Notes on Commandments to live by

Rough personal notes exploring many variations of "The Ten Commandments". Written by someone who knows little of religion or moral philosophy! It's part of an ongoing experiment in doing more early, rough, and personal background thinking in the open. I'd enjoy chatting with thoughtful people who are interested in this -- please post comments in the comment box below.

As a child, my brother and I would watch the movie "The Ten Commandments" every Easter. In the movie, the Director, Cecil B. DeMille, makes the comment that: "It is impossible for us to break the law. We can only break ourselves against the law"¹. This puzzled me. Surely the point of the Ten Commandments is to tell you what not to do. If the law is impossible to break, then why is the law needed? And, in any case, isn't it self-evidently possible to break the Ten Commandments? You can certainly steal, make graven images, and so on. I couldn't understand how DeMille, a thoughtful person and dedicated filmmaker, could have made such an obvious error. It wasn't until years later that I understood the truth underlying DeMille's assertion.

I recently came across the physicist Leo Szilard's personal version of "The Ten Commandments". Originally written in German, they were translated by Jacob Bronowski, and appeared in Frontiers in Science in 1971. The following notes contain a few of my resulting reflections on such lists. I also explored the question: what would I put on my own personal list? Not as an injunction to others! But rather as an exercise in thinking about my own beliefs. This turned out to be difficult but rewarding. It's easy enough to construct an impressive list, the right kind of list. But what I found more demanding was to construct a heartfelt list. I did not anticipate that moral exploration, if it is both heartfelt and public, is not only demanding but also exposing. These notes record a few of my explorations in that direction, in the hope that they may elicit connection in sympathetic others.

Szilard's Commandments

One could write volumes about these Commandments. But let me make just a few observations.

The archaic language is a mistake: "Touch no dish, except that you are hungry" is probably a joke – Szilard was rather portly by the time he wrote his Commandments – but is nonetheless better as "Don't eat unless you're hungry"². The philosopher A. C. Grayling used similarly archaic language in his "The Good Book", an attempt to write a modern secular Bible. It worked poorly. As I wrote elsewhere of Grayling's language, "I doubt he thinks this way, or has much experience writing this way… I'd guess he's aping some kind of style. It's as though he's writing a book in a language he only knows poorly. Doing so makes it harder to think well." Szilard's Commandments are better than Grayling's book. But they also suffer from this problem. Szilard doesn't think in this way; by writing in this way, he puts his thinking at a disadvantage. It's like driving with the hand brake not fully released.

Why Ten Commandments? Why not 10,000 Commandments? After all, life is complicated. This is why religious texts and self-help books contain detailed explanations and elaborations, not merely short lists of Commandments³! A story apposite for every situation! Presumably, underlying a short list of Commandments is an 80-20 belief: that much of the overall value is contained in a short list. This is, in part, an intellectual statement, about the content⁴. But it's also in part a statement about the practice required to achieve depth of understanding and actual embodied action. It is arguably better to engage dozens or hundreds of times with a list like Szilard's – applying it, critiquing it, thinking about how it applies to one's own life from many different angles, juxtaposing with other ideas, improving it – than to engage more shallowly with a longer document. It becomes a practice – something to return to and gradually deepen – rather than a large body of ideas to draw upon⁵.

The connection to self-help: Szilard's Commandments are recognizably part of the self-help genre. Self-help has a curious status in our culture. It's often dismissed as lowbrow by the intelligentsia, who may be scornful or scathing. "It's not scientific!" "It's just anecdotes". "Many self-help authors are grifters". All often true enough. And yet, it's not as though we can escape the questions addressed by self-help. What do you do if you're lonely? What do you do if you're desperate to lose weight? If you'd like to meet the love of your life? These are complicated issues – some people find them easy to solve, but if you've struggled with these problems over years⁶, then trying to solve them "scientifically" yourself may not be the right approach. You're probably much better off chatting with wise friends, who will likely give you a (personalized) lite version of self-help. You could, perhaps, take a year or two off to synthesize the relevant scientific literature, and run individual experiments. And what you'd find is partial and incomplete answers, which you'd need to fill in with informed guesswork from personal experience and conversations with experts and the wisest people you know. In short, in the best case you'd end up with something pretty much like the best self-help books. What Feynman said of psychiatry applies to self-help in general:

This is a reason I strongly dislike the dismissal of self-help as low-brow. Yes, large chunks are hokey. But the best bits are the best humanity has for some crucial life questions. It's foolish to look down our noses.

There is an exception to the dismissal of self-help as low-brow: if a text is old enough, it often gets a pass. The Bible, Marcus Aurelius, the Tao Te Ching, and others are all in this vein: all are widely (though of course not universally) treated as venerable sources of wisdom. And yet they come out of a similar tradition to modern self-help. The Ten Commandments have an exalted status in our culture – militant atheists aside – but they're not so different from a listicle from a content farm. They've merely got a reputation laundered by time and association with important institutions. This may seem crass and as though I'm dismissing the Ten Commandments, but that's not my intent. Rather: if you distilled and carefully remixed much of the best material from the content farms, I expect you'd get something surprisingly good, perhaps even rather original⁷.

There is enjoyment and insight in attempting to improve items in Szilard's list: This is part of why it's good that some statements are cryptic. The list is meant to stimulate conversation and reflection and (iterated) action, not to be a detailed instruction manual. Much like good song lyrics! Indeed, some items are arguably so misleading that they need to be completely rewritten or abandoned.

