productivity gains. I’ve learned a lot about how to work effectively with Claude in the 6+ months since its release. Here’s how I do it:

• At a high level, I think of myself as pair programming with Claude. I’m the navigator and Claude is the driver, though occasionally I ask Claude to help me navigate, as during planning, and I sometimes do a little driving in the form of jumping in to change small things in code.

  • I find this model superior to other ways of trying to use coding agents. In particular, I dislike letting agents run and submit a PR. I don’t trust the changes work, even if CI passes, and the iteration time is longer. Rather than have my attention repeatedly pulled in multiple directions, I want to work on one thing at a time, get it done, and then move on to the next one, just faster than if I was coding without an LLM.

• I almost always start with a plan, as in using planning mode. I give Claude some vague instructions about what I want and let it get familiar with the code base. It will come back with something. I then iterate with it on the plan until it looks basically right.

  • The most valuable thing I have that Claude doesn’t is context—context about the business, our users, their needs, how a feature should work, and, to a lesser extent, our engineering culture (it’s able to infer a lot about this from our code!). This is the most important stuff for me to tell Claude about when it’s planning.

  • If I’m working on something big or complex, I tell Claude I want to break the work into phases because I want to stop after each phase to create a PR in a stack of PRs. This helps keep it from getting lost because it’s trying to do too much, and because I actually do create a stack of PRs, which makes life easier come review time.

  • Sometimes I’m not sure if the initial plan Claude comes up with is good, so I ask questions. Claude can be a little overeager to make new plans, but if I keep pressing, it will explain its decisions and, if I think they are bad, I can work with Claude to come up with a better plan.

  • Speaking of which, I often dislike the initial plan. Sometimes I have an idea in mind of what we should do instead, but I will often feign ignorance and ask Claude for alternatives to avoid biasing it. Sometimes it comes up with a better idea than me, but usually it either comes up with the idea I had or makes a suggestion that helps me make my own idea better (and then I tell Claude to change the plan to do that).

• Once the plan is ready, I have Claude execute it. During this part I usually go do something else while it works. This might be working on another coding task, dealing with email or other messages, or catching up on Twitter if I can’t hold the context of two complex things in my head at once.

• When Claude is done, I carefully review its work. I trust nothing. I look over the diff and make sure I understand every part of it. If anything seems weird or I don’t understand something, I ask Claude what’s up with that part of the code.

  • Claude is biased towards action, so when asking questions, I often need to remind it that I’m just asking, not telling it to do anything yet. I’ll say something like don’t make any changes yet, but why is [some part of the code] like [whatever it’s like]?

  • I also ask Claude to plan changes before making them. This helps avoid a lot of rework and waiting for it to make changes that I’m going to undo. I sometimes go as far as putting Claude back into planning mode, but usually I just do planning in normal mode by being explicit with a prompt like don’t change anything yet. look at [part of the code] and give me options for how we might [make it better in some specific way]

    • I treat this process like a mini planning session, even if Claude isn’t in plan mode. So I ask lots of questions, try to understand tradeoffs, and come to a clear conclusion on what to do next before letting Claude write more code.

  • I’ll also sometimes go in and make edits. I can’t consistently get Claude to match my style preferences even with fairly aggressive prompting in CLAUDE.md, so I’ll go in and fix it manually. These are always small changes, though. For anything that’s going to take me more than a couple minutes I get Claude to do it.

    • When I do make edits, I always make sure to tell Claude I did this. Otherwise it’s liable to undo my changes. For this reason I often try to leave style edits until the very end, after I expect to make no more edits with Claude.

  • I find it useful to commit code as I go. This effectively gives me checkpoints in case Claude goes off the rails and does something weird that’s hard to recover from. I’ve regularly been saved from manual cleanup by being able to execute a quick git checkout .

• While I usually go do something else while Claude works, sometimes I find it useful to sit and watch it. I especially find this useful when I have low confidence that it will do the right thing, or when I’ve seen it wasting time doing extra steps after making code changes. Then I can interrupt it and correct course or stop it early and get on to the next thing.

