Overcoming Bias

Having been an academic researcher for over forty years now, I am well aware of academia’s many big failures as an intellectual system. By comparison, our system of public intellectuals has many advantages, including that it doesn’t try to be boring or hard to understand, and that it is willing to take on topics that ordinary people care about, even when those are out of academia fashion, or poor topics for demonstrating academic impressiveness. However, the system of public intellectuals has one huge failing, a failing so big that it threatens to cancel all its other advantages.

Academics have a key “situate” norm, which says that a paper should situate itself within a prior literature. That is, it should cite not only the actual sources which influenced it, but also the closest prior work in the same topic area; the author should have read and been influenced by those. A paper should also fairly explain its relation to these other papers, and respond to their relevant points on the theses of this paper.

The related norm of public intellectuals is instead to mention any other high profile public intellectuals who have discussed a topic lately. Any other sources or similar writings can be ignored. Yes, the academic situate norm has substantial costs, and we often fail to follow it into other disciplines or low prestige sources. But we follow it far more than do public intellectuals.

This academic norm to situate greatly encourages the accumulation of insight over time. Combined with the norm of novelty, it pushes academics to explain how our each paper adds to and extends the sum total of what we knew before. Public intellectuals, in contrast, can and do regularly repeat what others have said many times before. They fail to create a division of labor so that humanity can coordinate to more efficiently explore the vast space of possible topics.

The obvious solution here is to create or strengthen a situate norm among public intellectuals. Yes, ordinary readers couldn’t easily enforce such a norm, but public intellectuals already follow many norms not now enforced by typical readers. For example, editors of journalists usually enforce norms of spelling, grammar, and non-false quotes that typical readers can’t easily enforce.

If the many words required to situate an essay would detract too much from the flow of that essay, such words might perhaps be included in an available appendix or aside. If this is too big an overhead for short essays (e.g., X posts), we might excuse those.

Situate: I looked for but did not find someone who make this point before. Maybe Richard Posner says something similar in Public Intellectuals: A Study of Decline.

Added 13Feb: The fall of academia includes the decline of theory and efforts to find new abstractions, the rise of public intellectuals, the fragmentation of disciplines, and the rise of the prestige level below which academics feel free to ignore previous work.

Many have long described variation in political opinion in terms of two key dimensions, described either as economic freedom and personal freedom, or via rotated axes as left vs right and libertarian vs authoritarian. 2.5 years ago I posted on The Myth Of Left And Right, a book correctly arguing that while abstract thinkers have often proposed coherent concepts of what could be meant by left vs right, and political parties who are out of power sometimes embrace such concepts, political parties that hold substantial power use such labels with far less coherence. As a result, the positions widely seen as left or right have varied so randomly across space and time that the labels “left” and “right” offer little predictive power for the positions of future parties in other times and places.

It occurs to me that a similar situation likely also holds for the other libertarian vs authoritarian distinction. While abstract thinkers have often proposed coherent concepts of what this could mean, and while political parties out of power sometimes embrace such concepts, political parties that hold substantial power probably have positions that are far less predictable or coherent in terms of such abstract ideological concepts.

The modern world often sees itself as “liberal” and “tolerant” in the sense that some choices and areas of life are considered out of bounds for governments, firms, and widely shared social norms to intrude on, while other more “foundational” issues are seen as more appropriate for such collective choices. Initially it was just particulars of Christian religion that were seen as out of bounds, but over time freedoms in other areas have been variously seen as central to the “liberal” concept.

However, for the actual political parties in power and their policies, we have seen wide variations which which are the choices and areas for which we are to be tolerant, and which are the other areas where it is okay for a majority to impose its views and wills on minorities. We should expect to continue to see such variations, re both of the key dimensions of variation in political opinion. Abstract thinkers will continue to paint their clean pictures of principles that could organize policy positions, while people actually in power will more opportunistically do what they want, sometimes justifying those choices in terms of such abstract principles as convenient.

You like some people, food, and TV shows more than you do others. When those likes change over time, you can often notice changes in context, i.e., in the details of those things and your interactions with them. Allowing you to attribute your choice changes to such context. You might see that you tend to like people who are friendly, food that is spicy, or TV shows that are sexy. You conclude that you don’t just directly like particular people, food, or shows, but instead your preferences are about more deeper features, and you make particular choices to better obtain those fundamentals.

However, for other things that you choose, you made your first related choice early in life, and haven’t changed your choices much in a long time. Like being really into music. For example, you may have forgotten that you initially weren’t so into music, but then got into music more soon after associates praised and noticed you more when you were musical. Yes, today you get a lot of respect and attention for your music, but you tell yourself that those are incidental; music is now one of your fundamental values.

If all that attention were to go away, your interest would plausibly eventually decline greatly. But as the chance of that is low, you can easily deny or ignore such undermining counterfactuals. And maybe you have worked to “commit” to this choice, by making it harder to imagine ever liking music any less than you now do. Which would in fact slow such a decline, if it were to come.

The habits that you acquired early in life were also influenced by the community in which you were embedded. Had that community been different, those habits would also have been different. But you won’t notice that context dependence much unless you study how different communities vary in their habits.

Other than re timescale and visibility, I don’t see a fundamental difference here. Between the processes by which you make the kinds of choices where you can see yourself changing often, and so can attribute them to deeper values, and the choices you made long ago, and see as intrinsic to “who you are”, where you don’t notice that you would have made different choices had contexts been different. Yes, at each level we can distinguish influences that come from info at that level, and deeper influences. But all the levels make that distinction.

Yes, people like this idea that they have a “core value identity” which lets them see themselves as moral and connected to particular communities. But this makes them neglect and deny their actual past and potential future context-dependencies.

If we look as deeply as we can at the slowest and largest scale processes by which we change our choices, what we will find is evolution, both of DNA and of culture. Maladaptive choices are suppressed, while adaptive choices are enhanced. And so if there if anything that deserves the name of “your deepest value”, it is adaption. Whether you like the self-image of that statement or not, adaption is in fact what most deeply shapes your deepest personal choice processes, which shape your more shallow mental structures that you notice as the “values” that guide your most frequently changing particular choices.

Our universe is vast and mysterious. There is so much we don’t understand about what it holds, where it all came from, when or where advanced powers might have arise, and how they might travel. So I cannot confidently rule out vast powers out there somewhere, maybe even powers that are aware of and not entirely indifferent to humans. In this sense I am an agnostic, not an atheist.

I can’t even exclude the possibility that some such powers hang around our planet, keeping their existence mostly hidden, but weakly influencing the human trajectory through minor interventions that may add up to big changes in the long run. That is, after all, my best guess of what UFOs as aliens would imply.

But the common hypothesis that some of these powers listen to unspoken thoughts in human heads, and then tend to on average favorably change the local world around those humans to “answer their prayers”, that hypothesis I find far harder to swallow.