Consider "Do not lie without need". This is hopelessly vague. Most lies I'm aware of are said for what seem to the liar to be a good reason. I believe – certainly I hope! – that's true of most of my own lies. But even though the Commandment as stated is hopelessly vague, it still stimulates useful thought. Why and when are lies bad? Why and when are they truly "needed"? How do bad lies damage the liar? How do they damage the person or people lied to? How do they damage other people? The world? These are all good questions, worthy of much reflection. I've gradually come to believe, for example, that lies damage the liar much more than the liar is usually aware of. This is in part because each lie disconnects the liar a bit from their own understanding. Sometimes that damage is small, and outweighed by other benefits. But it's often surprisingly large⁸.

In a similar vein, many of the other Commandments are wrong, when considered literally. I violate "do not destroy what you cannot create" almost every time I take out the garbage. But although literally wrong, there is some grain of truth, and it's stimulating to work to uncover it⁹. Put another way: although the Commandments are imperative, good Commandments are (implicitly) questions. They challenge us: how would you improve this?

The trouble with Moral capital-S Systems: Consider: "Let your acts be directed toward a worthy goal, but do not ask if they will reach it; they are to be models and examples, not means to an end." I instinctively find myself in sympathy. I'll come back to the reason in a bit. First, though, a digression. One reason I felt awkward writing these notes is that I'm ignorant of modern moral philosophy¹⁰, and no doubt much of what I write will seem naive to people steeped in the field. I have tried reading into moral philosophy a little. And I've failed to make much headway through a head-on approach. I find myself rejecting or skeptical of many premises that seem unstated to me¹¹ – enough so that I find it almost impossible to read.

One instinctive challenge is that the people who engage in moral philosophy often don't seem like unusually good (or bad) human beings. Indeed, the "best" people I've met – the people who make me think "wow, this person has figured out how to be a good human being" – are usually not all that versed in moral philosophy. Now, I realize this may be read as an attempt to dunk on moral philosophers or people interested in moral philosophy! But that's not at all my intent. I'm just observing that the connection between such an interest and good behavior seems weaker than might perhaps a priori have been suspected. As so often, speaking well (or thinking well) and acting well are only loosely connected.

Just to overthink things¹², there seem to be two competing traditions here. One is moral philosophy, concerned with developing Moral Systems. I don't know the history in any detail, but this seems inspired in many ways by mathematics. Figure out some pretty reasonable basic axioms or models. And then try to explore and understand their consequences. It's not quite proving theorems – the Systems are rarely unambiguous enough for that. But "making arguments" to figure out what is right. And then there's the second tradition, which is what 99.99% of human beings use: just muddling through, talking with friends and family, watching other people, trying to figure out how to live in the world, to be a good person, to do right by others and by oneself.

I think the second tradition is usually much more powerful and reliable¹³. And the reason is that the world is immensely complicated, and as a result experience is much richer than any such System. It's a situation where, for now, simply exploring reality is in many respects far more challenging than such a System. It's the difference between attempting to deduce biology from a few simple ideas, and determinedly exploring the biosphere. The actual biosphere – the biosphere we can explore – is immensely complicated, and exploring it has (so far) been much more rewarding than attempting to understand things from theoretical first principles. With that said, a benefit of the Moral Systems is that one can push them in ways you can't (easily) in the world. Clever thought experiments, unusual questions – those are genuine generative benefits. But while I have no doubt that generates novel moral ideas, I have my doubts about whether it generates reliable moral insight¹⁴.

That's a lot of throat-clearing to say: I'm fundamentally very suspicious of any strong notion of a "Moral System" at all, or even of notions like consistency and implication in such a System. They're certainly useful generatively, to stimulate thought. But many people want to take them very seriously as a basis for extended lines of reasoning and action. I think that's usually a mistake, resulting in complex (and sometimes selfishly motivated) justifications for actions which would seem obviously wrong to any intelligent 10 year old¹⁵. When I say this, people interested in Moral Systems sometimes want to debate: such debate seems to me (with rare exceptions) a bad use of my time. They're insisting on using the first tradition, rather than the second. And it's much healthier to relate primarily to people's actual experiences, and only secondarily to Moral Systems or theories of what is good¹⁶.

You may wonder why I wrote these notes, if I think this way about Moral Systems(!) But I think of things like the Ten Commandments (and similar lists) rather differently. They're certainly not attempting to be a logical deductive basis for right action. It's much more helpful to think of them as a provocation and focus for moral thought or a moral practice, much like a good novel or movie or biography. They're not a logical system at all, but rather the answer to an implicit question: how can we 80:20 what we've learned in practice about behaving well?

