Lots of interesting insights, but their affirmative action take is a miss.
> Critics of affirmative action often commit the fallacy of letting a failure in one area doom the entire enterprise. This ignores the interdependent nature of affirmative action. [1]
Affirmative action sets up a zero-sum game where fixed resources like university admissions and employment offers are redistributed to people with the "correct" demographics. The conflict is not a disagreement over effectiveness. It's a misalignment between meritocracy and equity.
Do you disagree that some critics of AA are committing that fallacy?
AA is being used as an example of the failure mode where:
"The failure of a single component does not mean the program is fatally flawed; rather, it highlights the need for a comprehensive, coordinated approach"
Indeed, I'm sure the author would agree that part of the comprehensive solution is to increase the amount of university admission slots.
That feels more like a cop-out than a legitimate criticism of a fallacy.
If the author could propose an affirmative action program that didn’t have that “single component” at the core of how it operates then I’d be more interested in the argument, but as-is it just feels like an attempt to forcefully ignore valid criticisms.
I think something that often isn't considered with affirmative action is the benefits that are conferred to the people who are not in a minority. In other words it is a genuinely useful thing to go to a university with a broad spectrum of people and ideas.
In a purely meritocratic sense, all other beings equal a university that provides a diverse faculty and student body will better educate its students than a university that doesn't, all other things remaining equal.
The problem in practice is that these programs don’t actually select for diverse ideas, they select for demographic traits like gender or ethnic background.
If the team uses relational databases but someone shows up to an interview with a strongly held belief that NoSQL is the way to go, they’re likely to be rejected because their ideas don’t match the team’s. Diversity programs in practice never even attempt to push diversity of ideas, they ignore all of that and focus on things like gender and ethnic background.
This feels like a dangerous opinion to voice, but the workplace affirmative action programs I’ve seen in practice have been very poor in their implementation. At my last workplace that instituted diversity targets, HR would just start rejecting hires if they thought it would skew the diversity numbers in the wrong way. So you’d hit a wall where the only candidates you were allowed to hire couldn’t be, for example, men or of Asian descent or some other demographic trait they thought was over-represented. None of this improved diversity of ideas, it became a gender or skin color matter.
But those resources are already redistributed (from a distribution that somewhat aligns with demographics) with things like personal relationships (think legacy admissions or a father's buddy handshake internship). AA is meant to correct historical instances of this which snowball into familial / generational wealth and (most difficult to diffuse) social capital that was distributed unfairly.
That's the argument for it, not my belief. The argument for AA is that the so-called meritocracy had/has its own unequal distributions.
>AA is meant to correct historical instances of this which snowball into familial / generational wealth and (most difficult to diffuse) social capital that was distributed unfairly.
If that was the case it would be based on family wealth/income.
I think a major flaw of all these models is that they underestimate:
1. How easy it is to start fresh and shed your past reputation if you get caught doing something bad.
2. How forgiving people are and how tolerant they are to deception, abuse and immorality. I hate to say it but a lot of people are attracted to abusers. They keep going back to the same kinds of people who will abuse them over and over. These same people who tolerate abuse often seem to show disrespect and look down on good, honest people. I cannot overstate how powerful this effect is; and it seems to be getting worse over time! And these people keep coming up with narratives to gaslight themselves about their abusers "they're not so bad"... People will especially do this when their abuser has power over them (Stockholm Syndrome).
Once you factor these two things, cheating is the clear winning strategy. By a mile... It's objectively a superior strategy. If we just follow game theory; it will take us somewhere really dark. Game theory isn't what's keeping the world civilized. Society literally all rests on people's irrational emotions and moral principles.
The desire to do the right thing is completely irrational and is a net loss to the individual. If we continue with the current system and current assumptions, all moral individuals will be wiped out because they are at a HUGE disadvantage. To solve our social problems, we need to be more moral; we need to learn to judge ourselves and other people through the lens of morality and be very firm about it.
Writing from Japan. You are absolutely right about the "Finite Game".
If you can reset your reputation and start over, "Cheating" is indeed the winning strategy.
However, here in Japan, we have a different operating system called "Shinise" (companies lasting over 1,000 years). They play an "Infinite Game".
Their reputation is tied to a "Noren" (shop curtain) or a family name that has been built over centuries. You cannot simply discard it and respawn.
There is a movie hitting theaters here in Tokyo right now called "KOKUHO" (National Treasure). It depicts Kabuki actors who inherit a "Name" (Myoseki) with 400 years of history.