This practice of answering prayer seems to require far more knowledge and efforts than would be required to more strongly direct the overall human trajectory, which they apparently choose not to do. So what gains could result from all this extra effort?

By assumption, these powers could favorably change the world around those who pray, but instead tend to choose not to do so in the absence of appropriately “sincere” prayers. This gives advantages to humans who are more popular, and also to those who are richer, as it seems quite possible to pay money to induce more sincere prayers.

The act of prayer may cut stress in those who pray, make them more willing to cooperate, and give them joys of submission. But such gains seem also available if such people would just put similar faith into their local human powers. Which most humans in fact did through most of the farming era.

Such vast powers themselves could in principle just enjoy the praise and submission of humans, but then why not make themselves clearly known and get far more praise and submission? And why care so much about the opinions of such small creatures?

Yes, if you try hard enough you can probably come up with some scenario in which it all makes sense. But you will have to make a great many a priori unlikely assumptions to make all that work. As a result, I assign a very low prior to such scenarios.

In contrast, the idea of great powers who answer prayer seems quite likely to arise via superstitious wishful thinking, even if no such great powers existed. This seems to me a far more likely origin of this practice. Especially in light of the fact that randomized trials find no gains for people unaware that they are being prayed for.

(Yes there’s a vast literature on this, little of which have I read.)

We are parts of many social units. In those we see as “thick”, we are more okay being partisan and wanting other members to share many loyalties and cultural features with us. In “thin” units, we are instead more tolerant and tend to allow everyone who doesn’t violate basic norms.

Bryan Caplan has a new book You Have No Right To Your Culture, wherein he seems to me to say that we should see nations as relatively “thin” social units, and so be open to more immigration into them. Which made me curious about which kinds of units we see as how thick, and what unit features predict this view.

So I listed nine types of social units: families, clubs, firms, professions, churches, neighborhoods, cities, nations, and world. And I came up with three plausible factors that might predict thickness:

• Power - How much power a social unit has over its members.

• Competition - How easily might this unit be killed by competition. Which correlates with how many alternative units compete with each one, with how easy it is to leave the unit, and with voluntary entry to the unit. This tends to be stronger for smaller scale units.

• Dependence - How much unit members depend on each other, and so have externalities due to other member choices. Which creates more coordination gains within such units. Which induces such units to manage a wider range of aspects of member lives. Which results in members gaining deeper identities from such units, seeing these units as more sacred, discouraging internal diversity, and encouraging equal treatment.

To estimate which factors matter more, I asked six LLMs to give 0-10 scores for each of these nine units re levels of thickness, power, competition, and dependence, and then to do a regression estimating thickness from the three factors.

The next table shows all their coefficients, together with the median across all the LLMs. I also asked for the predicted thickness for nations from their regression model, which is the last row. (The median of those happens to equal the median of nation thickness scores.)

The medians tell a simple story: social unit thickness, i.e.., how okay we are with requiring unit members to share loyalties and cultures, as opposed to tolerating differences, depends little on how much power those units have over us, nor on how much competition those units face, including how easy units are to leave or how voluntary to enter. Thickness is instead mainly seen as resulting from dependence, i.e., feeling that member outcomes depend a lot on choices made by other members. Nations are seen as ~2/3 toward thick on a thin-to-thick scale.

So people wanting to be more careful than tolerant re nation immigrants seems quite predictable, given that people think that nation member outcomes often depend a lot on what other nation members do. Which suggests three ways to change their minds:

1. Show them that nation member outcomes do not in fact depend so much on what other nation members do.

2. Convince them to set unit thickness levels on something other than how much unit member outcomes depend on other member actions.

3. Convince them that the immigrants they worry about are in fact likely to take actions that will give them good outcomes.

Our dominant world monoculture seems to be drifting into maladaption, and if we don’t fix this it will have much less influence over the future. When natural selection controls future change, one must package features one values with adaptive packages if one wants a hope of their lasting into that future. And I want some of our unusual monoculture features, like open abstract inquiry, to last.

One big cause of our modern cultural drift into maladaption is: a weakening of selection pressures. Compared to centuries ago, cultures (as units) today just don’t die much due to wars, famines, or pandemics. But one “selection” pressure remains strong: people want to join and copy cultures that seem to be “winning”, to gain the respect that comes from being and associating with “winners”.

So maybe if we could get more of us to see (natural selection type) adaption as “winning”, our behaviors might get more adaptive faster. Yes, one big obstacle is that after WWII many (somewhat falsely) blamed “social Darwinism” for Nazism, creating a taboo against trying to consciously win at adaption. But even if we can overcome that, a perhaps bigger obstacle is that we find it hard to see clearly who is winning in this sense, or which policies would help win, and so find it hard to encourage our leaders to help us so win. (Nazi strategies worked out quite badly, for example.) Thus even if we could raise the status of adaption, people might instead pursue other ends while giving lip service to adaption.

I have a solution idea: use speculative markets to create clear well-informed estimates of the future adaptive success of groups who could plausible coordinate to adapt better. That is, create “adaption futures.” With visible, focal, respected market estimates of group adaption, groups could reward leaders who improved their adaptation, or use conditional adaption estimates to pick policies that would so improve. And if we could get down to N=1 “groups”, we might even advise personal decisions re adaption.

To make this work, the main thing we need is good-enough ex-post measures of group adaptive success. With that, we could create assets which pay monotonically in such measures, and subsidize trading in such assets. After which the market prices of such assets can give our desired clear estimates of group adaption. And prices in conditional-called-off adaption markets could give us estimates of which policies promote group adaption.

Now if all we wanted was a measure of DNA adaption success, we’d only need to count the number of descendants of each group perhaps a century or two later. For example, we might measure the DNA of later folks and compare those to DNA measures of current folk. Or maybe just use records of who parents whom. We might also want to weigh final descendant counts by estimates of the future adaptive success of those later descendants, such as from their health, wealth, or status. These could give groups clear signals of how well a group did in the adaption competition.

However, once we realize that most natural selection of humans today is done via cultural evolution, not via DNA evolution, we will want a way to also count cultural descendants, not just DNA descendants. These are counts of how many people later inherited how much of the culture of a group. How can we measure that? Not only accurately, but also canonically, to avoid suspicions that those who define our measures, or who implement their measurement, might do so in ways that favor some groups or policies over others?

I suggest that we take detailed surveys of current cultural behaviors and markers, described in ways that we hope could also apply well to people in the future. Maybe hundreds or even thousands of diverse cultural features per person. And also surveys of who is how much in what groups today. Do such surveys now for a big random sample of people around the world, and also commit to later (say in a century or two) doing such a survey again, of people then. If we see the possible cultures of a person as a big vector space of points x, these surveys could allow us to estimate two normalized distributions over x, n(x) for now, and f(x) for the future. And for each group g, a normalized distribution g(x) of how that group is distributed over x today.