Extended digression aside, let's get back to Szilard's Commandment: "Let your acts be directed toward a worthy goal, but do not ask if they will reach it; they are to be models and examples, not means to an end." My instinct is that my life has (mostly) been better when I've followed this Commandment. And this is true of other people I know well, too. But there, of course, caveats. I think that going to the Andromeda Galaxy is a terrific goal; but it's not a realistic goal, right now. I suspect it's better, rather, to soften the statement. Something like:

One could, and likely should, iterate many more times upon this idea. Let's try a couple more times. One critique: depending on your personality, it may be worth worrying a great deal about whether or not your acts will reach your goal. Some people need that security; others need it much less. A second critique: worrying about whether you will reach a goal often helps improve the actions you are taking. That is: there is a kind of useful worry one can do. I believe that what Szilard is pointing at is a kind of useless rumination it's easy to get stuck in. So perhaps we can modify it as:

It's getting wordy! And vague, with weasel words like "occasionally" and "most". Let's see if we can improve it by being more concrete:

Not too bad. I think it's better than Szilard's initial formulation! But my instinct is also that there is something slightly off about Szilard's initial formulation, but there is an important idea inside it. So let me take a step back, and try to express the underlying idea in my own words:

That's also too wordy, and veering into cliche. At the same time, it seems to me to be more correct. In any case: I believe these ideas could be improved much further through many more iterations. I won't do that now – I'm just exploring, trying to get a sense of how to proceed.

Our new AGI Overlords' Commandments

Earlier, I mentioned in passing the idea of distilling the wisdom of the web. In that vein, I asked ChatGPT for a recent update to the Ten Commandments:

Not bad! Worth pondering, though the moral authority is diffuse and rather obscure. There's also perhaps more overlap between the Commandments than seems ideal. It's the wisdom of the ancients, in a modern, curve-fitting edition. It'd be fun to edit the Commandments, and iterate on them with ChatGPT's help. Fortunately, ChatGPT didn't come back with: "I am the Lord thy AGI, and thou shall have no other AGI before me". So I tried it again with Claude, the similar system from Anthropic. The same prompt gave the following response:

Also interesting! An intriguing point about the large language models underlying these systems is that their training is based on learning to predict the next token in long streams of existing text. But, as I've written elsewhere, "a big part of thinking is to create meaning, by finding ways of violating expectations. The language models can seem instead like ways of rapidly generating nearly content-free cliches, not expectation-violating meaning." Put another way: it often seems challenging to use them to find really original, fundamental ideas¹⁷. Later in these notes, we'll consider Christ's second Commandment: "You shall love your neighbor as yourself." Whoever first wrote that line was a moral genius, capable of an extraordinary leap of moral imagination. Not that the idea is necessarily correct or that no-one had previously had similar thoughts, but: it is an extraordinary condensation of a deep and difficult moral idea. It's interesting to ponder what would be required for an AI system to have similar moral insights. More generally, moral progress has often depended on leaps of moral insight, often arising out of some moral crucible – think of the struggles underlying the overthrow of slavery, or the movement toward civil rights. Moral imagination is a real thing. I feel this vividly in, for example, famous passages from Lincoln and Martin Luther King, Jr. They contain remarkable ideas, focused and distilled by Lincoln and King, and no doubt shaped by the moral crucible in which they lived and worked.

The Abrahamic Commandments

Let's switch now to the "standard" Ten Commandments. To distinguish them from the previous Commandments, I'll refer to them as the Abrahamic Commandments. The full text is rather long, and contains much that does not seem relevant to modern audiences. Let me instead give it in the abbreviated form suggested by Wikipedia, and based upon the King James Bible. It begins with "I am the Lord thy God", and continues:

External focus: Many of Szilard's Commandments are about your relationship with yourself. The Abrahamic Commandments are much more about your relationship with God and with other people. One possible explanation is that we live in a more individual-focused age now. If you view effectiveness or goodness as primarily individual qualities, then advice may fruitfully be aimed at improving individuals. If you view effectiveness or goodness as primarily collective qualities, then the more external focus makes a great deal of sense. A very different explanation – which, on balance, I find more plausible – is simply that much of the origin of the Ten Commandments is in earlier codes of law. Codes of law are concerned with ordering society and relations between people, and naturally have a social focus.

It's fascinating to compare to the Four Noble Truths of Buddhism, which are – even more than Szilard's Commandments – focused on internal and individual experience (with the partial exception of the fourth Noble Truth). Indeed, insofar as the Noble Truths express the core ideas of Buddhism, and the Commandments express the core ideas of Christianity, there is surprisingly little overlap in domain of applicability, and a natural syncretism possible.

They are about what is good, but [mostly] not about power: This is true of Szilard's Commandments, the Ten Commandments, and the Four Noble Truths. Indeed, much of the Bible and many other religious texts either ignores power, or is quite sympathetic to power¹⁸. Indeed it's striking that the Gospels (only a tiny part of the Bible, of course) violate this: Jesus has considerable sympathy for the underdog. This sympathy is perhaps most strikingly exemplified in the Beatitudes, e.g., in Matthew:

It's possible to read this cynically: as justification for keeping people down; or apathetically, as a way of reconciling people to their fate. Insofar as the Bible is the work of a committee, there may be some truth to this (for some committee members, any way)! But given the remainder of the Gospels, I believe it likely it also expresses some genuinely original thinking about power relations and inequality.

The Abrahamic Commandments are far more imperative than Szilard's: By comparison, Szilard's lists might well be called "The Ten Suggestions".

The Abrahamic Commandments are (mostly) far clearer than Szilard's: Think about Szilard's first Commandment: "Recognize the connections of things and laws of conduct of men, so that you may know what you are doing." I think I know what he's trying to say, but I'm not 100% sure. It's a little too obscure. By contrast, the Abrahamic Commandments are clear, but still invite contention and discussion. That combination – clear, but inviting discussion – seems to me ideal.

