Worth linking to here—on this website, specifically? Are you sure?? Personally, I’m happy to take the bait as part of my desperateandvehement Free Speech for Shared Maps campaign, but I’m guessing a lot of other users would rather we silently Just Not? That is, having Less Wrong posts about censorship as a general phenomenon (perhaps illustrated with fictional or non-contemporary examples, as tradition requires) is clearly within-charter of advancing the art of human rationality, but having Less Wrong posts about alleged censorship of specific topics might be off-charter if it were the case that those topics are censored and conversations about whether or not they’re censored are too likely to run afoul of the censorship, or if they’re not censored but talking about whether they were would leak too much info about censored topics in nearby counterfactual worlds.
Okay, probably you can guess from the previous paragraph that I disagree with Jackson and Winston about the taboo being mythical (I tried to write the last sentence of the previous graf in a way that paid lip service towards Glomarization, but I doubt anyone bought it), but more importantly and fundamentally than that, I think the “taboo: mythical or real?” framing suffers from a pretty bad Sorites problem which I think we could do a pretty good job of dissolving (albeit with more effort than I have to spare right now) by going into more detail about specifically which coalitions are trying to suppress what information about which topics, by what means, in which venues, and how well they’re succeeding. For example, Jackson and Winston are correct to point out that Eysenck/Jensen/Rushton/Lynn/Gottfredson kept their tenured jobs, but that’s only setting an upper bound on the strength of incentives against hereditarianism (the tenure system still works!), not showing that the incentives are mythical.
I don’t think this paper is being very fair to the scientific claims of hereditarian researchers. For example, Jackson and Winston write,
a high heritability index [...] had no bearing on the source of any Black–White differences in intelligence test scores, as the sources of within-group variation could not be used to explain between-group variation (Lewontin, 1970).
Now, it’s true that within-group heritability doesn’t directly imply between-group heritability—Lewontin’s seed analogy is germane and competent hereditarians absolutely know this—but I think “no bearing” is false in a Bayesian sense: there’s a relationship between within-group and between-group heritability such that the former is relevant (if not in itself dispositive) evidence about the latter. Jensen covers these issues in a lot of detail in Ch. 12 of his The g factor: basically, if two groups differ by one standard deviation in some phenotypic trait, and the within-group heritability is 0.7 but you think the between-group heritability is zero, that implies that the groups’ environments differ by 1.83 standard deviations (in whatever aspects are relevant to the development of the trait). But that’s a specific in-principle-falsifiable empirical prediction! If you want to argue that the environments do differ that much (perhaps due to factors like structural racism that we don’t know how to adequately measure), fine!—but that’s not the same thing as “no bearing”.
Hereditarian researchers still call for establishing a two-tiered educational system for White and Black people (Cofnas, 2020, p. 134).
I have a 1000+ karma non-anonymous lesswrong account (not this one obviously) and I can tell you that if lesswrong got a reputation as a hereditarian hangout, I would delete that account and move my blog posts to WordPress or something instead. I don’t want to take that risk to my reputation, to my current and future jobs, and to my current and future relationships.
Sadly, for people who need a job, the only free speech is anonymous speech.
There should be a big warning against using your real name in the account creation dialog. I feel more comfortable online since I stopped using my full name.
Rationalists[1] already have something of a reputation for being hereditarians[2].
[1] By which I mean something like “people who use the term ‘rationalist’ for themselves in internet discussions”. (The term has a number of other uses.)
[2] By which I mean something like “people who think it likely that there are significant differences in important psychological characteristics between groups that approximate the popular idea of races”. (The term has a number of other uses.)
The relevant metric is probably how easy it is for an average HR person to make the connection.
If you use your full name on a website X, and the website X is classified as wrongthink in Wikipedia or RationalWiki or whatever source the HR person might use—case closed. If the HR person needs to communicate their judgment to someone else, they can just copy the relevant paragraph describing website X, and a screenshot of your article confirming that you indeed are associated with website X.
But “rationalists” are just one small group among many. Most HR people probably never heard about it. Even if they did, it would be difficult to communicate to an outsider why sharing too many articles about Bayes Theorem on your Facebook page proves that you are an evil person.