One simple approach would be to define cultural adaptive success per x as a(x) = f(x)/n(x), i.e., how many future people there are at each “culture” x, per each current person with that culture. Then the average success of group g could be the group average of a(x) over x, i.e., a_g = Int_x g(x) a(x) dx. Group market assets m_g might pay according to some monotonic transform of a_g, such as m_g = m(a_g) = ln(a_g). A more advanced sort of combinatorial market might even produce consensus functions that estimate a(x) for all x, not just for particular groups g. That actually seems technically feasible to me.

This approach assumes that the main way that cultural behaviors x change over time is that the people now at x create and influence more future people to be at x later. But what if there are also ways that behaviors x tend to change over time due to internal processes, or due to ways that behaviors depend on changed shared external factors, such as world wealth, peace, health, or tech? If so, the above approach would credit people at x that happen to be toward where the world moves with unusual cultural influence, when in fact such folks needn’t of had much influence at all.

To deal with this, I suggest that we also collect whatever data we can on such systematic cultural change processes, and estimate a change function c(x) that says how points x today tend to change into points y = c(x) later, independent of cultural influence from others, and then estimate cultural adaptive success via a(x) = |Det Dc(x)| f(c(x))/n(x). That first term with a matrix determinant of derivatives corrects for x volume changes due to a complex change function c(x). But keep the analysis to find c(x) simple, so we can stay canonical; when in doubt, attribute change to f/n, not c.

Okay, that’s the basic idea. Now let’s consider some pesky details. We need to make this adaption measure a(x) canonical, so as to avoid suspicions of biases re groups or policies. This is a key problem and I don’t claim to have solved it. But here is a plausible start. First, include as many cultural behavior variables as people are willing to pay to measure. Transform such variables to be more normally distributed, then use mean-zero unit-variance transforms of those. If stat and market methods prefer low dimensionality, do a factor analysis of all these culture variables and focus on the largest factors. Then use standard normality-based stat methods to build best fit models of g(x), n(x), f(x) from the collected data.

As n(x), f(x) are normalized, the above method only estimate relative cultural success. But they can easily be combined with any other estimates, market or otherwise, re the overall future success of humanity. If AI descendants later matter re future success of groups today, then cultures of such future AIs can be included in the later surveys.

As discussed above re estimates of DNA adaptive success, we might improve cultural success estimates via estimates of adaptive success of the population after the time of the future survey, using parameters like individual wealth or status. We might even commit to making a new set of markets then, and use their prices for such estimates.

There might be big disputes re the relative weighting of key culture factors, or of culture vs DNA vs org influence. In this case it seems okay if we just estimate up to a dozen different dimensions of adaption separately. People can then use different functions to combine these into their preferred adaption estimates. There could still be strong incentives to increase these dozen measures of “winning”.

Markets might find it easier to use assets with bounded asset values, obtained via using payoff functions m(a) that are also bounded.

To pay for all this, group representatives might pay survey owners for the right to create and trade assets based on group adaption estimates a_g from the surveys. Such representatives might also pay for market maker subsidies to make those markets more informative, and also for conditional markets to advise particular group decisions. Other funding might come from philanthropy.

You might worry about the long time scales. But stock prices can reveal expectations about commerce issues decades into the future even if individual traders don't hold each stock for more than a year. They just have to expect to sell to other traders with similar expectations.

A strong criticism of this proposal is that firms with a market price do in fact have an estimate of their long term adaptiveness, but this doesn’t stop cultural decay from consistently killing them, even when CEOs have great incentives, powers, and knowledge re preventing this. So without enough macro culture variety, markets estimating the adaptive success of each one might still not be enough to prevent maladaptive macro culture drift.

But let’s turn this critique around: working to prevent firm cultural decay is quite economically valuable in its own right, and succeeding more at that should raise our hopes for also succeeding with analogous efforts for macro cultures.

In our standard model of natural selection, organisms encode stable behavioral strategies, and pass those encodings on to their kids, only some of whom manage to make more kids, causing a drift over time toward strategies that tend to promote successful reproduction. In our standard model of decision theory, agents have fixed preferences, start with “prior” beliefs, update those beliefs based on info observed, and then pick the actions that max expected preferences.

Culture is humanity’s superpower, and in our best models of culture, humans combine these two approaches. DNA encodes brains that act much like standard decision theory agents for decisions of small to modest scale and scope. Larger decisions are handled by agent preferences and priors, which are encoded in culture. That culture is given in childhood by parents and teachers, but such transfer also continues through our lives. For example, we continue to assimilate to the culture of our elites as they change, to cultures of those who conqueror our places, and to cultures of cities, firms, clubs, and families to which we choose to expose ourselves. For example, over the last few centuries much of the world copied a great many features of successful rich West cultures, including their Christian religion.

I’ve continued to ponder how best to combine deliberate decision strategies with cultural inheritance. And in this post I want to prod such thoughts by focusing on an especially dramatic case:

Consider someone who, like me, now expects descendants of today’s Amish, Haredim, and other insular fertile fundamentalist religious cultures to, in a few centuries, “win” by becoming much more culturally influential than descendants of today’s dominant world monoculture. Such a person might today plausibly try to respect those future winning cultures similar to how they’d respect a culture that had recently conquered their place. So they might try to make themselves open to assimilating into that future winning culture, such as by believing in the Judeo-Christian God. The reasoning is similar; in both cases a “winning” culture has shown substantial evidence of its adaptive superiority.

The general idea is that if natural selection is going to continue, and if you want to influence the longer-term future, you will have to find a way to combine the features you love with other adaptive features, to create a package with a better chance of success, to give your loved features their best chance to survive and thrive.

The reason I expect the Amish, etc. to win is that they have grown fast and maintained insularity for over a century, and survived many big change challenges in that time, while the leaders of our decaying world monoculture have far less incentive, knowledge, and power to change that culture, compared to CEOs re firm cultures, yet such CEOs consistently fail to stop firm cultures from decaying and killing firms.

Nine counter arguments:

1) What if I don’t care about influencing our long term future?

Then you are excused. But do expect people with your attitude here, and those with correlated features, to decline over time in the future.

2) Natural selection should not encourage a culture to have members promote the death of that culture, compared to others.

Cultural assimilation usually isn’t all or nothing; you retain something of your origin. Sometimes the best way to promote your culture is to merge a part of it into another more adaptive culture. While you can’t save all of your culture this way, this might still be your best shot.

3) We often try to resist, not assimilate into, a conquering culture.

Yes, when we think there’s a decent chance such resistance could succeed, getting our entire culture back seems better than having a modest influence over an invading culture. But when the chance of successful resistance falls too low, abject submission seems a better strategy.