Watching it, I realized: In their world, cheating doesn't just mean losing a job. It means "killing the Name" for all ancestors and future generations. The penalty is infinite.
When the "Reset Button" is removed from the game, "Honesty" and "Sanpo-yoshi" (Three-way satisfaction) naturally become the mathematically dominant strategies.
Cheating only works when you plan to exit.
> 1. How easy it is to start fresh and shed your past reputation if you get caught doing something bad.
True, but this is a necessary feature of a society or workplace to discourage cheating and abuse.
If a person could easily shed their reputation and start over on an equal footing with everyone else, cheating would be a zero-cost option. Cheat until you get caught, then start over and repeat.
This is why trust and reputation are built over time and are so valuable. It’s frustrating for newcomers or those who have lost reputation somehow, but it’s a necessary feature to discourage fraud and cheating.
Cooperation has been "invented" in evolution many times independently and is long term stable in many species.
If your comment was true that fact wouldn't exist.
We may consider the world we live in today competitive, but at the end of the day, humanity is a globe spanning machine that exists due to cooperative behavior at all scales.
Comments such as yours are really missing the forest for the trees.
I suspect that it's really the fact that cooperation is so powerful and pervasive that makes it normal to the point where any deviation from it feels outrageous.
So you focus on the outrageous due to availability bias (seeing the trees rather than the forest).
This doesn't make sense to me, our current prosperity is founded on an enormous mountain of collaboration and shared beliefs. Usually not out of selflessness of course, often guided and forced by strong leadership and/or strong institutional structures to bend selfishness into selflessness (like capitalism to a degree).
Poor countries tend to stay poor not due to fundamental resource constraints but due to self-reinforcing loops of desperate crab-bucket like behavior, where everyone is cheating one another out of necessity (or culture). Broad collaboration and institution building is always the only way out of the hole, although the hole can be very deep and collaboration can be very costly until you get out.
You are right though, that for an individual living in a good collaborative system, often cheating is very effective, it's just that the system can only handle a certain amount of that behavior before it collapses.
As is discussed in the first scene of Plato's The Republic (surprisingly entertaining to modern tastes), the best play tends to be "to be unjust while seeming just". If people are going to be assholes, it is actually much better if they are discrete about it and keep a pretense of civilization. When people start acting conspicuously like assholes, out of a weird sense of honesty, that's when it propagates and the whole thing collapses, like a bank-run. It's an ancient story that we are still living.
> Poor countries tend to stay poor not due to fundamental resource constraints
Sometimes highly shrewd rich countries infiltrate the power structure of poor countries through N-pronged strategy to keep them stuck in a rut so that they don't become future threat, also extract their resources in the meantime.
Indeed, the way out of that is also broad collaboration, sometimes not peaceful or clean.
And the last century showed that this also works at a large scale, we all got a lot richer as a global community by letting poor countries develop and doing business with them, instead of exploiting them to death.
Yes, strongly believe this is the case. The corrupt leaders are rarely chosen by the people; they are installed by foreign powers. There are many cases you can dig into which are absolutely atrocious; like people getting paid big money by western leaders to assassinate their friends to take power and pass laws which facilitate the extraction of resources by foreign corporations.
Like the story of Thomas Sankara's assassination by his trusted childhood friend Blaise Compaoré is quite disturbing. It seems like Compaoré was leader for a very long time and is still in politics... I cannot think of a more morally deprived individual. If game theory was as claimed; nobody should want to work with such deeply disloyal and psychopathic individual. It's just like I say; people have a strong tolerance, even attraction to abusers. If you look at the real story, you notice this pattern over and over... but we are so badly gaslit about such things (aka 'PR') that we don't notice.
Ed Witten here : "So first of all thanks very much. I'm very honored to have the chance to give this talk. Of course Nima and I both wish we could do more for
peace than just to give talks at an online meeting for peace. Unfortunately we know that there are lots of bad things happening in the world and we hope that there will be better days ahead. Hopefully as one would say in Hebrew [..] which means soon in our own day.
> Poor countries tend to stay poor not due to fundamental resource constraints but due to self-reinforcing loops of desperate crab-bucket like behavior, where everyone is cheating one another out of necessity (or culture)
This doesn't seem true and I'd be interested in any stats that back this up. It reminds me of a very interesting result (that most never internalize) which is that the number one way to avoid corruption is to pay public servants handsomely such that the job rivals the private sphere. Most developing countries can't do that, and that's why most of them have issues with corruption.