Back to DeMille: So what did DeMille mean when he said: "It is impossible for us to break the law. We can only break ourselves against the law"? Both Szilard's and the Abrahamic Commandments are prescriptive or normative statements – statements about how you should behave. But they're grounded in an understanding of how the world is. Consider: "do not covet what you cannot have". This is a prescriptive statement, but it's based in descriptive facts about human psychology. If you covet what you cannot have you will waste time you could have better used in other ways. You are likely to make yourself miserable. You are likely to be driven into actions destructive of self and of others – perhaps you will steal, or violate other people's trust, or the trust of society. These are (admittedly, oversimplified) descriptive facts¹⁹ related to coveting what you cannot have. Insofar as they encode a true understanding of the world, you can't break them, they simply are. You can, however, break yourself upon them, by trying to behave as though they are not true. I believe DeMille was saying that underlying the Ten Commandments are some descriptive facts about the world, and the prescriptive content of the Commandments mostly expresses how you ought to behave in response to those descriptive facts²⁰. In a similar way, Szilard's Commandments are prescriptive, but reflective of underlying descriptive facts. The first two of the Four Noble Truths, by contrast, are closer to being descriptive facts about the world, especially about human psychology. The final two of the Four Noble Truths express what one should do in response.

Christ's Commandments

In the Gospels of Matthew and Mark, Jesus is asked which is the most important of the Commandments. He names two, with the first called out as "the greatest and first Commandment":

These capture and distill much (though not all) of the Ten Commandments. Indeed, in many regards they actually strengthen the Ten Commandments. They're less specific and legalistic, but also deeper. Reflect upon them enough and you may well arrive – or at least consider – ideas like abolitionism, civil rights, the evil of racism, the need for gender equality, and many others. Perhaps, even, veganism – at the least, one begins to wonder who, exactly, counts as your neighbor. Does it include animals? Peter Singer's hypothetical drowning child on the other side of the world? Or in the future? They're ideas with enough depth to reward a lifetime of thought.

Christ's Commandments are intimidating. "You shall love your neighbor as yourself". Who lives up to that? Certainly, I know many people who could and likely would make intelligent-sounding arguments against it. But it's also hard not to suspect those arguments are often rooted in motivated reasoning, no matter how impressive-sounding they are intellectually. I will say one thing in modification: we do have some responsibilities toward ourselves that cannot be assumed by others. And so the naive reading of treating others as we would treat ourselves (and vice versa) must be wrong.

As an atheist, it's tempting to reject Christ's great Commandment, or at least to try a considerable variation. One obvious thing is to take the pantheist route, making it: "You shall love the Universe with all your heart, and with all your soul, and with all your mind… Love the Universe above all else." It's unsatisfactory, though. Christ is setting up "the Lord your God" as a personal embodiment of all that is good, of good action, and of order. We're good at loving people – and it's very good for us to love people²¹ – but not so good at loving things, especially rather abstract things like "the Universe". It's also (very) tempting to reject. But I think that's a mistake, for much the same reason: it's very, very good for us²² for us to love other people selflessly and wholeheartedly.

So I'm unwilling to set the Commandment aside. Instead, I tried a few other variations. Everything I found was inadequate: "Love the Good" [too abstract]; love some saintly person who actually lived [too subject to their peculiarities and weird power dynamics]; love a near-mythical person. The last is probably the best, and is arguably the Christian approach, with Jesus-as-mythologized (as opposed to any actual historical human). You get the benefits of anthropomorphization, but at the same time Jesus is a stand-in for God. There is, of course, the problem that you know you're reading what is, at best, very stylized history. Still, it turns out we humans are pretty good at doing that, too.

Feynman's "Commandment"

Christ's great Commandment made me ponder related statements. There's a striking example, in a very different vein, in the opening to The Feynman Lectures²³. Emphasis mine:

Feynman's sentence – it's not really a Commandment, it is simply about how the World is – is quite a peculiar statement. It hasn't always been true: up to about 370,000 years after the Big Bang, the Universe was too hot for atoms to form bound states, and we simply had a plasma. Indeed, go back far enough and you get to the pre-hadronic era, where even protons hadn't formed! Still, Feynman is right to say that in that one sentence there is an enormous amount of information about the world.

The Ten Commandments (and Christ's Commandments) are about how you should live, for social order and harmony; the Four Noble Truths are perhaps best described as being about how your personal experience actually is. By contrast again, Feynman's sentence is about how the world is. And so, again, it's interesting to ponder the potential syncretism.

Feynman and Szilard were both theoretical physicists. Perhaps that merely reflects the fact that I am a theoretical physicist, and so am likely to be aware of statements in this vein by theoretical physicists²⁴. But I suspect there is a kind of attractiveness to the notion of (if not the details of) the Ten Commandments, if you're a theoretical physicist. Such a venture attempts to pack as much meaning into a few powerful statements as possible. That's something theoretical physicists are strongly temperamentally inclined toward. A Unified Theory (of Life). That said: while I very much like Feynman's sentence, I don't actually much like Szilard's Commandments. They feel ad hoc and uncareful as a group. By contrast, there is an underlying unity of thought to the Ten Commandments, the Four Noble Truths, Christ's Commandments, and many other such collections. Those unities are different in each case. But they are there.

Miscellaneous observations on the "Commandments" so far

Obviousness: bug or feature? The idea "love your neighbour as yourself" is surprising. It's possible to dismiss cheaply: I suspect (though do not know) that Ayn Rand would have done so, for instance. But if you avoid a cheap dismissal, then it is an idea to struggle with. It is, I think, relatively easy to deepen one's understanding of the idea, and to improve the verbal formulation; but it is extremely difficult to come to terms with.