4) Our current habits are largely of copying cultures that have recently been clearly successful, not ones that seem likely to succeed in the far future.

But the logic of copying success doesn’t care when exactly the success will be achieved, only when such success becomes sufficiently clear.

5) We can’t actually choose to believe something just because we think it would be good to believe it.

Yes, beliefs aren’t simple dials to turn in our head. But we can deliberately change many influential aspects of the contexts of our belief changes. Otherwise there would be little point to the vast literature on the rationality of beliefs.

6) But don’t we need culture to evaluate which cultures “win”?

Sure, cultures tell you which virtues to count how much in estimating cultural success, and that may influence your estimate of a culture’s adaptive success. But in most cases, including this case, that doesn’t change the answer much.

7) There’s no particular evidence that the Judeo-Christian religions of those societies is what would make them win.

We usually don’t know that much about which particular cultural features are more responsible for a culture’s success. Which is why we evolved the habit of copying culture packages wholesale.

8) How could the adaptive success of a culture count as evidence that its religion is true?

It seems that on average, all else equal, cultures that believe more true things are more likely to succeed. To bet otherwise, you’d need particular evidence that this particular claim is an exception to this general trend.

9) But if the correlation between a cultural feature and cultural success is low, success is only weak evidence re that feature.

Yes, but as we typically have great uncertainty over the future adaptiveness of cultural features, usually most of our evidence is weak. Nevertheless, if we care enough about adaptiveness, then even weak evidence will be enough to tip our actions in the direction of our best clues so far, even if those clues remain weak.

To study perceptions of causes of cultural change, I started with this posted list of the top 240 novels ever. I then asked (paid versions of) three LLMs to, for each novel, see if a main character is shown to have a stance of support or opposition to some social movement, and if so to pick the most central character like this. Re this set of novel characters, I ask what %-political (vs cultural) was the movement, how the character’s stance re it changed in the novel, and to pick from 8 possible causes of change.

Out of the 240 novels, ChatGPT found 9, Gemini found 35, and Claude found 99 where characters took a stance re a social movement. Of these, 5, 15, and 61, respectively, characters changed their stance. So the LLMs had rather different standards re how to decide those.

Their median movement-%-politics estimate was 85%, 80%, and 58%, respectively. For novels where the character changed their stance, those medians were 90%, 100% and 54%. The fraction of novels with a movement where it was <20% political is 0%, 3%, and 6%. That 3% is one novel, Don Quixote, while the 6% are On the Road, The Age of Innocence, Confessions, Brideshead Revisited, The End of the Affair, and The House of Mirth.

The fraction of characters whose stance changed who came more to support their movement was 80%, 40%, 34%. And “seeing unexpected events or facts in the world” was said to cause their stance change in 100%, 77%, and 82% of cases where a cause was identified. The next most common cause was “Change resolved inconsistency in prior norms”, at 0%, 15%, 7%. Here are the other 4 possible specific causes that LLMs rarely thought described novels:

• Saw opportunity to gain power, status, attention

• Saw prior associate or prestigious model change, copied them,

• Gained new associates or prestigious models, copied them

• It just felt right

Thus top novels are overwhelmingly focused on political, not cultural, change, and on change driven by characters seeing unexpected things in the world, and not driven by feelings, consistency, or copying associates. Typical real process of cultural change seem largely invisible to top novel authors.

I’ve read a bit about cultural change both in corporations and in our larger macro cultures. And one thing I’ve noticed is that firm culture change less often involves youth movements. Yes, when there are larger cultural youth movements, that does influence behavior in firms. But we don’t so much see youth movements particular to specific firms. Why?

I see four main explanations. The first is that most corporations just don’t last very long. As business change is often enacted via old firms falling and new firms rising, there is less need for youth to visibly fight for change there. Youths can instead just switch to other firms, or wait for new firms to arise. Macro cultures, in contrast, will only change if folks push for change, and youths can be more sure to win eventually if they just wait til the old are gone.

A second explanation is that, compared to larger societies, the hierarchical nature of a firm more structures its communication and choices. So young people with ideas for change tend to privately persuade supervisors to adopt them, supervisors who then become the face of such changes. Both because youths have more access to firm leaders, and because they can face stronger retribution for visibly opposing leaders. In macro cultures, it is harder for youths to meet and persuade leaders to support changes. Plausibly firm leaders, compared to macro leaders, have stronger incentives to adopt changes even when they are hard or threaten prior leader investments.

A third explanation is that different cultural units use somewhat different status markers. In a firm, it is easier to demonstrate concrete achievement influencing firm success, achievement reflected in hierarchical position. We are less sure who contributed how much to a macro culture’s success, less confident that rank there reflects achievement, and so are more willing to listen to people with more indirect markers of quality, like education, articulation, popularity, and energy.

A fourth explanation is that art, intellect, and morality abilities are more influential re macro culture changes, and in those areas we more have a myth of the genius who is visible early in life via their impressive public opposition to the old. So in those areas of life org leaders and supporting personnel take backseats to the foregrounded youthful geniuses, who become leaders of youth movements.

Each of these explanations suggests a corresponding approach to reducing the influence of youth movements in our macro cultures. We could make it easier to switch macro-cultures, give macro leaders better incentives and make their contributions clearer, and deconstruct the myth of genius.

Our culture—the values we hold, the norms we follow, the virtues we admire —shapes nearly every aspect of our lives. Cultural norms determine how we raise children, structure relationships, pursue careers, and find meaning. When these norms shift, the effects spread across billions of lives for generations. Yet the process by which we change our culture seems alarmingly inadequate to the task at hand.

Consider how cultural change actually happens. Many people notice what they see as problems with existing norms. Some discuss these with friends. But only a few—perhaps a handful per significant cultural shift—successfully articulate a compelling alternative that spreads widely, usually via youth movements. These “cultural entrepreneurs” write the essay, give the talk, or create the art that reframes how we think. They are the bottleneck through which nearly all cultural innovation must pass.

Who are these bottleneck people? They tend to be writers, academics, journalists, and creators—those with platforms and communication skills. They’re generally well-educated and economically secure enough to take reputational risks. They’re often in their late twenties through forties, young enough to feel friction with existing norms but old enough to be heard. They’re insiders enough to have credibility but outsiders enough to have something new to say.

The problem is not with these individuals per se, but with the staggering mismatch between the difficulty of their task and the resources they bring to it. Evaluating and redesigning cultural norms is an extraordinarily complex undertaking. It requires understanding how a proposed change will interact with dozens of other cultural elements, how it will affect diverse populations in varied contexts, and what second and third-order consequences might emerge over at least decades, maybe even centuries. It demands knowledge of history (why do current norms exist?), cross-cultural perspective (how have others solved this?), and systems thinking (what are the hidden dependencies?).