Rich countries also have crab-bucket like behavior. You don't have to look twice at the current US administration to see lots of corruption and cheating and fraud, for example.
Pretty clear trend: low-trust societies have low gdp and high-trust societies have high gdp, regardless of resource distribution. Africa/South America are resource rich, japan/iceland are resource poor.
I didn't mean it so literally. Having robust taxation and well supported institutions is what I mean by "broad collaboration" and an effective "culture", as in a social operating-system a set of values and habits, that continually support and self-heal such constructions.
I don't meant that everybody should be nice, and that poor countries are somehow culturally nasty, absolutely not. Real collaboration cannot be just founded on morals and good faith, it's not sustainable, it's more about incentives engineering.
In terms of references, the main one that comes to mind is the economics Nobel price from 2024: "for studies of how institutions are formed and affect prosperity".
> the best play tends to be "to be unjust while seeming just"
Yep this is a huge problem now. I think wealth inequality is also making this worse because people often turn a blind eye to the bad behaviors of people who have power over them. This is an extremely powerful effect; it's everywhere. For example, Christians turning a blind eye to certain negative character traits of God as he appears in the old testament. Employees turning a blind eye to the immoral actions of their boss and coming up with justifications to keep them on a pedestal...
The social structure is not determined by morality; it's the other way round; morality is determined by the social structure.
It reminds me of an old French fable in which a lamb tries to reason with a wolf why he should let him live... The wolf listens to the lamb's logic but then he eats it anyway and the story ends with a sentence like "The reason of the strongest is always the best one."
My point (and Plato's) was rather that some people will definitely cheat, because it's locally rational, and it's actually better for everyone if they are "classy" about it and don't flaunt it too much. A minority will get away with terrible things, but somewhat bounded by conspicuousness, and at least the majority remains blissfully (willfully?) unaware and propping up the civilized system which is so much better for all of us.
It is quite a cynical point of view of course. It's a hard balance, when it gets bad sometimes it's better to air the dirty laundry and go through the pain of purging those cheaters.
But the worse thing is to have people be loud and proud cheaters, which is happening more and more. That's a deadly virus to a civilized society, everyone starts thinking they are dumb for not cheating, and we quickly go back to the dark ages.
It's a bit like calling out the bank for being a fraud because they don't have all the money in a vault, and rushing to get your cash out. If people start taking the red pill and shouting that society is just a game of pretend, which it kind of is, then our very real prosperity can vanish overnight.
>> then our very real prosperity can vanish overnight.
This sentence assumes a certain degree of shared prosperity. I think this is increasingly an illusion. IMO, Social media tends to create filter bubbles which create illusions of shared prosperity. Most of the social bubbles I participate in, the view is much more like 'monopolized prosperity' than 'shared prosperity'.
I've been in a unique position to have mingled with billionaires/millionaires and also normal people and the contrast is significant. In some circles; it's like even the company cook is getting rich... In others, it's like there are some really talented people who keep failing over and over and can't make any money at all from their work; like they're suppressed by algorithms.
I think most people wouldn't mind seeing the whole system collapse as they don't feel they have any stake in it; their experience is that of being oppressed while simultaneously being gaslit about being privileged! It's actually deeply disturbing. I don't think most people on the other side have any idea how bad it is because their reality looks really wonderful.
I suspect reality is probably that the oppression which used to be carried out at a distance in Africa is now being applied to specific groups within the same country; and filter bubbles are used to create artificial distance.
Perhaps you'd like https://en.wikipedia.org/wiki/Blockout .
Yours sincerely, a TGM-fan
> Critics of affirmative action often commit the fallacy of letting a failure in one area doom the entire enterprise. This ignores the interdependent nature of affirmative action. [1]
Affirmative action sets up a zero-sum game where fixed resources like university admissions and employment offers are redistributed to people with the "correct" demographics. The conflict is not a disagreement over effectiveness. It's a misalignment between meritocracy and equity.
[1]: https://nonzerosum.games/unlockingsolutions.html
AA is being used as an example of the failure mode where:
"The failure of a single component does not mean the program is fatally flawed; rather, it highlights the need for a comprehensive, coordinated approach"
Indeed, I'm sure the author would agree that part of the comprehensive solution is to increase the amount of university admission slots.
