While it’s really good to see that you care about helping people find high-impact jobs inside or outside of EA, I don’t think this idea is likely to work.
Creating a high-quality impact analysis for any given job is very difficult and time-consuming. Doing so for many different jobs in different fields would be nearly impossible; I’m not aware of any crowdsourced resources, even outside of EA, that routinely tackle problems of that caliber.
(Yes, Wikipedia and StackExchange are great, but Wikipedia mostly copies and aggregates content from elsewhere, while StackExchange users have access to the formidable tools of disciplines like mathematics and computer programming. Impact analysis is a much newer field with fewer tools, and requires many different approaches depending on what you want to analyze.)
Consider, for example, this 80,000 Hours post on how many lives a doctor might expect to save. (Note that this post may not reflect 80K’s current thinking, since it came out seven years ago.) They do come up with a number, but it relies on lots of loose estimates and assumptions, and applies only to an “average” doctor (which isn’t much help if you’re deciding between dermatology, brain surgery, and earning-to-give at a bank). And of course, saving lives/QALYs isn’t the only way a job can have an impact, so even this time-consuming 80K analysis is only the beginning of what you might want to produce for this single, broad career path.
Of course, it’s probably better to have the 80K analysis than nothing. But I think “doctor” is one of the easiest professions to think about in this way, because there are lots of doctors who do very similar things. Something like “work in the EU commission” would be much harder to analyze.
I don’t expect the impact analysis to be as good as what 80K aims for, but don’t you think we could improve over the current situation?
For example, imagine someone disagrees with the 80K analysis of being a doctor or due to some combination of skills, contacts and location wants to earn to give as a doctor anyways. While it’s hard to estimate a job’s impact in general, on the job board, people could estimate impact relative to an average doctor. By default, the search would multiply this by 80K’s overall low priority of medicine, but if someone wants to be a doctor for other reasons, they could edit the overall priority to take that into account.
Another example is avoiding a negative direct impact while earning to give or building career capital, or working in an area that 80K hasn’t (yet) evaluated (e.g. cybersecurity). Personally I would find a larger job board helpful even if most positions had only quantitative impact estimates and relied on the law of large numbers for reasonable results.
In general, 80K has a good (or at least decent) estimate of absolute impact for a small number of positions, but given the difficulty of getting one of them, my impression is that there’s space for a much larger list with just relevant areas and relative impact for each position.
I definitely think there’s room for improvement! In particular, I’d be interested to see people who have or want to obtain Job X write about their assessment of the job’s impact—ideally on this very website, though a site cataloging all such reviews and presenting them with nice formatting could also be useful.
(The Forum is a good place to try “content-only” versions of this kind of idea while features like dynamic sorting and web scraping get figured out.)
I’m not sure crowdsourcing from people without direct interest would work very well, but that’s largely based on my experience with other EA crowdsourcing projects; it’s easy to get initial enthusiasm, but few such projects keep going over the long run.
More measures of job impact are on my long list of “things I wish existed, but which seem not to be getting generated by EA’s collective consciousness”. If people can be convinced to make more of them exist, that would be wonderful, but I expect that getting solid writeups will be difficult.
That said, I hope you convince people to try! I’ll definitely read any job-impact posts published on the Forum, and I will appreciate the authors’ efforts, even if they only make loose/”law of large numbers” estimates.
While it’s really good to see that you care about helping people find high-impact jobs inside or outside of EA, I don’t think this idea is likely to work.
Creating a high-quality impact analysis for any given job is very difficult and time-consuming. Doing so for many different jobs in different fields would be nearly impossible; I’m not aware of any crowdsourced resources, even outside of EA, that routinely tackle problems of that caliber.
(Yes, Wikipedia and StackExchange are great, but Wikipedia mostly copies and aggregates content from elsewhere, while StackExchange users have access to the formidable tools of disciplines like mathematics and computer programming. Impact analysis is a much newer field with fewer tools, and requires many different approaches depending on what you want to analyze.)
Consider, for example, this 80,000 Hours post on how many lives a doctor might expect to save. (Note that this post may not reflect 80K’s current thinking, since it came out seven years ago.) They do come up with a number, but it relies on lots of loose estimates and assumptions, and applies only to an “average” doctor (which isn’t much help if you’re deciding between dermatology, brain surgery, and earning-to-give at a bank). And of course, saving lives/QALYs isn’t the only way a job can have an impact, so even this time-consuming 80K analysis is only the beginning of what you might want to produce for this single, broad career path.
Of course, it’s probably better to have the 80K analysis than nothing. But I think “doctor” is one of the easiest professions to think about in this way, because there are lots of doctors who do very similar things. Something like “work in the EU commission” would be much harder to analyze.
I don’t expect the impact analysis to be as good as what 80K aims for, but don’t you think we could improve over the current situation?
For example, imagine someone disagrees with the 80K analysis of being a doctor or due to some combination of skills, contacts and location wants to earn to give as a doctor anyways. While it’s hard to estimate a job’s impact in general, on the job board, people could estimate impact relative to an average doctor. By default, the search would multiply this by 80K’s overall low priority of medicine, but if someone wants to be a doctor for other reasons, they could edit the overall priority to take that into account.
Another example is avoiding a negative direct impact while earning to give or building career capital, or working in an area that 80K hasn’t (yet) evaluated (e.g. cybersecurity). Personally I would find a larger job board helpful even if most positions had only quantitative impact estimates and relied on the law of large numbers for reasonable results.
In general, 80K has a good (or at least decent) estimate of absolute impact for a small number of positions, but given the difficulty of getting one of them, my impression is that there’s space for a much larger list with just relevant areas and relative impact for each position.
I definitely think there’s room for improvement! In particular, I’d be interested to see people who have or want to obtain Job X write about their assessment of the job’s impact—ideally on this very website, though a site cataloging all such reviews and presenting them with nice formatting could also be useful.
(The Forum is a good place to try “content-only” versions of this kind of idea while features like dynamic sorting and web scraping get figured out.)
I’m not sure crowdsourcing from people without direct interest would work very well, but that’s largely based on my experience with other EA crowdsourcing projects; it’s easy to get initial enthusiasm, but few such projects keep going over the long run.
More measures of job impact are on my long list of “things I wish existed, but which seem not to be getting generated by EA’s collective consciousness”. If people can be convinced to make more of them exist, that would be wonderful, but I expect that getting solid writeups will be difficult.
That said, I hope you convince people to try! I’ll definitely read any job-impact posts published on the Forum, and I will appreciate the authors’ efforts, even if they only make loose/”law of large numbers” estimates.