Yet the typical cultural entrepreneur spends perhaps a few hundred hours thinking about their proposal. They write an essay, maybe a book. They discuss ideas with friends who share similar backgrounds and perspectives. There’s rarely systematic investigation of how the proposed norm change worked in other contexts, rigorous consideration of edge cases, or serious engagement with the strongest counterarguments. The thinking that goes into reshaping cultural norms affecting millions is often less thorough than what a committee might spend redesigning a corporate office.

This wouldn’t necessarily be problematic if cultural entrepreneurs were selected for wisdom, breadth of knowledge, or careful systems thinking. But they’re not. They’re selected for persuasiveness—the ability to craft compelling narratives that resonate emotionally. They succeed by coining memorable phrases and pointing to patterns that “click” for readers, not by rigorous analysis.

Their reasoning tends to be analogical rather than logical, working through evocative examples rather than formal arguments. It is more intuitive than academic, more humanities than STEM. Terms stay deliberately vague to maximize applicability. It is anchored more strongly in strong moral intuitions than in fundamental issues of cultural adaptiveness. And while all this is what allows ideas to spread, it means success in the cultural marketplace correlates weakly with validity.

We’ve seen this play out in recent history. Advocates of “following your passion” in career choice articulate real problems with soul-crushing conformity but neglect how this interacts with credential inflation and student debt. Proponents of radical authenticity identify genuine costs of self-monitoring but underestimate the exhaustion of constant emotional exposure. Champions of optimization and efficiency capture real waste but miss the essential functions of slack and redundancy. In each case, articulate people identify legitimate problems and solutions that resonate—but lack the time, tools, or perspective to adequately map the full terrain.

This creates a disturbing dynamic: cultural change is both quite consequential and greatly under-theorized. We invest more careful analysis in designing a smartphone than in redesigning norms that will shape how millions of people live. The bottleneck people are doing only a small fraction of the thinking that the task actually requires, yet their proposals can reshape society within a generation or two.

What makes this especially concerning is the absence of local feedback mechanisms. By the time we can clearly see the consequences of a cultural shift, the change has already diffused widely and become entrenched. There’s no recall process for bad cultural innovations, no systematic post-mortems, no institution charged with learning from mistakes.

We seem to face a structural problem in how human societies adapt. Our process for allowing cultural innovations to emerge and spread—relying on articulate individuals to notice problems and propose alternatives—may be fundamentally inadequate for the complexity of its task. Plausibly resulting in the problem that I’ve been highlighting for a few years now: maladaptive cultural drift.

The ancients suffered this problem less, as their societies were simpler, with lower rates of change of all sorts. And they could more rely on their much higher levels of cultural variety and societal selection pressures to filter out their worst mistakes.

Maybe cultural change is too important to leave to the happenstance of who can write a compelling essay. It seems that cultural conservatives, who have long warned against changing culture too much too fast, were roughly right. But alas, just doing what they suggest today can’t fix most of our past mistakes. What can we do?

It seems we must find a way to create more systematic analytic processes for evaluating proposed norm changes. I’ve suggested one. Until we do, we seem likely to suffer continued cultural decay, plausibly resulting in the decline and replacement of our dominant world civilization, following by a continuing rise and fall of civilizations.

In the last month, many who want to kill Polymarket have agreed on a common strategy: claim that Polymarket allows illegal “insider trading”. I’ve received roughly a dozen media inquiries about this in the last few weeks. E.g., some articles: 1, 2, 3, 4, 5.

This is the strategy: In 2011, the US CFTC issued Rule 180.1 prohibiting any commodity market trades accompanied by “deceptive” talk. Since then the CFTC has interpreted this rule as prohibiting trades based on info obtained via a breach of a duty of trust or confidentiality. Some recent profitable Polymarket trades have been accused (without evidence) of being based on such “inside info”. But Polymarket is a crypto system, where they cannot usually know who are their traders. QED: Polymarket must be shut down.

I see much misleading talk in my world, so a rule against deceptive talk relevant to trades seems to me to not actually be enforced much. And also quite hard to fairly enforce, and in any case not clearly needed. The US Supreme Court has struck down laws against lying on free speech grounds, unless such lies produce “material gain”. I’m not convinced we should distinguish material vs non-material gains re lies.

In any case, trading on info that you promised to keep secret does not remotely “deceive” markets re your or their trades. Furthermore, I don’t think it makes sense to generally assign to all our social institutions the task of preventing anyone from revealing secrets that they instead promised to keep. If you come to my store to buy a dress for your wife, and reveal to me her dress size, I don’t think it should be my job to check that it was okay with your wife for you to tell me her dress size. The two of you should be in charge of figuring out how to enforce your info promises to each other.

As a closer analogy, both journalism and speculative markets are info institutions, which reach out to collect info from the world, aggregate that info into useful summaries, and then spread those summaries into the world so that people can act on them. Both are private for-profit institutions, depending on voluntary participation.

We should treat all such info institutions the same re any duty to ensure that the info they collect doesn’t break secrecy promises. Are you willing to require that journalists either don’t talk to anyone who might know a promised secret (e.g., any govt officials), or keep records of who they talked to on what, available to law enforcement to check on suspicions that a news article revealed a secret? If not, we shouldn’t requite analogous rules of speculative markets either.

Yes, it can sometimes be hard for people and orgs to keep secrets. People who promise to keep secrets might instead tell them to others. But assigning to everyone the task of making sure that no one reveals promised secret seems way too hard, and also doesn’t give good incentives re how many secrets are worth keeping. Instead let people and orgs figure out how to use their voluntary powers to keep their secrets, via contract, reputation, relationships, or something else. When someone tells you something, it shouldn’t be your job to check if they are thereby revealing a secret.

Re civ decline, my basic story is that a key contribution to the fall of civilizations is plausibly that as a civ gets big, rich, and peaceful, its local cultural evolution process parameters get worse, at least at the culture level. There is typically lower variety, weaker selection pressures, and faster environmental change. So the process goes bad, its cultures drift into maladaption, and the civ falls. Our current civ will plausibly suffer this fate, with the added problem that due to modern cultural activism we also have higher rates of internal cultural drift.

What does this say about Christianity? Early on it was a small sect competing with many others, and the fact that it won against them suggests that it was unusually adaptive then, at least in that context. Then it took over the Roman Empire and most of Europe, and became securely in pace for millennia. And while securely in place, Christianity substantially changed its character many times. So doesn’t my theory suggest that those changes would on average have been maladaptive?

Well first notice that most Europeans didn’t know much about Christian doctrine re how to live ordinary peasant lives until about 1600 or so. Before then Christianity mainly influenced elites, cities, and larger institutions. Also, there was often lots of competition within Christianity; the cultural evolution problem would only be re cultural features that were imposed on everyone in Christianity, allowing little local deviation.