Hmm, maybe. Presumably the scenario you have here is applying for a job; I don’t think anyone’s going to get fired for being associated with Less Wrong unless things get much worse than they are now, one way or another. Even then, this seems to me some way from being a real danger at present.
[EDITED to add:] I do completely agree, though, with your point that seeing that someone uses Less Wrong is much easier than seeing that someone is a rationalist when all you’re doing is a cursory web search or whatever.
I don’t think anyone’s going to get fired for being associated with Less Wrong unless things get much worse than they are now
Yep, neither do I. It’s just this feeling that things today are worse than I would have imagined 10 years ago, so maybe it is prudent to assume that they get even worse 10 years later, and start mitigating the risks now while it is relatively easier.
Specifically, I believe that the privacy-invasion technologies will only get better. I could imagine that in 10 years there will be some machine-learning software that will (1) connect your pseudonymous writings across the internet, based on various “fingerprints” such as rare words you use, the typos you make, frequency of using capital letters and diacritics, etc; and maybe also (2) automatically select the potentially most controversial content. I can also imagine that once this technology becomes generally known, it will quickly become used everywhere, simply because not using it would be considered negligence on the side of the HR person.
Also, god knows what could be considered controversial 10 years in the future; the politics can change in any direction. Perhaps writing on an atheist website will already be quite damning, but that can of course be made much worse by algorithmically cherry-picked quotes. Maybe the future will be more woke, maybe it will be a backlash against wokeness, maybe the definition of what is considered woke will change unpredictably. (That is, even people who feel safe today have a good reason to fear the future.) Maybe in the future it will be acceptable to get fired not only for what you wrote, but also for what your parents wrote, in which case my actions today are mostly hurting my kids. Maybe.
I wish we could have a norm that what happens outside of workplace stays outside of workplace, especially when we are talking about fucking opinions.
Luckily, now we have a scientific paper that smugly insists that all my worries are silly.
(And by the way, my actual object-level opinion on race vs intelligence is “I don’t know”. Which is already quite damning, because all good and decent people already know the right answer, so pretending to not fully understand the actual mechanism of intelligence is surely just an excuse for something evil. I know people in real life who were called out as racist—luckily, outside their jobs—for saying there was a genetic component of intelligence.)
Worth linking to here—on this website, specifically? Are you sure?? Personally, I’m happy to take the bait as part of my desperate and vehement Free Speech for Shared Maps campaign, but I’m guessing a lot of other users would rather we silently Just Not? That is, having Less Wrong posts about censorship as a general phenomenon (perhaps illustrated with fictional or non-contemporary examples, as tradition requires) is clearly within-charter of advancing the art of human rationality, but having Less Wrong posts about alleged censorship of specific topics might be off-charter if it were the case that those topics are censored and conversations about whether or not they’re censored are too likely to run afoul of the censorship, or if they’re not censored but talking about whether they were would leak too much info about censored topics in nearby counterfactual worlds.
Okay, probably you can guess from the previous paragraph that I disagree with Jackson and Winston about the taboo being mythical (I tried to write the last sentence of the previous graf in a way that paid lip service towards Glomarization, but I doubt anyone bought it), but more importantly and fundamentally than that, I think the “taboo: mythical or real?” framing suffers from a pretty bad Sorites problem which I think we could do a pretty good job of dissolving (albeit with more effort than I have to spare right now) by going into more detail about specifically which coalitions are trying to suppress what information about which topics, by what means, in which venues, and how well they’re succeeding. For example, Jackson and Winston are correct to point out that Eysenck/Jensen/Rushton/Lynn/Gottfredson kept their tenured jobs, but that’s only setting an upper bound on the strength of incentives against hereditarianism (the tenure system still works!), not showing that the incentives are mythical.
