I think others have suggested this, but have you thought about putting your 10K GBP into a donor lottery or otherwise saving up to getting a larger donation size? I’d like to see research address that question (e.g., is a 100K donation >10x better than a 10K donation?).
I’ll note that the team should still allocate its time to answering the object level question, so that if they win the lottery they know where they’ll give the money.
It’s hard to recommend this over playing the lottery upfront, so that you actually know whether your research will be directly used or not.
If you think it’s important that the research is done even if the money is not used, would you recommend just doing the research project even with merely a hypothetical £10k that isn’t actually donated at the end?
(1) A 10% chance of donating $100K should be roughly as motivating to a risk-neutral EA as a 100% chance of donating $10K (not taking into account arguments that the risk-neutral utility of money may be nonlinear).
(2) Research around whether to donate $100K or $10K (or how to donate $100K conditional on winning the lottery) would be useful.
“A 10% chance of donating $100K should be roughly as motivating to a risk-neutral EA as a 100% chance of donating $10K (not taking into account arguments that the risk-neutral utility of money may be nonlinear).”—that’s not how human psychology works.
As a piece of presentational feedback, I found it a little frustrating to have a title like this one, and yet not have the term “lego brick” specifically and directly explained until something like the third paragraph :)
Another key reason against looking for “lego bricks” that I don’t think you addressed is that marginal thinking is much more generalizable. You’re publishing all your research work, and if I come along afterwards with £1k or £100k, a conclusion you made based on marginal thinking is much more likely to be useful to me than one tailored to your exact donation size.
My guess is that the value of your research in how it informs and influences others may even exceed the value of the £10k directly: if that’s modestly likely to be true, it seems a strong recommendation to avoid “exact fit” opportunities.
I guess strictly speaking this kind of motivation falls out of scope for your project, which aims simply to find the best way to spend the £10k. But it’s certainly a reason I’m glad you made this choice :)
One additional reason why I ask is that it seems like a lot of OxPri (do you have a preferred abbreviation? OPP is naughty and conflicts with OpenPhil) research to date has been of the form “field X looks interesting and may have some promising opportunities but it doesn’t look like we could do anything for 10K GBP” which makes it feel like the donation size may be limiting.
I think others have suggested this, but have you thought about putting your 10K GBP into a donor lottery or otherwise saving up to getting a larger donation size? I’d like to see research address that question (e.g., is a 100K donation >10x better than a 10K donation?).
I’ll note that the team should still allocate its time to answering the object level question, so that if they win the lottery they know where they’ll give the money.
I definitely agree. But it might be good to see research around whether a donor lottery is worthwhile and how to donate $100K.
It’s hard to recommend this over playing the lottery upfront, so that you actually know whether your research will be directly used or not.
If you think it’s important that the research is done even if the money is not used, would you recommend just doing the research project even with merely a hypothetical £10k that isn’t actually donated at the end?
That would defeat the purpose of the project. I think that the purpose is to spur research and the money is there for extra encouragement.
I don’t think that’s true for two reasons:
(1) A 10% chance of donating $100K should be roughly as motivating to a risk-neutral EA as a 100% chance of donating $10K (not taking into account arguments that the risk-neutral utility of money may be nonlinear).
(2) Research around whether to donate $100K or $10K (or how to donate $100K conditional on winning the lottery) would be useful.
“A 10% chance of donating $100K should be roughly as motivating to a risk-neutral EA as a 100% chance of donating $10K (not taking into account arguments that the risk-neutral utility of money may be nonlinear).”—that’s not how human psychology works.
How easy is it for an EA to overcome that?
Also, if there’s a motivation—impact trade-off, how can we navigate that?
As a piece of presentational feedback, I found it a little frustrating to have a title like this one, and yet not have the term “lego brick” specifically and directly explained until something like the third paragraph :)
Another key reason against looking for “lego bricks” that I don’t think you addressed is that marginal thinking is much more generalizable. You’re publishing all your research work, and if I come along afterwards with £1k or £100k, a conclusion you made based on marginal thinking is much more likely to be useful to me than one tailored to your exact donation size.
My guess is that the value of your research in how it informs and influences others may even exceed the value of the £10k directly: if that’s modestly likely to be true, it seems a strong recommendation to avoid “exact fit” opportunities.
I guess strictly speaking this kind of motivation falls out of scope for your project, which aims simply to find the best way to spend the £10k. But it’s certainly a reason I’m glad you made this choice :)
What’s the time period?
One additional reason why I ask is that it seems like a lot of OxPri (do you have a preferred abbreviation? OPP is naughty and conflicts with OpenPhil) research to date has been of the form “field X looks interesting and may have some promising opportunities but it doesn’t look like we could do anything for 10K GBP” which makes it feel like the donation size may be limiting.