Okay, but Christianity influenced marriage much earlier, from about 1200, promoting monogamy and banning cousins marriage, and that suppressed family clans. Christianity also pushed individualism early on, via consent in marriage and the freedom to write wills, especially donations to the Church. And those wills funded many big monasteries, which slowly took over lots of European land. Also, conflicts between the church and crowns started early and weakly contributed to the lack of a single power taking over all of Europe. (Though note: most of these things didn’t actually change much over the history of Christianity.)

The protestant revolution created more competition among forms of Christianity, but it also induced record high level of religious hostility and destruction. But then the strangest thing happened: after the thirty years war (1648+), Europe suddenly agreed to great religious tolerance, at least among variations on Christianity, inducing world record low levels of religious destruction.

It turns out that religious tolerance, individualism, suppressing family clans, and preventing a single empire, were important enablers of modern capitalism, which enabled the Industrial Revolution. Which seems to have been quite adaptive in many ways, and least on the timescale of a few centuries. So does this show that some pro-adaption process was driving changes in Christianity over millennia? Or did it just get lucky?

I think Europe just got lucky. The accumulation of land by the church, and the increase in religious hostility, seem maladaptive, and suppressing family clans was also probably maladaptive at first. Then Europe got lucky in that religious hostility dramatically (and puzzlingly) reversed, crowns grabbed back most of that church land, and then individualism, suppressed family clans, and no central empire together turned out to be very good for capitalism and industry.

But maybe to get this lucky, Europe needed to make some big changes from prior cultures, and an out-of-control drifting Christian culture is part of what gave Europe the ability to make such big changes. Usually big random changes go badly, but they sometimes allow evolution to make big jumps to new peaks, in ways that wouldn’t be possible without them. Yes, this means maybe we today will also get lucky in a similar way. But don’t count on it.

One of the greatest obstacles to humanity using abstract thought to help on its biggest issues is this: our main professional specialists in abstract though applied to big issues are academics, and they refuse to think abstractly about how they allocate resources.

Academics are divided into an inherited structure of disciplines and subfields, a structure that changes only slowly. The focuses within each subfield fluctuate with changing fashions, the relative size of fields changes as outside funding changes, and sometimes funding allows the creation of new fields.

The (high-dimensional) space of all possible abstract topics, however, is only sparsely populated by this structure. Most of that population is in dense “urban” clusters, with few residents of the vast “rural” tracks outside of those clusters. Most of sources of resources and prestige for academics are tied to those clusters, which discourages them from venturing outside of those clusters.

Academics are so tied to their little “cities” that they have only vague concepts of most of the space outside of it. In fact, only ~2% of academics could give a coherent intelligible (but not necessarily correct) answer to this question: “Why is your particular research nearly the most cost-effective among the options available?” (This has long been my personal experience.)

Few academics can justify their research relative to anything besides other nearby research. The sort that might cite them, and which they might cite. The question of why that whole area is funded is of little interest to them.

A big factor that causes academics to clump strongly in topic space is that they strongly prefer to use the most prestigious methods available in each area. As research is mainly judged for its prestige-potential, not its social value. This heavily favors the topics that better allow one to show of one’s mastery of prestigious methods. Other topics are neglected.

The intrinsic value of topics in this vast space of possible topics is only weakly correlated with the density of academics near them. Which means that most of the important abstract topics are not near the center of clumps. And so even if someone does venture out into the sparsely populated “rural” topic areas, and finds a topic very important to humanity, other academics aren’t much interested in evaluating what they claim to have found. Most likely, there aren’t places to publish it, there aren’t programs to fund it, and there aren’t jobs to hold specialists in it. Why bother?

Yes, if you have gained enough prestige working in the center of an academic city, more academics will listen to you to what you’ve found outside of cities, and may even be willing to stretch the city boundaries in those directions. But the few qualifying academics here usually have many other more rewarding ways to spend their time.

Yes, some kinds of non-academics also specialize in more abstract reasoning. Such as journalists or managers. But when they claim to have found something important out in the academic wilderness areas, their lack of academic credentials is usually suggested as a reason to doubt them. After all, if wouldn’t the nearest academics had said something, if there really was something to be found there?

You might think that outsiders should hold academics accountable, but a major trend of the modern era has been for prestigious professions to wrest control from outsiders. Academics succeeded at this via promoting grants, tenure, and peer review. Today funders primarily buy prestige by association with academics who decide internally who is prestigious.

Most likely, our world civilization will peak in population in about three decades, and its economy will peak soon after, or maybe even before, depending on how bad Africa is at substituting for everyone else in the world economy. From then on rates of innovation will shrink faster than does the economy, and our civ will continue to fall and slow for centuries, until replaced by rising insular fertile religious subcultures (eg Amish, Haredim).

When future folks look back on this, the fall of the greatest civ so far, who will they blame? You, actually. In two polls I asked re a generic civ that rises and falls, people in which period are most to blame for that fall. In one poll I divided the civ’s history into 4 parts: start to rise, finish rise, start to fall, finish fall. In another, I divided it into 5 parts: flat turns to rise, mid rise, peak where rise turns to fall, mid fall, fall turns to flat.

For the 5 part split, 61% most blamed peak, with equal 19% going to each of the two adjacent periods. For the 4 part split, 52% say start fall, and 35% finish rise. Thus the most blame is given to those around just after the fall begins. And as most of my readers are below age 50, and about half of age 50 readers should be alive in 35 years, that’s you.

Maybe think about what you are doing to hinder our civ’s fall?

The public has apparently consistently disliked modernist artistic styles in many areas, including architecture:

Americans prefer traditional/classical buildings to modern ones by about 70% to 30% (regardless of political affiliation!). In a poll of America’s favorite architecture, 76% of buildings selected were traditional/classical. A study of courthouse architecture determined that … ‘most non-architects dislike ‘modern’ design and have done so for almost a century.’ Yet 92% of new federal government buildings are modern. (More)

Which raises the question of why such styles are so often chosen. The obvious proximate answer is that elites choose them to impress other elites, who tend to prefer specialist not popular evaluations, even when they represent democratic governments. But why do they do this?

Scott Alexander, author of the above quote, suggests as explanations: elites trying to hide not flaunt wealth, getting out of touch with popular tastes, becoming more Protestant, seeing timeless aesthetic truths, responding to rising labor costs, splitting off from popular tastes, or signaling taste instead of wealth. Samuel Hughes says:

Modernism was driven by artists & the intellectual circles that surrounded them. In the case of some arts, … music … literature, it never really left those circles: … In … other art forms, rich people were eventually drawn in to some degree. (more)

Which seems right, but raises the question of why artists diverged and how they were able to get away with it.

When search for explanations of social trends, it seems better if possible to focus on trends which one well understands in fundamental terms, so one needn’t ask what caused those trends. And one such key trend for understand the modern world is: the rise of professions.