The implicit argument is that AA's largest challenge is a coordination problem. It's not. It's a clash in values and a fight over zero-sum rewards.
If the author could propose an affirmative action program that didn’t have that “single component” at the core of how it operates then I’d be more interested in the argument, but as-is it just feels like an attempt to forcefully ignore valid criticisms.
In a purely meritocratic sense, all other beings equal a university that provides a diverse faculty and student body will better educate its students than a university that doesn't, all other things remaining equal.
If the team uses relational databases but someone shows up to an interview with a strongly held belief that NoSQL is the way to go, they’re likely to be rejected because their ideas don’t match the team’s. Diversity programs in practice never even attempt to push diversity of ideas, they ignore all of that and focus on things like gender and ethnic background.
This feels like a dangerous opinion to voice, but the workplace affirmative action programs I’ve seen in practice have been very poor in their implementation. At my last workplace that instituted diversity targets, HR would just start rejecting hires if they thought it would skew the diversity numbers in the wrong way. So you’d hit a wall where the only candidates you were allowed to hire couldn’t be, for example, men or of Asian descent or some other demographic trait they thought was over-represented. None of this improved diversity of ideas, it became a gender or skin color matter.
That's the argument for it, not my belief. The argument for AA is that the so-called meritocracy had/has its own unequal distributions.
If that was the case it would be based on family wealth/income.
1. How easy it is to start fresh and shed your past reputation if you get caught doing something bad.
2. How forgiving people are and how tolerant they are to deception, abuse and immorality. I hate to say it but a lot of people are attracted to abusers. They keep going back to the same kinds of people who will abuse them over and over. These same people who tolerate abuse often seem to show disrespect and look down on good, honest people. I cannot overstate how powerful this effect is; and it seems to be getting worse over time! And these people keep coming up with narratives to gaslight themselves about their abusers "they're not so bad"... People will especially do this when their abuser has power over them (Stockholm Syndrome).
Once you factor these two things, cheating is the clear winning strategy. By a mile... It's objectively a superior strategy. If we just follow game theory; it will take us somewhere really dark. Game theory isn't what's keeping the world civilized. Society literally all rests on people's irrational emotions and moral principles.
The desire to do the right thing is completely irrational and is a net loss to the individual. If we continue with the current system and current assumptions, all moral individuals will be wiped out because they are at a HUGE disadvantage. To solve our social problems, we need to be more moral; we need to learn to judge ourselves and other people through the lens of morality and be very firm about it.
However, here in Japan, we have a different operating system called "Shinise" (companies lasting over 1,000 years). They play an "Infinite Game". Their reputation is tied to a "Noren" (shop curtain) or a family name that has been built over centuries. You cannot simply discard it and respawn.
There is a movie hitting theaters here in Tokyo right now called "KOKUHO" (National Treasure). It depicts Kabuki actors who inherit a "Name" (Myoseki) with 400 years of history. Watching it, I realized: In their world, cheating doesn't just mean losing a job. It means "killing the Name" for all ancestors and future generations. The penalty is infinite.
When the "Reset Button" is removed from the game, "Honesty" and "Sanpo-yoshi" (Three-way satisfaction) naturally become the mathematically dominant strategies. Cheating only works when you plan to exit.
True, but this is a necessary feature of a society or workplace to discourage cheating and abuse.
If a person could easily shed their reputation and start over on an equal footing with everyone else, cheating would be a zero-cost option. Cheat until you get caught, then start over and repeat.
This is why trust and reputation are built over time and are so valuable. It’s frustrating for newcomers or those who have lost reputation somehow, but it’s a necessary feature to discourage fraud and cheating.
If your comment was true that fact wouldn't exist.
We may consider the world we live in today competitive, but at the end of the day, humanity is a globe spanning machine that exists due to cooperative behavior at all scales.
Comments such as yours are really missing the forest for the trees.
I suspect that it's really the fact that cooperation is so powerful and pervasive that makes it normal to the point where any deviation from it feels outrageous.
So you focus on the outrageous due to availability bias (seeing the trees rather than the forest).
Evolution does not work maximizing individual success.
Poor countries tend to stay poor not due to fundamental resource constraints but due to self-reinforcing loops of desperate crab-bucket like behavior, where everyone is cheating one another out of necessity (or culture). Broad collaboration and institution building is always the only way out of the hole, although the hole can be very deep and collaboration can be very costly until you get out.
You are right though, that for an individual living in a good collaborative system, often cheating is very effective, it's just that the system can only handle a certain amount of that behavior before it collapses.
