I like this idea, but there are a lot of different things it could mean:
Try to create a “Journal of Effective Altruism”, like an academic version of this here EA Forum, for researchers to exchange ideas and build up the field. This idea could be interesting but it’s vulnerable to the trade-offs expressed by So-Low Growth, and it’s not clear why we’d have to start by buying a prestigious journal—surely we could just start our own journal from scratch.
Buy a mainstream journal and just exert a slight EA bias insofar as we try to highlight higher-impact papers and subjects than the previous journal owners, thereby subtly changing the incentives of a large research community. This is a cool idea and it inspired my “fund a university department” answer.
Buy a journal as an exercise in metascience, in order to implement better scientific standards such as requiring preregistration of hypotheses, etc. This idea could of course be combined with having a slight EA tilt; your answer seems like it is saying we should do a combination of these last two bullet points.
Another big question is, what kind of journal (or university department) should we be aiming to buy? Maybe “Nature” is the most prestigious journal overall, but surely we’d want to target our efforts at a specific field—one that seems like it could be particularly high-impact, but perhaps also one that’s currently misaligned from EA values and seems to have poor scientific standards (that way we’d have lots of room to improve things). Some things that come to mind:
Buy a computer science journal and mostly keep running it like normal, except try to ensure that worries about AI safety are treated as pressing and important rather than niche and weird.
Buy a journal about bioethics and try to steer it away from obsessing about blame, privilege, whether anyone can truly give “informed consent” for anything, etc, in favor of utilitarian-style public health examination of which policies would minimize deaths and suffering during a pandemic, or how a faster approval pace for drugs might save more lives even if it means we let some duds sneak through the process.
Buy a journal about aging, Alzheimer’s, or gerontology and try to steer it towards higher-impact research about the root causes of aging rather than just classifying finer and finer details about the various downstream effects of the decaying health that comes with age.
Buy some kind of journal related to biology or animal behavior and subtly try to get everyone studying the kinds of animal-welfare topics relevant to organizations like Rethink Priorities.
Try to turn an existing journal of economics into a journal about Progress Studies.
Buy a journal in a problematic field, like gain-of-function research or hypersonic missile design or AI design (if such obviously net-negative journals exist) and try to sabotage it by jacking up prices or shutting down the journal entirely.
And so on and so forth. Are some of these better targets than others? (For instance, maybe in biology all the incentives flow from the journals down, but in computer science changing the journals would not do as much to change the overall field.) How much does it cost to actually buy a journal, and is that something it’s possible to actually do?
I like this idea, but there are a lot of different things it could mean:
Try to create a “Journal of Effective Altruism”, like an academic version of this here EA Forum, for researchers to exchange ideas and build up the field. This idea could be interesting but it’s vulnerable to the trade-offs expressed by So-Low Growth, and it’s not clear why we’d have to start by buying a prestigious journal—surely we could just start our own journal from scratch.
Buy a mainstream journal and just exert a slight EA bias insofar as we try to highlight higher-impact papers and subjects than the previous journal owners, thereby subtly changing the incentives of a large research community. This is a cool idea and it inspired my “fund a university department” answer.
Buy a journal as an exercise in metascience, in order to implement better scientific standards such as requiring preregistration of hypotheses, etc. This idea could of course be combined with having a slight EA tilt; your answer seems like it is saying we should do a combination of these last two bullet points.
Another big question is, what kind of journal (or university department) should we be aiming to buy? Maybe “Nature” is the most prestigious journal overall, but surely we’d want to target our efforts at a specific field—one that seems like it could be particularly high-impact, but perhaps also one that’s currently misaligned from EA values and seems to have poor scientific standards (that way we’d have lots of room to improve things). Some things that come to mind:
Buy a computer science journal and mostly keep running it like normal, except try to ensure that worries about AI safety are treated as pressing and important rather than niche and weird.
Buy a journal about bioethics and try to steer it away from obsessing about blame, privilege, whether anyone can truly give “informed consent” for anything, etc, in favor of utilitarian-style public health examination of which policies would minimize deaths and suffering during a pandemic, or how a faster approval pace for drugs might save more lives even if it means we let some duds sneak through the process.
Buy a journal about aging, Alzheimer’s, or gerontology and try to steer it towards higher-impact research about the root causes of aging rather than just classifying finer and finer details about the various downstream effects of the decaying health that comes with age.
Buy some kind of journal related to biology or animal behavior and subtly try to get everyone studying the kinds of animal-welfare topics relevant to organizations like Rethink Priorities.
Try to turn an existing journal of economics into a journal about Progress Studies.
Buy a journal in a problematic field, like gain-of-function research or hypersonic missile design or AI design (if such obviously net-negative journals exist) and try to sabotage it by jacking up prices or shutting down the journal entirely.
And so on and so forth. Are some of these better targets than others? (For instance, maybe in biology all the incentives flow from the journals down, but in computer science changing the journals would not do as much to change the overall field.) How much does it cost to actually buy a journal, and is that something it’s possible to actually do?