More generally it would probably be worthwhile trying to rewrite what we think of as central LW insights without using LW-specific jargon, in fact using as little technical jargon at all as you can get away with.
(That amounts to an exercise in "inferential distance golf" - we know that it's possible to get insight X across by showing a smart reader a path which uses N terms of jargon each of which bridges an inferential gap, but if it is possible to explain X to the same smart reader in fewer than N steps requiring new terms, that's a win.)
"inferential distance golf"
That's an excellent phrase. Could you (or someone) please write a post laying out the game of inferential distance golf, with some real-life examples? (A university degree is a more-or-less finely-tuned set of games of inferential distance golf, for example.) Could be most useful as a way of thinking.
More generally it would probably be worthwhile trying to rewrite what we think of as central LW insights without using LW-specific jargon, in fact using as little technical jargon at all as you can get away with.
I couldn't agree more. The key phrase from Roko's post is:
Linking to posts explaining jargon is NOT OK.
This is one of the main problems with the sequences: even posts that don't use technical jargon often link to posts that do.
This is one of the main problems with the sequences: even posts that don't use technical jargon often link to posts that do.
Would it even be possible to rewrite the sequences so that every post is standalone? I don't think so.
They don't have to be standalone, but instead of links they could include short summaries of the ideas they depend upon so that the reader understands the relationship between concepts without breaking the "flow" of reading.
I doubt one can rewrite the entire sequences that way. Things do build.
However, could one write the basic insights that way? I strongly suspect so. (I may be cheating with my definition of "basic," but I think if you put together a list of standalone ideas in the sequences, you would have a pretty good list of articles.)
It would, in fact, be possible.
You'd just have to target them more at people who are already well-versed in traditional academia, i.e. people who have degrees from top universities or PhDs or are in some equivalently intellectually demanding position (e.g. Investment banker, Lawyer, CEO, etc).
If you narrow the target audience you can assume more in terms of speed of uptake, and you can take as granted various norms of traditional rationality (such as what counts as a sound argument, the need for empiricism, the weight of peer-reviewed journal publications, etc) and perhaps even some domain knowledge of basic math and economics.
At the same time, you would want to frame the material in a more serious, less cranky, less jargon-y way.
Isn't Eliezer writing a book? I imagine that that would be closer to that. Like, while chapters/sections aren't standalone the book is organized sequentially enough that each section only depends on the ones before it.
it doesn't provide a good error theory, i.e. it just states the correct answer without explaining, in detail, why our cached answers and instincts are wrong.
You don't get anywhere by doing that. People just backwards rationalize why the cached answer was correct anyway.
It is also too long.
I may present a small prize to the winner, if (s)he desires it!
If the point of the essay is that we should approach charitable work with the same "hardheadedness" that we do with other types of financial transactions then you should tell us how much the winner will receive.
you should tell us how much the winner will receive.
Maybe he should. But it doesn't follow from
we should approach charitable work with the same "hardheadedness" that we do with other types of financial transactions
Would you take a job from someone you didn't know for a salary that wouldn't be told to you until after you finished the job?
I understood that being hardheaded about charitable work means carefully choosing which work to do based on how much expected good it produces. Not choosing the job that pays you the most money. Unless you're planning to donate it, anyway.
There's nothing wrong in general about wanting to know how much you're going to get paid, of course.
Two other important points (perhaps for a different article) are 1) your intuition about charitable projects is very bad 2) many people do not realize how hard it is to evaluate charitable projects in these terms. GiveWell talks about these issues a fair amount, but I am not sure if they have an essay I can point to. The reasons are related. 2 is because the charity givers are not usually on the receiving end of charity so they don't get feedback on which to evaluate what improves people's lives and what doesn't without explicitly trying to measure this. 1 follows from 2 because you don't get good feedback for similar decisions to help calibrate your intuition. It is also exacerbated by the fact that the people you are trying to help have very different circumstances from you (in many ways wealth level, culture, economic opportunities etc.).
I'd like to clarify that your intuition is bad not only because you're not built to think about large situations (though that's certainly an issue) but because you are not the person who gets to see the results.
And sometimes the results or non-results aren't very "see-able". For example, if you donate to a puppy rescue charity instead of, say, donating to the running of Givewell, it is kind of hard to visualize what was lost.
9 points is a lot of points to make in one article. 1800 words might not be enough to make all these points.
Well, we'll just have to see what points on the tradeoff-front the competitive forces find. If several different tradeoffs emerge (e.g. some too long, some missing 2-3 points), none of which look obviously best, I'll make several of my friends read each of them and see what the results are.
Points (2) and (8) are very similar and can be amalgamated I think.
We could also have an 1800 word version and an extended 2500 word version. I feel that 1800 is what someone smart can reasonably be asked to read in one sitting, though.
Roko,
Great to see you again!
I may make an attempt of my own over the next week or so.
Regarding:
It is usually better to work a highly-paid job and donate because if you work for a charity you replace the person who would have been hired had you not applied
Certainly this will be true for some people, but my impression is that skilled labor is often a limiting factor in the philanthropic world and this makes the situation more murky. I'll have a lot more to say about this in a future post.
Excellent. I think you would write an excellent article, and Jsalvatier I see has put up some $.
Regarding $ vs time, typically if you have 10 equally skilled people who can each either donate $30k/yr or offer $30k/yr worth of labour, you end up best off having 7 donors and 3 workers, because you have to pay salaries to the workers (even if small salaries), plus overheads, plus money spent on tangibles like advertising. And probably at least one of your 3 workers should be a dedicated fund-raiser/donor relations.
So even if someone can offer skilled, ideologically dedicated labour, 66% of such people should just become donors.
Probably the best we have, but it isn't what I'm after, because it is optimized to be a sequence article rather than a standalone. e.g. talking about AI at the beginning is going to really put people off. Also doesn't speak about purchasing warm fuzzies and utilons separately, and doesn't present the evidence from GWWC. But thanks for the link.
Also too Jargon-y: What is "fun theory"?
I have a friend who is currently in a lucrative management consultancy career, but is considering getting a job in eco-tourism because he "wants to make the world a better place" and we got into a debate about Efficient Charity, Roles vs. Goals, and Optimizing versus Acquiring Warm Fuzzies.
I thought that there would be a good article here that I could send him to, but there isn't. So I've decided to ask people to write such an article. What I am looking for is an article that is less than 1800 words long, and explains the following ideas:
but without using any unexplained LW Jargon. (Utilons, Warm Fuzzies, optimizing). Linking to posts explaining jargon is NOT OK. I will judge the winner based upon these criteria and the score that the article gets on LW. I may present a small prize to the winner, if (s)he desires it!
Happy Writing
Roko
EDIT: As well as saying that he will pay $100 to the winner, Jsalvatier makes two additional points that I feel should be included in the specification of the article:
6. Your intuition about what counts as a cause worth giving money to is extremely bad. This is completely natural: everyone's intuition about this is bad. Why? Because your brain was not optimized by evolution to be good at thinking clearly about large problems involving millions of people and how to allocate resources.
7. Not only is your intuition about this naturally very bad (as well as cultural memes surrounding how to donate to charity being utterly awful), you don't realize that your intuition is bad. This is a deceptively hard problem.
And I would also like to add:
8. Explicitly make the point that our current norm of ranking charities based upon how much (or little) they spend on overheads is utterly insane. Yes, the entire world of charities is stupid with respect to the problem of how to prioritize their own efforts.
9. Mention the point that other groups are slowly edging their way towards the same conclusion, e.g. Giving What We Can (GWWC), Copenhagen Consensus, GiveWell.