I don’t think this paper is being very fair to the scientific claims of hereditarian researchers. For example, Jackson and Winston write,
Now, it’s true that within-group heritability doesn’t directly imply between-group heritability—Lewontin’s seed analogy is germane and competent hereditarians absolutely know this—but I think “no bearing” is false in a Bayesian sense: there’s a relationship between within-group and between-group heritability such that the former is relevant (if not in itself dispositive) evidence about the latter. Jensen covers these issues in a lot of detail in Ch. 12 of his The g factor: basically, if two groups differ by one standard deviation in some phenotypic trait, and the within-group heritability is 0.7 but you think the between-group heritability is zero, that implies that the groups’ environments differ by 1.83 standard deviations (in whatever aspects are relevant to the development of the trait). But that’s a specific in-principle-falsifiable empirical prediction! If you want to argue that the environments do differ that much (perhaps due to factors like structural racism that we don’t know how to adequately measure), fine!—but that’s not the same thing as “no bearing”.
Notably, Cofnas denies this.
I have a 1000+ karma non-anonymous lesswrong account (not this one obviously) and I can tell you that if lesswrong got a reputation as a hereditarian hangout, I would delete that account and move my blog posts to WordPress or something instead. I don’t want to take that risk to my reputation, to my current and future jobs, and to my current and future relationships.
Sadly, for people who need a job, the only free speech is anonymous speech.
There should be a big warning against using your real name in the account creation dialog. I feel more comfortable online since I stopped using my full name.
Rationalists[1] already have something of a reputation for being hereditarians[2].
[1] By which I mean something like “people who use the term ‘rationalist’ for themselves in internet discussions”. (The term has a number of other uses.)
[2] By which I mean something like “people who think it likely that there are significant differences in important psychological characteristics between groups that approximate the popular idea of races”. (The term has a number of other uses.)
The relevant metric is probably how easy it is for an average HR person to make the connection.
If you use your full name on a website X, and the website X is classified as wrongthink in Wikipedia or RationalWiki or whatever source the HR person might use—case closed. If the HR person needs to communicate their judgment to someone else, they can just copy the relevant paragraph describing website X, and a screenshot of your article confirming that you indeed are associated with website X.
But “rationalists” are just one small group among many. Most HR people probably never heard about it. Even if they did, it would be difficult to communicate to an outsider why sharing too many articles about Bayes Theorem on your Facebook page proves that you are an evil person.
Hmm, maybe. Presumably the scenario you have here is applying for a job; I don’t think anyone’s going to get fired for being associated with Less Wrong unless things get much worse than they are now, one way or another. Even then, this seems to me some way from being a real danger at present.
[EDITED to add:] I do completely agree, though, with your point that seeing that someone uses Less Wrong is much easier than seeing that someone is a rationalist when all you’re doing is a cursory web search or whatever.
Yep, neither do I. It’s just this feeling that things today are worse than I would have imagined 10 years ago, so maybe it is prudent to assume that they get even worse 10 years later, and start mitigating the risks now while it is relatively easier.
Specifically, I believe that the privacy-invasion technologies will only get better. I could imagine that in 10 years there will be some machine-learning software that will (1) connect your pseudonymous writings across the internet, based on various “fingerprints” such as rare words you use, the typos you make, frequency of using capital letters and diacritics, etc; and maybe also (2) automatically select the potentially most controversial content. I can also imagine that once this technology becomes generally known, it will quickly become used everywhere, simply because not using it would be considered negligence on the side of the HR person.
Also, god knows what could be considered controversial 10 years in the future; the politics can change in any direction. Perhaps writing on an atheist website will already be quite damning, but that can of course be made much worse by algorithmically cherry-picked quotes. Maybe the future will be more woke, maybe it will be a backlash against wokeness, maybe the definition of what is considered woke will change unpredictably. (That is, even people who feel safe today have a good reason to fear the future.) Maybe in the future it will be acceptable to get fired not only for what you wrote, but also for what your parents wrote, in which case my actions today are mostly hurting my kids. Maybe.
I wish we could have a norm that what happens outside of workplace stays outside of workplace, especially when we are talking about fucking opinions.
Luckily, now we have a scientific paper that smugly insists that all my worries are silly.
(And by the way, my actual object-level opinion on race vs intelligence is “I don’t know”. Which is already quite damning, because all good and decent people already know the right answer, so pretending to not fully understand the actual mechanism of intelligence is surely just an excuse for something evil. I know people in real life who were called out as racist—luckily, outside their jobs—for saying there was a genetic component of intelligence.)