• I always test Claude’s changes, just like I’d test changes if I made them directly. This should be obvious, but I own the code, not Claude, so it’s up to me to test it and make sure it works.

  • That doesn’t mean I can’t ask Claude for help, in particular to write automated tests. I’ll usually give it a vague description of what I want tested and what cases I care about, and it will code them up.

  • The tests Claude produces aren’t perfect, but writing tests is tedious, and I otherwise don’t write enough of them. Claude reduces the friction enough that I produce a larger number of useful tests in my code, even if those tests don’t meet the same quality bar I set for the application code.

  • I also test the code for real. I run it in dev and staging. Claude makes mistakes all the time, and I need to be sure the code actually does what I think it does. It’s not uncommon to discover that code looks correct but does something subtly different than intended or behaves contrary to expectations I didn’t realize I had until seeing the code in action.

• I use Claude to review the code. I start a new session and ask it to review the code. Don’t tell Claude that I wrote the code with its help, because that can bias it. I just ask it to review the changes. It can look at the diffs in the commit, read more code, and spot issues the original Claude instance missed.

  • In practice I also use other automated code review tools that run as part of CI, and sometimes I skip local Claude review because of that. The point is that LLMs are great at noticing mechanical mistakes in code that humans often miss, and I rely on a fresh set of LLM eyes to catch those before they make it to production.

• I use Claude to refactor the code it produces. The initial run is usually full of code smell or exposes latent smell, and this is where a huge chunk of my productivity boost comes from. Rather than shipping more technical debt on top of existing debt, I’m able to continually pay down debt with every PR because the cost of doing refactors with Claude is dramatically lower than doing them by hand.

  • For example, code organized into the wrong directory structure? That’s a pain to fix by hand. Claude can do it in a few minutes while I get a fresh cup of coffee.

  • Sometimes I need to talk with Claude about the refactor and plan it. The process looks like this:

    • hey, this part of the code looks weird to me. what’s up with it? don’t try to fix it, just explain to me why it’s like this

    • Claude answers

    • okay, what are some options for how we could improve it

    • Claude proposes some solutions

    • let’s go with option B, but change it to include feature 2 of option C

    • or actually, i don’t like any of those, let’s do [something better] instead

  • Again, if I need more information to make a decision, I ask for it.

    • Frequently I have to explicitly ask Claude to look at docs or otherwise search the web for it to get correct information about how, for example, a library or API works. This is basically always worth the time it takes.

• Finally, I never get mad at Claude. Sometimes it’s kinda dumb or includes a change that I don’t notice until a reviewer asks me about it and then I look dumb because I didn’t know it was there, but that’s my fault. As I said, I own the code, not Claude, and that means I’m both responsible and accountable for what Claude does. I find that this mindset is critical to using Claude Code effectively.

Can You Become a 5x Engineer?

Yes. Probably. Maybe. If you use Claude like I do, the quality of the results you get will depend heavily on how good of a programmer you are.

From what I can tell talking to other people, I work well with Claude because I have a lot of programming experience—over 30 years’ worth! I’ve developed strong intuitions about what good code looks like, and have clear ideas about what patterns will and won’t work. I can steer Claude towards good solutions because I know what good solutions look like. Without this skill, I’d be lost.

Now, I earned my intuitions the hard way, by spending years writing code by hand, carrying my punch cards uphill both ways. If you’re in that situation, great, you can pair with Claude Code the way I do to good effect. If not, is there anything you can do?

I think so. You can learn by talking to Claude and asking it to explain things to you. If you have a few basic motions to ask about tradeoffs and to ask for help understanding why Claude thinks some change is a good idea, you can use it to bootstrap yourself and gain a lot of experience rapidly. This might not be a total replacement for years of experience, but it’s better than nothing, and realistically, it’s what you’re going to have to do if you’re just starting out and want to get productive enough to get and stay employed.

Because if there’s one thing you should come away from this post with, it’s not that there’s a bunch of ways to use Claude or other coding agents to increase your productivity, it’s that you must start using these tools to become more productive so that you aren’t left behind.