By contrast, "thou shalt not kill" – an idea arguably implied by "love your neighbour as yourself" – isn't so surprising. It's ambient in our society as a strongly held norm (within a certain scope), and as part of our legal system. But because it's been so strongly established as a norm, it's faded into the background, and become something we don't think much about. Intellectually, of course, it still raises many interesting questions. Is it okay to kill people in war? Is it okay for the state to kill criminals? What about citizens killing in self-defense? Where and under what conditions does the norm break down? But for me, and I expect most (though not all) people reading this, it's not an idea one usually spends much time struggling with. In this it's very different from "love your neighbour as yourself", which is an idea to struggle with, in an ongoing fashion, to iterate and improve. In that sense, the latter question is perhaps a better foundation for an ongoing reflective practice.

I'm not sure what I think about this. The curious thing about ambient conditions is that they: (1) mostly tend to become pretty boring and taken for granted by people within them; yet (2) they can be vastly different (there are historic times and societies where killing has been common); and (3) they often express profound ideas (indeed, the idea you should not kill is profound, and requires many related ideas and infrastructure); yet (4) because the ambient conditions don't change, it's difficult to really engage all the time, as part of a routine reflective practice.

I suspect it's mostly in changes and contrasts of ambient conditions that interest lies. You see this with the Covid pandemic: the idea of freedom of movement was pretty dull-seeming in 2019. Not an interesting subject to discuss at a party! But quarantine and stay-at-home orders have shown to everyone that this is a profound idea, connected to many other profound ideas. People now are engaged. In a similar way, the idea of pacifism arises most naturally in the context of a war. Indeed, that makes "thou shalt not kill" very salient. In any case, this is an axis to reflect upon: whether a principle reflects a desired ambient condition or whether it's a principle to struggle with in an ongoing way. From the point of view of the construction of society it is the former that is most interesting; but as the foundation for a reflective practice it may be the latter.

The comparison to Silicon Valley and "ambitious" self-help: Lacking from all the lists discussed so far is much mention of individual effectiveness or impact²⁵. None talks (directly) about how an individual can become important, or powerful, or wealthy. They're instead about how to have a good life and just society, and how to do right by the people around you. These things aren't unconnected to individual effectiveness. But they're not the same, either. In this, these lists are quite different in focus from much ambition-oriented modern self-help literature, in the vein of Dale Carnegie. It's tempting to think such ambition-oriented literature must be a modern invention. But it's not. Consider, for example, "The Instructions of Shuruppak", an example of Sumerian "wisdom literature" dating to circa 2500 BCE(!!), which includes such admonitions to the reader as: "You should not make a well in your field: people will cause damage on it for you"; and "You should not vouch for someone: that man will have a hold on you; and you yourself, you should not let somebody vouch for you, that man will despise you". Such ideas are not far from ambitious modern self-help, and seem to have been common in ancient texts. I find it very striking that such individual orientation is largely missing in many religious lists.

Veneration or iteration? Most religious texts are static: you can argue with the Bible, but it's unlikely you'll change it. It's as though we were still using the Principia as the source of scientific understanding. This makes little sense. It's good to have a canon of classics. But it's not good when that canon ossifies. The Protestant Bible removed 7 books from the Catholic Bible – a half millennium ago. By contrast, rapid iteration would mean modifying verses every week, replacing books every year. In practice, of course, religions do change. But the changes are usually quite slow compared to many other communities of practice. Maybe what you want isn't the Ten Commandments, but rather the Temporary Commandments.

Something I admire about the Effective Altruists is that they (at least in principle) don't assume they know what is good in advance; rather, they seek to figure it out²⁶. As I've said elsewhere²⁷, they're trying to do for the question "What are good actions?" what science has done for the question "How does the world work?": construct a community that iteratively improves the best answers. I admire that, even if in practice EA has many issues²⁸. At the least it suggests a more dynamic approach to determining "what is good".

My Commandments? Your Commandments?

In the spirit of Szilard it's enjoyable to attempt to develop your own personal Ten Commandments. It feels hubristic: you're making up your own version of one of the most venerable documents in human history! Who are you to be attempting this? But of course it's of intrinsic value to consider such questions. It would be a little like regarding cooking as hubristic because you are not a Michelin Chef: you still need to eat, and even modest efforts are of value. My experience of developing my own commandments has been that it became more enjoyable once I began to treat it as a fun game, not to be taken too seriously. I suspect Szilard treated it this way.

What should such a list be for? Should it be to improve your social relations and relationship to God, like the Ten Commandments? (The latter part is challenging, to put it mildly, as an atheist.) Or maybe individual self-development, more like the Four Noble Truths or Szilard's Commandments? Something else entirely? Without dwelling upon the question too much, I will make a choice: if you – generic you! – could get 15 year-old you to discuss a list of 10 things with your friends each week for the rest of your life, what would that list be? For the purposes of benefiting your community and society, of doing good in the world, and of having a good life? These should be ideas deep and stimulating enough to reward a lifetime of thought and engagement.

