Strongly upvoted. I’d love to see more work like this: well-powered, preregistered, and directly answering an important question.
John Salter
Karma: 699
2 points I’d raise:
-
We’re really bad at predicting how a change would affect our happiness. We overwhelmingly overestimate the impact. If you know this, schlep won’t seem as bigger sacrifice
-
I think we, as a community, need to incentivise schlepping by granting it more social status and discriminating towards people who do it.
-
Forget marginal funding week, here’s what I can do with 6H of your time.
Some great things about this post:
Concise and concrete; is quick to understand precisely what you did and why
Scout mindset; just presents the data without shilling for an idea
Directly relevant to doing good effectively
I’d love to see this exact experiment done on a larger scale with more robust outcome data. Kudos to the author. I hope they post more stuff like this.
Sam Altman / Open AI Discussion Thread
I think this a cool idea that should be tried. If net-negative, it can be undone fairly easily by simply deleting the website. I’m up for being a beta user.
Upvoted. It’s great when people put their requests on the forum so they can be scrutinized thoroughly by the community.
I know nothing about community building. Regardless, here are some things I think would make your case stronger to the average reader:Listing some notable Polish EAs, especially if your outreach brought them into the movement. You’ve mentioned some organisations, but it’s costly for the reader to figure out who exactly you’re referring to and what your role was.
There are some places where vagueness could be concealing strength/weakness. For example, you’ve referred to your slack participants doubling. This could be from two people to four or from 10000 to 20000.
It seems as though you’re pursuing five strategies at once as a fairly small organisation. This is generally a bad idea, because doing one thing well is hard as it is. Can you elaborate on why you think such a broad focus is called for? (e.g. perhaps they are synergistic, or other similar orgs have had success with this approach)
It’s not clear why your plans for 2024 have been made. Can you comment a little on each and why you feel they are high EV?
I suppose your major competitor for this type of funding is Meta Charity Funders themselves. Why should the reader donate to you rather than to them? After all, they are supposedly the experts on what meta charities deserve funding. Wouldn’t it be better to give them the money and letting them choose the best charity that’s gone unfunded from their applicant pool?
The size of the effect makes me wonder if there’s something multiplicative going on. If two independent bad things co-occuring will kill you, then reducing the likelihood of both those things by 30% would reduce the resulting deaths by 51%.
= 0.51
It was three things that needed to co-occur, then 30% reductions would yield a 66% reduction in resulting deaths.
I think this would generalise somewhat to a more complicated causal landscape, so long as a sizable proportion of the deaths are caused by these multiplicative interactions. This could perhaps explain why the reduction in deaths amount to more than the sum of their parts?
Examples of relevant bad things:
- Malnutrition
- Parental illness
- Disease
- Hard times with money
- Doctor burnout / Limited hospital capacity
Very nicely written!
That being said, here are some changes I’d like to see in who we hire / invite:
- Someone who went to Oxford despite being born poor is likely much smarter than their richer peers; familial social economic status should be considered..
- Greater focus on people who’ve sacrificed more. It’s a costly signal of value-alignment
- Prioritise people are cannot have achieved their position through office politics / bullshiting / being carried by their teams e.g. soloprenuers >> management consultant.
I shared this idea pretty strongly a few years back, but have changed my mind due to personal experience of running an organisation with lots of people. I think a community has enough parallels for it to be a useful comparison.
Here’s what changed my mind:
1. The number of one-to-one relationships increases with the factorial of the number of people in a group. Shit gets complicated and it creates a breeding ground for bad behaviour; you can behave horribly and then move on to a new group of people without facing ramifications, because it’s unlikely that those two groups are talking to one and other.2. At least within an organisation, having a lot of people necessitates hierarchy which itself has major drawbacks e.g. it increases the links in the Chinese whisper chain that is organisational communication, affecting information flow both on the way up and on the way down.
3. You’d think that the productivity of a team would be n * average productivity, but it’s more like n * productivity of the worst person; mediocre people lower the standards people hold themselves to and higher performers leave in disgust. New hires then conform the ever lowering standard.
4. A lot of processes are at least somewhat serialized. You can’t make them go faster by increasing the number of people, only increasing the quality will move the needle (e.g. a team of 100000 sprinters will still reach the finish line slower than Usain Bolt)
5. A person is smart. People are stupid. When you have large groups, reputation starts to become decoupled from reality. Rumor becomes the dominant mode of information transfer. People start having very strong opinions about people they’ve barely met or interacted with.
6. The importance of work is power-law distributed. The top priority is usually more important than the tenth through to the nth combined. What really matters is nailing the most important stuff.
Strongly upvoted. It’s really important that people talk about their failed projects publicly and talk about what went wrong.
Lets assume human genetic engineering is wholy morally permissable for sake of discussion.
Wouldn’t it be more effective to clone scientific geniuses? It seems like you could do a lot more good with a lot less effort this way.
This is a fantastic comment
I used YC endorsement as a filter to decide what to include, as that way I know it’s a common enough mistake to justify talking about. Do you watch their youtube channel?
Good point.
You were right! I typo’d—have fixed now, ty.
Re: 2, I don’t. I actually think they are more prone to it. I think funders, for this reason, should prioritise projects that sound unfun to work on because it’s less likely to have be proposed for non-impact reasons.
e.g. if choosing between funding two similar projects, one that sounds horribly boring to execute and another that sounds like a nerd’s wet dream, you should fund the former rather than the latter.
(apologies if I’ve misread your comment)
You should talk to @Brad West