Medieval guilds declined greatly in power after ~1600. Since ~1800 one’s profession has become a more central part of one’s status and identity, as religion, ethnic, and national identities have faded. The rise of work knowledge and specialization, and of living density, made it harder to pick professionals using a local network and reputation, and so people became more accepting of professions gaining more autonomy to control the quality of their members.

For example, from 1800 to 1950, academics wrested control from outsiders via replacing prices with grants, giving professors tenure, and using peer review for evaluations. Doctors gained the power to control malpractice liability, medical schools, and professional licensing. Judges prevented legislators from regulating the practice of law, so law controlled its own rules and licensing. Clergy have long had such autonomy, and engineers gained quite a lot.

The same modern attitude of indulgence toward professions seeking to judge and control their own quality also made people more willing to defer to artists on artistic quality. This supported a move from individual patrons to salons, galleries, museums listening to curators, critics, and journals.

So the move where artists successfully pushed modernism even when most customers didn’t care for it would just not have been possible a century or more before it happened. Of course this doesn’t explain the more precise timing and direction of the modernist change, which I’ve suggested is partly a response to recent social mixing of different cultures undermining tradition, the rise of abstraction encouraging that in art, and the rise of high school inducing the rise of activist youth movements.

Animal brains compute DNA-encoded behavioral strategies, using DNA chosen by simple natural selection. Such brains can “reason”, but not very abstractly, and mostly about how to execute their DNA-given instincts. Animals don’t reason about where such strategies came from.

In humans, DNA encodes sufficiently-sophisticated brains to let us copy others’ behaviors. As a result, in humans a cultural form of natural selection allows us to evolve behaviors far faster than do other animals. This is humanity’s superpower. It induces us to reason more, about how best to generalize from the behaviors we see, and also about which humans to copy which behaviors from.

The invention of language let humans more easily transfer many things between our brains, to do so more abstractly, and to reason more easily about it all. We humans thus became more ambitious about the scope and abstraction of our reasoning.

We humans eventually came to realize that we could try to reason critically about the behaviors which we have inherited via cultural evolution. But as natural selection didn’t document its code, we mostly made up plausible-sounding reasons for such behaviors. While some of us were sometimes willing to change behavior based on such reasoning, this typically went badly, as we didn’t understand why evolution had made its choices. Thankfully, most others didn’t much respect or heed such reasoning.

During the Industrial Revolution, humanity saw many big fast changes to tech and social practices, and initially tried to preserve prior cultural values, norms, and status markers in the face of such changes. But then around roughly 1900, elites suddenly began a big new “modernism” project to try to remake culture via abstract reason. Abstract reason was then high status, it had been credited for much of those tech and business changes, and was the central focus of the new schools that most everyone now attended into until adulthood, and which decided much about their status in life.

The cause of this sudden change seems to be a combination of a rise in mixing of folks from different cultures making their prior traditions seem less compelling, plus a rise in the autonomy of professions, here in particular of art/culture specialists. Young adults kept together in schools created youth cultures, which began youth movements to change culture, movements justified in terms of abstract reasons, and which consistently won against older opponents.

Though abstract reason has fallen somewhat in status since then, cultural change driven by young activists has continued at a rapid pace. Now if these youth activists had focused their reason on which practices, norms, and status markers could make our culture more adaptive, we might have stayed adaptive even in the face of rapid tech/social change. Alas, young activists did not much consider cultural adaptiveness in picking changes to advocate. They mainly based their abstract reasoning on strong moral intuitions, and were largely uninterested in reflecting on the cultural processes that created such intuitions.

As a result, fast cultural change largely uncorrelated with adaptiveness has been a key driver of our cultures drifting into maladaption. That and fast tech/social change, the huge fall in deep cultural variety due to increasing ease of travel, travel, and talk, and the fall in selection pressures due to increased wealth, peace, and health. This drift will plausibly cause our civilization to fall, and likely be replaced by now insular fertile religious cultures. More similar rises and falls may follow.

Our choices are stark. Either 1) do nothing and slowly go bust, 2) quit embracing abstract reason so much, returning to a world of tradition, ignorance, war, poverty, and strong selection pressures, or 3) find and adopt ways to go all-in on using abstract reason to choose adaptive cultural elements. Or maybe do this indirectly via adopting a very competent form of government (e.g., futarchy) and assigning to it a sacred goal that aligns with adaption over centuries (e.g., 1M living in space ASAP).

Alas going bust seems by far our most likely outcome. If so, our current industrial era will be remembered a unique dreamtime, with horror by descendants see us as having indulged a great excess of reason, and with a wistful sense of loss by descendants who try to recreate and improve on it many centuries later.

I have long posted on studies that try to make sense of how human values differ, and how they have changed over time. A new paper in Social Science Quarterly tries to compare four different proposed values models, by modeling the correlation network structure of the variables they see as central. This diagram shows their key result: three clumps of 2-4 named clusters, each which has 3-10 correlated variables:

Alas that paper doesn’t give names for the particular variables, and some of the clump names seem overly opaque. But what I can see is still interesting. The lower right clump (called “virtuous agency”) matters most to modern elites, while the middle clump (called “power and social order”) mattered more in the farming era. Did the upper right clump (called “hierarchical individualism”) matter most to foragers, and also matter more now due to a toward-forager modern trend?

I’ve suggested that many modern differences are due to moderns seeing themselves as higher status, due to their higher wealth. If so, that virtuous agency would also have mattered more to farming era elites, while the power and social order clump should matter more to non-elites today.

A small but real part of the variation in the policy positions taken by individuals and organizations is explained by two key factors: an econ and a (social/identity/) culture factor. Two diagrams show distributions of stated positions re these factors:

Some “stylized facts” to consider:

1. Each side (i.e., left or right) varies less on one factor more “central” to it than on the other: the left less on econ, the right less on culture. So there are far more (upper left of diagrams) “authoritarian” left econ and right culture folks than there are (lower right of diagrams) “libertarian” right econ and left culture folks. The median left, median right, and libertarian positions roughly form a triangle.

2. The more educated vary less on the less central factor. So the very educated are more polarized, and are more a 1D distribution, while others are more 2D.

3. Policy typically sits more in the middle of this space, compared to median opinion. So policy is more libertarian, and also closer to the median educated.

4. Over centuries, policy has moved toward the left on both axes, though it has moved more in culture than econ over the last half century, and moved more on econ than culture in the prior half century.

5. Compared to high level govt policy, govt policy details, and also individual behavior, sit even more toward the middle of this space. The right divorces more and gives more to charity, while the left conserves less energy and more blocks low income housing.

I’m tempted to explain these patterns this way:

1. In terms of the many details of policy, our implied positions on these axes, averaging over these details, are roughly distributed as independent Normals.

2. But to take a visible abstract position in that middle feels too conformist. Most want instead to look abstractly principled and different.