One consequence of these design choices (and the brief nature of the list) is that the ideas in the list will inevitably seem simple and obvious, perhaps simplistic. One comment I often hear about self-help advice is that it's "obvious". While that is often true, people also often have trouble following the advice, even when it is good. The benefit of a practice focused on simple-but-fundamental-ideas is that it really is worth returning to fundamentals over and over. It's like the basic forms in yoga or Tai Chi or scales for the piano. Ideas like the Four Noble Truths may be criticized as simplistic; but it's difficult to develop a good relationship to them²⁹, requiring continual practice and engagement over many years.

All that said, I'll now just note down some things I instinctively feel are directionally correct. At first, I shall make no attempt to debug them, nor to consolidate styles, though I'll then make two passes to improve them. One challenge I found in the construction: it feels cliche and redundant to include ideas like "Don't kill other people" or "Don't steal". On balance, I've included some such cliches anyway: an idea may be a commonplace, and yet still reward ever-deeper engagement.

I feel surprisingly self-conscious and uncomfortable about the notes. They are too much intellectual analysis, not enough fully (or even partly) metabolized feeling and lived experience. They are emotionally guarded, when such an endeavor is best done unguarded; this is understandable, given that sincere public pursuit of such an endeavor is easy to mock; I suspect few things damage our own understanding more than the desire to appear knowledgeable and sophisticated. In any case, it seems likely that wise Commandments must come from heart and mind and community; while I have leaned too much on mind alone. Certainly, a pattern that I (strongly) recognize in myself is to turn this into an intellectual game. I also know that the people who can talk most compellingly about doing the right thing are rarely the people doing the right thing. When I saw Szilard's Commandments I thought writing these notes would be a brief, amusing (for me) intellectual diversion. But of course that's incorrect: it's more like scaling every mountain in an enormous mountain range, and I have just barely begun to approach the foothills. The notes are a tiny step in a long-term practice of living.

Enough throat clearing! Let's take a janky, instinctive first shot, prioritizing instinct over impressiveness. Turns out I wanted Eleven Commandments, not Ten, at least in this instinctive list:

Let's do a round of criticism, both of items in the list, and of the list overall. What is missing? What can be improved?

I found it surprisingly difficult to improve the traditional entries. To return briefly to the Abrahamic Commandments, consider: "Thou shalt not kill". As mentioned earlier, this may be contested in many ways. Is it okay to kill in self-defense? What about in War? Is it okay if the State kills? How broadly does the proscription apply? Is it meant to apply to all humans, or just your in-group? Does it apply to other primates? To animals in general? To aliens? To artificial life? And so on. Of course, you can modify the Commandment to address these issues. Perhaps a determined pacifist would leave it unaltered as "Thou shalt not kill". But most people would add provisos, along the lines of: "Thou shalt not kill, except in the last need of self-defense". And perhaps add a scope: "Thou shalt not kill any human, except in the last need of self-defense". Through multiple rounds of revision you can make the Commandment much "better". But while more accurate, it would not be as stimulating as a moral provocation. If you think of Commandments as Questions, rather than Answers or Directions, then many traditional Commandments seem pretty close to local optima. That's why I found it so difficult to improve the traditional Commandments on my list.

But while traditional Commandments are often local optima, that doesn't mean higher local optima can't be found. The reason I didn't include things like "Thou shalt not kill" on the list is because "Love your neighbor as yourself" seems (mostly) much better than "Thou shalt not kill". It's a more powerful, more accurate, and more thought-provoking idea. It (arguably) lacks for concreteness: you need to think a bit to realize that it entails ideas like not stealing and killing. But overall it seems a major improvement. That said: it again has many problems. For instance, some people don't love themselves, or love themselves in unhealthy ways. Second, it's unclear how tightly scoped this should be. Is it referring to all people, or just literally your neighbor? Does it include the people you're at war with? Does it include animals? Aliens? Who exactly is included? Who is a who? It's easy to make modifications to take such issues into account. Maybe it becomes: "Love other people as yourself", broadening the scope. And then: "Love other people as you ought to be loved", to take into account some people not loving themselves. But when I make these kinds of changes I again get things which are more accurate, but less thought-provoking. "Love your neighbor as yourself" is, again, pretty close to a local optimum.

Let's run through my prior list, and for each item spend some time pondering local improvements. I note in advance, with some abashment, that I personally often fail at many of these in practice. Working on it! Furthermore, of course one does not merely think oneself into changed behavior: it only comes through repeated action and reflection³⁰.

Piet Hein said "problems worthy of attack prove their worth by fighting back". And so I have found this exercise. Of course each of the above ideas could easily be the subject of a book³³, not to mention thousands of hours of lived experience. In any case: let's collect the revised version. But let's reorder them, placing them into three groups: relationship to the social other; relationship to self; and relationship to the universe. I'll also en passant make some further improvements to the text. The result is 7 ideas or principles:

Looking at the above: they're not very good. There are many local issues. There are tensions and even contradictions between them; there is no sense of overall unity³⁴. The style is all over the place. I suspect I could have done better – in the sense of superficial appeal, playing to an audience – simply by taking 30 minutes to pound out Ten Commandments that "look good". The benefit of this exploration is not in the end product. It is rather in wrestling with some of these ideas from a different place than I have done before. And, I hope, to have found a place to stand, to wrestle with them further in the future, the Ten (Seven?) Commandments as a kind of Living Practice. Let me conclude with a few thoughts for future actions:

If you've followed me to this point, thank you! I'd be very interested to hear your opinion in the comment section below. I'm especially interested in your comments on the following questions:

Acknowledgments

Thanks to David Chapman, ChatGPT, Laura Deming, Andy Matuschak, and Howard Nielsen, for conversations directly related to these notes.

