But I think when prioritizing their reading lists people should add a “neglectedness in the EA community score” to avoid echo chambers.
Yeah, I agree with this.
Consider how much original insight and valuable disentanglement research you can really add if you spend years reading the same 50 pop non-fiction books that everyone else has read.
As noted in my other comment, I really don’t think that all things on this list are widely read in EA.
And I think that reading (let’s say ) 5-30 books on lists like this one (of which there will hopefully be more in future!) can also be seen as somewhat akin to doing an undergrad unit or two to get up to speed on a new field. It seems worth noting that:
Many fields have a set of works that most people working in that field are expected to have read some fraction of
although people can each read different particular works from that set
EA contained a lot of very unfamiliar ideas to me when I first joined.
I actually did almost immediately start thinking of original research and post ideas, but it turned out that most were reinventing the wheel or missing key considerations. I think reading books from this list really did help me “get up to speed” and start contributing in better ways.
I acknowledge that it’s possible I could’ve gotten similar or better gains from reading books that are currently less often recommended. But I do think some of the often recommended books are unusually useful for the sort of work I want to do. And I think I would’ve almost certainly been worse off if I’d had no recommendations from EAs (as opposed to “the standard recommendations plus additional, carefully chosen but less common recommendations”; I’d be keen to see us move towards that state, as noted elsewhere).
It may be worth noting that I got into EA from Western Australia and without having studied much university-level econ, philosophy, math, computer science, etc. The process I went through might be less necessary for someone based in Oxford, London, or San Francisco, or someone with a more math-y background.
Generally, people should read more papers and write more literature reviews themselves than reading more popular non-fiction.
For me, there’s hardly a tradeoff between these things: I listen to audiobooks in times of my day when I can’t do much else, e.g. when doing chores or on public transport. I could consume papers during this time using text-to-voice, but obviously text-to-voice isn’t as good as an actual voice actor. (I do spend a decent chunk of the rest of my time reading papers.)
Also, that comment seems to presume that most or all readers of this list will want to be researchers? I think a lot of EAs should be doing things other than research. And for them, it may really make sense for them to:
mostly consume nicely packaged and engaging summaries of key ideas from a wide range of fields
sometimes supplement that by reading papers on particular things
rarely or never write literature reviews.
(Less importantly, although this reading list does lean heavily towards popular non-fiction, it isn’t entirely popular non-fiction. E.g., it includes MacAskill’s thesis and The Strategy of Conflict.)
Also, that comment seems to presume that most or all readers of this list will want to be researchers? I think a lot of EAs should be doing things other than research.
I see your point and agree to an extent. My point was that I recommend people to focus more on active learning is often better than passively consuming content, even if they do not want to be a researcher. Just like at university you do not merely read things but also write essays.
I think the best way to learn things is roughly:
write a review of something yourself
read papers
read (popular) non-fiction books
listen to podcasts
But I agree that podcasts and non-fiction books can be more entertaining and not as cognitively demand especially when you have some time to while doing chores etc.
The point about active rather than passive learning, even just for learning’s sake rather than producing original work, is a good one. But I think there are many more ways to do that than writing literature reviews.
One way that seems especially time efficient is making Anki cards (as I suggest in this post), since that can be done quickly in little gaps while doing chores etc.
Another is writing up “key updates” from a thing one has read—not just copying key passages, but saying how the ideas in the book have changed one’s beliefs or plans. This is something I’m now trying out, and an example can be seen here.
Another way would be writing relatively low-effort commentaries, criticism, analysis, original thoughts, etc. as EA Forum posts, without doing proper literature reviews.
So maybe we can imagine a dimension from very active to very passive learning, and another dimension for how much non-background time is required, and we’d like people to find activities that hit the best tradeoffs on those two dimensions for the various parts of their day/week.
But I agree that podcasts and non-fiction books can be more entertaining and not as cognitively demand especially when you have some time to while doing chores etc.
These are indeed the main benefits of podcasts for me, but one other benefit is that they sometimes contain ideas that haven’t yet been properly written up anywhere. (That obviously doesn’t apply to non-fiction books.)