As is discussed in the first scene of Plato's The Republic (surprisingly entertaining to modern tastes), the best play tends to be "to be unjust while seeming just". If people are going to be assholes, it is actually much better if they are discrete about it and keep a pretense of civilization. When people start acting conspicuously like assholes, out of a weird sense of honesty, that's when it propagates and the whole thing collapses, like a bank-run. It's an ancient story that we are still living.
Sometimes highly shrewd rich countries infiltrate the power structure of poor countries through N-pronged strategy to keep them stuck in a rut so that they don't become future threat, also extract their resources in the meantime.
And the last century showed that this also works at a large scale, we all got a lot richer as a global community by letting poor countries develop and doing business with them, instead of exploiting them to death.
Like the story of Thomas Sankara's assassination by his trusted childhood friend Blaise Compaoré is quite disturbing. It seems like Compaoré was leader for a very long time and is still in politics... I cannot think of a more morally deprived individual. If game theory was as claimed; nobody should want to work with such deeply disloyal and psychopathic individual. It's just like I say; people have a strong tolerance, even attraction to abusers. If you look at the real story, you notice this pattern over and over... but we are so badly gaslit about such things (aka 'PR') that we don't notice.
https://youtu.be/Ta5Dx327KQc?t=4899
This doesn't seem true and I'd be interested in any stats that back this up. It reminds me of a very interesting result (that most never internalize) which is that the number one way to avoid corruption is to pay public servants handsomely such that the job rivals the private sphere. Most developing countries can't do that, and that's why most of them have issues with corruption.
Rich countries also have crab-bucket like behavior. You don't have to look twice at the current US administration to see lots of corruption and cheating and fraud, for example.
Pretty clear trend: low-trust societies have low gdp and high-trust societies have high gdp, regardless of resource distribution. Africa/South America are resource rich, japan/iceland are resource poor.
I don't meant that everybody should be nice, and that poor countries are somehow culturally nasty, absolutely not. Real collaboration cannot be just founded on morals and good faith, it's not sustainable, it's more about incentives engineering.
In terms of references, the main one that comes to mind is the economics Nobel price from 2024: "for studies of how institutions are formed and affect prosperity".
Yep this is a huge problem now. I think wealth inequality is also making this worse because people often turn a blind eye to the bad behaviors of people who have power over them. This is an extremely powerful effect; it's everywhere. For example, Christians turning a blind eye to certain negative character traits of God as he appears in the old testament. Employees turning a blind eye to the immoral actions of their boss and coming up with justifications to keep them on a pedestal...
The social structure is not determined by morality; it's the other way round; morality is determined by the social structure.
It reminds me of an old French fable in which a lamb tries to reason with a wolf why he should let him live... The wolf listens to the lamb's logic but then he eats it anyway and the story ends with a sentence like "The reason of the strongest is always the best one."
It is quite a cynical point of view of course. It's a hard balance, when it gets bad sometimes it's better to air the dirty laundry and go through the pain of purging those cheaters.
But the worse thing is to have people be loud and proud cheaters, which is happening more and more. That's a deadly virus to a civilized society, everyone starts thinking they are dumb for not cheating, and we quickly go back to the dark ages.
It's a bit like calling out the bank for being a fraud because they don't have all the money in a vault, and rushing to get your cash out. If people start taking the red pill and shouting that society is just a game of pretend, which it kind of is, then our very real prosperity can vanish overnight.
This sentence assumes a certain degree of shared prosperity. I think this is increasingly an illusion. IMO, Social media tends to create filter bubbles which create illusions of shared prosperity. Most of the social bubbles I participate in, the view is much more like 'monopolized prosperity' than 'shared prosperity'.
I've been in a unique position to have mingled with billionaires/millionaires and also normal people and the contrast is significant. In some circles; it's like even the company cook is getting rich... In others, it's like there are some really talented people who keep failing over and over and can't make any money at all from their work; like they're suppressed by algorithms.
I think most people wouldn't mind seeing the whole system collapse as they don't feel they have any stake in it; their experience is that of being oppressed while simultaneously being gaslit about being privileged! It's actually deeply disturbing. I don't think most people on the other side have any idea how bad it is because their reality looks really wonderful.
I suspect reality is probably that the oppression which used to be carried out at a distance in Africa is now being applied to specific groups within the same country; and filter bubbles are used to create artificial distance.