Footnotes

This quote appears in more than one variation on internet quote sites. Unfortunately, my digital version of "The Ten Commandments" – purchased on DVD many years ago, during the heyday of DVDs – has a glitch which makes playback unreliable. Pure stubbornness makes me refuse to purchase another copy. I'd appreciate if anyone can verify or correct the quote.↩︎
In some versions of Bronowski's translation, Szilard's fifth Commandment is rendered as "[untranslatable pun]". This is the case, for example, in Spencer Weart and Gertrud Weiss Szilard's 1978 book "Leo Szilard: His Version of the Facts". I haven't seen Bronowski's original document, but suspect "[untranslatable pun]" was probably Bronowski's choice. Still, the same remark applies to the English translation of many of the other Commandments. I do not know if this is also true of the original German.↩︎
It's interesting to ponder the Tao Te Ching as an intermediate. It's roughly 5,000 Chinese characters, in 81 chapters. It reads as something like the 500 Observations. It's mostly not very Commanding.↩︎
This 80:20 Pareto structure of knowledge is fascinating. Of course, similar power-law type behavior occurs in many phenomena. But usually it's possible to make an argument – typically from preferential attachment – for why we expect such behavior. There is not obviously any such argument here: it seems to be a fact about the structure of knowledge. There is a related metascientific question that puzzles me: why we see power law (or similar) behavior in citations to papers? Preferential attachment in citation dynamics is the obvious explanation. I believe that's a real effect, and there's certainly evidence such preferential attachment does occur. At the same time, it seems surprising if preferential attachment is the dominant effect. More fundamentally: some papers are genuinely far more important than others, that this is due to facts about the (actual) structure of knowledge, not merely the social processes we use to construct knowledge, and this is what drives (massive) differences in citation. But I don't understand why knowledge should have that structure. Is it some fact about the representations we use for knowledge? To be clear, some of the reason is certainly social – e.g., as mentioned, preferential attachment dynamics. But it may also be a kind of "temporal social compression". Consider the Schroedinger equation. it's very short and very important. But part of the reason it's compact is because people have previously realized the importance of certain ideas – like mathematical equality or derivatives – and decided they should have compressed representations. Still, I do not believe this is the entire effect. The world is shockingly well ordered, and it's a priori surprising that there are laws of considerable generality. To put it in terms of algorithmic information theory: we should – post hoc anthropic arguments aside – a priori expect no such compression to be possible at all. That it is is something of a miracle – the miracle of why the world may be understood at all.↩︎
Of course, one can do both. But the idea of the practice is, I think, important.↩︎
There are, of course, standard solutions as well. I knew one physicist who maintained a self-righteous fury at people trying to lose weight, announcing surprisingly frequently and contemptuously that all such people had to do was: "consume fewer calories than they burn!" Well, yes. This is much like telling a basketball player to "just shoot the ball through the basket", or a 100 meter runner to "just get to the end of the track faster than everyone else". True enough, but does rather misunderstand the underlying problems. I can give you a very specific example to illustrate the point: for some years, I struggled to keep my weight down. I presumed I lacked the discipline to consume fewer calories. Turns out: I had a mild autoimmune condition; once it was identified and treated, I "magically" lost most of the excess weight, about 10 kilograms or so. The condition had previously made me extremely sore while exercising, and messed with my energy in other ways (poor sleep, immune system constantly revved to 11) that affected my diet. "Consume fewer calories" didn't address root causes; once the root cause was identified and addressed, losing weight was easy. I don't, of course, mean to imply that others will have the same experience, it's merely an illustrative example.↩︎
This is only in part what's going on – there are other forces at work, too. If a modern listicle-writer had Michelangelo and Bach and Tolstoy riding artistic point for them, they'd end up looking pretty good.↩︎
This often occurs with things other than lying, of course. I was once working on a research project where I was being asked to rapidly re-explain the project, several times a week, to a group with little relevant background. I tried to put it in terms which made sense to them, but there was insufficient shared background and time for me to make much progress. After this went on for several months I realized to my shock that it had (badly!) damaged my own understanding of what I was trying to do. It's because the core creative ideas were vague, half-felt whisps, which I was trying to slowly develop, while my frequent explanations were (by contrast) very concrete and definite; the latter was overwhelming the former in my own understanding, but all the real core creative content was in the half-felt whisps. I ceased attending the meetings, and the project gradually got back on track.↩︎
I wonder if this is a case of: "Debugging considered helpful". That is, if debugging is an important part of how we engage with a set of ideas, then maybe those ideas actual benefit from having some small errors in them.↩︎
And religious thought. I'll focus on moral philosophy here, since I feel it more acutely for moral philosophy, for reasons I don't entirely understand! But similar remarks apply to religious thought.↩︎
That's a big part of why I'm writing these notes.↩︎
If you don't like overthinking, you're on the wrong website.↩︎
Certainly not infallible. It makes you very dependent on your culture. As Hannah Arendt points out, one of the most horrifying things about a criminal like Adolph Eichmann is just how normal he was. In a healthy culture, Eichmann-as-a-boy might even have developed into a useful contributing member of society. In Nazi Germany he was a monster.↩︎