Another way would be writing relatively low-effort commentaries, criticism, analysis, original thoughts, etc. as EA Forum posts, without doing proper literature reviews.
I agree that active learning and writing doesn’t have to be a literature review-and all these formats actually also work. Perhaps we’re coming full circle and it does actually connect to the point in the other thread: we need to encourage people to write more commentaries.
Yeah, I agree with this.
As noted in my other comment, I really don’t think that all things on this list are widely read in EA.
And I think that reading (let’s say ) 5-30 books on lists like this one (of which there will hopefully be more in future!) can also be seen as somewhat akin to doing an undergrad unit or two to get up to speed on a new field. It seems worth noting that:
Many fields have a set of works that most people working in that field are expected to have read some fraction of
although people can each read different particular works from that set
EA contained a lot of very unfamiliar ideas to me when I first joined.
I actually did almost immediately start thinking of original research and post ideas, but it turned out that most were reinventing the wheel or missing key considerations. I think reading books from this list really did help me “get up to speed” and start contributing in better ways.
I acknowledge that it’s possible I could’ve gotten similar or better gains from reading books that are currently less often recommended. But I do think some of the often recommended books are unusually useful for the sort of work I want to do. And I think I would’ve almost certainly been worse off if I’d had no recommendations from EAs (as opposed to “the standard recommendations plus additional, carefully chosen but less common recommendations”; I’d be keen to see us move towards that state, as noted elsewhere).
It may be worth noting that I got into EA from Western Australia and without having studied much university-level econ, philosophy, math, computer science, etc. The process I went through might be less necessary for someone based in Oxford, London, or San Francisco, or someone with a more math-y background.
For me, there’s hardly a tradeoff between these things: I listen to audiobooks in times of my day when I can’t do much else, e.g. when doing chores or on public transport. I could consume papers during this time using text-to-voice, but obviously text-to-voice isn’t as good as an actual voice actor. (I do spend a decent chunk of the rest of my time reading papers.)
Also, that comment seems to presume that most or all readers of this list will want to be researchers? I think a lot of EAs should be doing things other than research. And for them, it may really make sense for them to:
mostly consume nicely packaged and engaging summaries of key ideas from a wide range of fields
sometimes supplement that by reading papers on particular things
rarely or never write literature reviews.
(Less importantly, although this reading list does lean heavily towards popular non-fiction, it isn’t entirely popular non-fiction. E.g., it includes MacAskill’s thesis and The Strategy of Conflict.)
Yes agree with much of this!
I see your point and agree to an extent. My point was that I recommend people to focus more on active learning is often better than passively consuming content, even if they do not want to be a researcher. Just like at university you do not merely read things but also write essays.
I think the best way to learn things is roughly:
write a review of something yourself
read papers
read (popular) non-fiction books
listen to podcasts
But I agree that podcasts and non-fiction books can be more entertaining and not as cognitively demand especially when you have some time to while doing chores etc.
The point about active rather than passive learning, even just for learning’s sake rather than producing original work, is a good one. But I think there are many more ways to do that than writing literature reviews.
One way that seems especially time efficient is making Anki cards (as I suggest in this post), since that can be done quickly in little gaps while doing chores etc.
Another is writing up “key updates” from a thing one has read—not just copying key passages, but saying how the ideas in the book have changed one’s beliefs or plans. This is something I’m now trying out, and an example can be seen here.
Another way would be writing relatively low-effort commentaries, criticism, analysis, original thoughts, etc. as EA Forum posts, without doing proper literature reviews.
So maybe we can imagine a dimension from very active to very passive learning, and another dimension for how much non-background time is required, and we’d like people to find activities that hit the best tradeoffs on those two dimensions for the various parts of their day/week.
These are indeed the main benefits of podcasts for me, but one other benefit is that they sometimes contain ideas that haven’t yet been properly written up anywhere. (That obviously doesn’t apply to non-fiction books.)
I agree that active learning and writing doesn’t have to be a literature review-and all these formats actually also work. Perhaps we’re coming full circle and it does actually connect to the point in the other thread: we need to encourage people to write more commentaries.