3. So most prefer to visibly identify with one of the two popular principled and different abstract positions: left econ or right culture. (Fewer take the also principled and different abstract libertarian position.)

4. The more educated more want to be abstractly principled and different, and also accept abstract arguments for correlated econ and culture positions.

5. Being less visible, most policy details, as well as detailed party positions on policy, are not very consistent with abstract positions, and so are pretty random and also more toward the middle of those independent Normal distributions.

6. Being even less visible and more random, most personal actions are on average even more toward that middle.

7. The arguments for the left econ position are more abstract than for right culture, which is why policy changed more on econ during the peak of abstraction, and but has more recently moved more on culture while abstraction was in decline.

Alas, this account suggests that the more abstractly described is futarchy’s outcome measure, the more likely it would be to be pulled toward “authoritarian” left econ and right culture positions, and the more it might might jump back and forth between left and right as different political coalitions took power. An outcome measure that was more a mess of details might be more stable and closer to detailed policy preferences, though alas that also seems more open to gaming and corruption.

Vitalik Buterin (VB) on culture:

Zuzalu in 2023 was an experiment: bring ~200 people, from multiple communities -Ethereum, longevity, rationalism, AI - together in one place for two months, and see what happens. … fall away over time: novel governance designs, and a search for legal autonomy. … Over time, popups would get shorter in duration, smaller in scope, and more generic in substance, … I have started advocating for Zuzalu-inspired communities to start having permanent nodes. … I always fear the “regression to the mean” that they will turn into glorified coworkingspaces, and lose all of their cultural or experimental interestingness. …

We have too much of a two-level structure: individuals, very powerful large-scale actors like states, and nothing else. … a successful “intermediate institution” … needs to be some kind of neo-tribe … that focuses, and meaningfully innovates, on the thing that humans do that isn’t generic: culture. … people make the mistake of thinking culture is something that can be explicitly laid down by mission statements and top-down edicts. …

Mistake of over-identifying culture with the purely aesthetic, subjective and group-identity-oriented parts of culture: food, music, dance, dress, architecture styles, and ignore the parts that are functional, whose success or failure drives the success and failure of civilizations. …

Many things … require “immersion” to succeed: lifestyle habits, local public goods such as air quality, work habits, lifetime learning habits, limitations on use of technology, etc. Doing anything truly interesting and unique requires “depth”, and substantial collective investment and effort to create an entire environment oriented around better serving those needs. These things cannot easily be done by an individual …

Culture is a big complicated blob where actions, consequences, statements by leaders and theories by intellectuals all influence each other in every direction. …

What we want is a better “world game” for cultural evolution: an environment where cultures improve and compete, but not on the basis of violent force, and also not exclusively on low-level forms of memetic fitness (eg. virality of individual posts on social media, moment-by-moment enjoyment and convenience), but rather on some kind of fair playing field that creates sufficient space to showcase the longer-term benefits that a thriving culture provides. …

Cultural innovation works better when it arises out of a collection of habits, attitudes and goals that are shared by a particular group, and adapted to the group’s needs. …

So far, I have told two disjoint stories. One is about smaller-scale community-driven projects, and experimentation in culture. Another is about larger-scale politics and business-driven projects, and experimentation in rules.

… I predict that in general the “market structure” will split [these] tribes and zones into distinct categories, because these are different things that require different specialties that are complementary …

People just have to get off their butts and actually create these alternative cultures and environments, and doing it is hard. Startups are also hard. But startups have had a multi-billion-dollar capitalist optimization machine figuring out all the most optimized ways of doing them and rapidly growing them to scale, and turned them into cookie-cutter standardized playbooks. Culture does not have the same profit motive, and culture is inherently not easy to scale. …

I do not literally expect we are going to see a world where most people live in tribes, or even zones. … But I do expect a world that is somewhat more dynamic in both economic and political rules and in cultural dimensions, and that gives people more options. Such a world would be a world where (i) people have more meaningful freedom, both to escape persecution and to choose the kinds of environments that they truly enjoy living in, (ii) we get better innovation both in economic and political rules and in culture, and (iii) instead of the innovation and creativity of the world being concentrated in a few super-centers of global economic and political power, it is globally distributed everywhere across the world. This is a world that I want to live in.

It seems that VB has learned much about culture. But I fear that he, and most of you, still don’t quite get just how severe is our culture problem.

Some aspects of culture, like clothes and food, let people feel enjoy a distinct cultural identity, but are shallow, and don’t much influence biological adaptiveness. Other aspects are deep, like those that set our attitudes and behaviors to family, fertility, death, war, community, and democracy.

Each aspect of culture has a (context-dependent) scale above which it can easily vary, and below which it cannot. Aspects with low scales allow for much variety and experimentation, and thus effective selection, but aspects with high scales allow for far less, and this is where we have a big problem. Deep aspects tend to have higher scales.

When aspects vary one at a time, then to win, innovations must be attractive given the usual distribution over all the other aspects. But when aspects can be varied as packages, parts can win that would not seem attractive by themselves. This is like the evolution of species compared to organisms, or of firm cultures compared to within-firm innovations. These species- and firm- level innovation processes actually matter more than do within-species and within-firm level innovation processes, as we have seen more innovation in fragmented habits and industries.

When we try to innovate in language, tech, business practice, law, and governance, we can often A) see which factors were held fixed while others changed, B) distinguish cause from effect among related factors, and C) see which particular factor changes most contributed to which particular local outcomes of interest. These abilities greatly aid us in identifying promising changes and adapting them to new contexts.

But doing these is just far harder with culture. Which is why culture has far less structured experimentation, or learning of useful lessons from the variations we see. And this is why cultural evolution has long been much more of a simple Darwinian process of natural selection: variations happen, and then some win out over others.

It is thus far easier to promote innovation in say software or governance, compared to culture. To innovate in culture, we can mainly just induce many new cults, cults that try out whole packages of deep high-scale cultural features, that try to last for generations, and that build up sufficient insularity against outside influences to have a shot at actually retaining kids and preserving distinctiveness for generations.

Most of the cultural experiments that VB celebrates have far too little insularity to plausibly serve in this role. More important, due to several centuries of easier talk, travel, and trade, the world now has many orders of magnitude a) too little variety, at cultures level, and also b) too few attempts to start and grow new cults. So I’m afraid that the amount of added variety that VB could plausibly induce in these parameters, even if he induced 100x more from others, is just vastly insufficient to the need.

But VB, or people like him, could plausibly single-handedly induce far more experimentation and thus innovation in governance. And there is a decent chance that we could find far more competent forms of governance, which we might then assign to the task of fixing and improving our processes of cultural evolution, including its variety. Yes, even this seems a long shot, and so my best prediction is that we will fail, and our civilization will decline, to be replaced by descendants of the Amish, Haredim, etc. But its the best shot I can see.