These ideas have certainly been stimulated in part by David Chapman's ideas about systems (I don't intend to imply Chapman would agree – but I certainly wouldn't have had this line of thought without exposure to his ideas.). A brief summary (and some further pointers) may be found here.↩︎
Many of the "clever" bits of long-termism seem like this to me. In most cases they seem like very clever people using their giant brains to make and defend arguments which are challenging to see through, but whose real motivation is emotional. If you convince yourself that your work is contributing a 0.0001% improvement to the future of humanity, and there are 10 to the 30th potential future humans, then you're justified in giving yourself a billion dollar raise, if that shifts the amount to 0.00011%. The example is artificial, but I've heard arguments apparently seriously put forward that are closely analogous. I've found it personally interesting to debug these arguments, but almost useless to debate with long-termists: the essential point is that the motivation appears to be emotional, not rational, and the rational argumentation is a mere veneer.↩︎
This is, incidentally, also one reason science has been so much more successful than philosophy, and why getting drawn into philosophical debates is so often a waste of time. People who want to have such debates will say "prove me wrong!" But the way one wins such a debate is (usually) to not play. The puzzle-solver's response to getting tied up in moral conundrums like the trolley problem and "the repugnant conclusion" is to try to solve the problems. I think a better response may be to wonder if maybe you did something wrong in the first place. Nothing says morality should be a logical or consistent system at all. I suspect you learn far more by volunteering to help other people. Moral experience is so much more valuable than moral theories!↩︎
I'm far from sure of myself here. Ask them to tell a joke or write a script or construct a mashup scenario and the systems can do extremely well (with some work), and provide "novelty" by any reasonable measure. But if you ask venerable questions, the answers tend to be condensations of conventional wisdom, not novel and insightful. To get things which are novel and insightful usually seems to require injecting some seed for that insight into the prompt, as things presently stand. Perhaps what I'm wondering is whether it's possible to build a "bias for novel, deep insight" into such systems.↩︎
There are, however, frequently reminders to be grateful for any lofty station one attains in life. E.g., practices of fasting, the need for charity, and so on, all speak to this. But these practices are still fundamentally sympathetic to power.↩︎
I dislike the term "descriptive fact". It's difficult to make unambiguous sense of the term "fact" at all. But I just mean something (approximately, plausibly) true about the world, not containing any direct assertion about how one "should" behave.↩︎
One is reminded of the relationship between behaviorism and theories of psychology based more upon internal states than external behavior.↩︎
And to be loved by people. So many people in the world feel loved by no-one (even when they are). One of the other profound ideas of Christianity is "Jesus loves you".↩︎
This may seem a selfish framing justifying an (ideally) selfless act. But I mean "us" in both an individual and collective sense.↩︎
The connection perhaps occurred to me in part because a former friend at Caltech once drily observed to me that at Caltech, Feynman and God sometimes seem to be interchangeable.↩︎
I haven't worked fulltime as a theoretical physicist in 16 years; perhaps I should say "was trained a theoretical physicist". Still, deep habits of thought die hard, and theoretical physics left a deep imprint upon me. In many respects I feel more like a theoretical physicist than anything else, in habits of mind, even though those habits are now applied to other subject matter.↩︎
This is not quite true: the fourth of the Four Noble Truths, the Eightfold Way, in part addresses these questions.↩︎
Sometimes very ineffectively. (As I write I am very preoccupied with watching EA wrestle with racism. It is revealing some very deep issues with EA.)↩︎
Michael Nielsen, Notes on Effective Altruism (2022).↩︎
A standard EA response to this is: "Oh, how could we do better?" But that is a project of many tens of thousands of hours, and not something I wish to pursue. Indeed, I cannot resist one small critique: what EA considers good seems mostly determined by internal politics and power and community standing and sheer repetition, and only secondarily by the quality of a person's evidence. This is a set of structural factors that actually makes it extremely undesirable to participate. It's bad if one has no standing within the community (you are not heard); it is perhaps even worse if you have high standing within the community, especially if you have a lot of power (your bad ideas will be heard and acted upon). I see enough of both patterns to be extremely skeptical.↩︎
This doesn't necessarily mean practicing them. It may mean rejecting them, iterating upon them, replacing them; or, perhaps, coming to deeply accept and understand and practice them.↩︎
One feels hubris in this exercise if it's understood as being instructive of others. But if viewed as being primarily for oneself, it's not so much hubris as an essential part of living well.↩︎
It's interesting the extent to which this statement depends on surrounding context. As Adam Smith and others have so rightly pointed out, with the right institutions establishing lots of positive-sum games, selfish action can help the community a great deal. This is reflected later in the list.↩︎
I have this problem with stoicism too. Life isn't about dulling oneself, and feeling nothing, which sometimes seems to be the message of both Buddhism and Stoicism. Incidentally, in Huston Smith's "Religions of the World", Smith describes attachment-as-suffering as an essentially negative view, and asks if there is a more positive framing. As I understand, his solution is the view of the enlarged the self. You may be attached to getting that promotion at work. And if someone else is promoted instead, that may cause you suffering. But if you have a sufficiently large view of self, you will feel joy at that outcome too. It's an intriguing idea, and raises many questions. Perhaps most of all: to what extent are we free to (and is it healthy to) set the boundaries of our self?↩︎
I find it retrospectively rather humorous that I didn't realize this consciously in advance.↩︎
Perhaps the underlying unity is that of yin and yang.↩︎