Broadly agreed with this, but I’m a bit worried that contests with large prizes can have distortionary effects. That is, they might pull EAs towards using their time in ways which are not altruistically/impartially best. This would happen when an EA switches her marginal time to some contest with a big prize, where she otherwise would have been doing something expected to be more impactful (e.g. because she’s a better fit for it), but which doesn’t stand to win her as much money or acclaim.
For instance, I think the creative writing prize was a really great idea, but my own experience was one of feeling like I really ought to at least ‘buy a raffle ticket’, sinking a few hours into trying to write something halfway decent, then missing the deadling and feeling a bit deflated. I don’t mean this in an arrogant way, but in retrospect I think it’s likely I could have done something better with my time, impact-wise.
One fix could be leaning towards more and narrower contests, which require specialised skills, rather than fewer, highly general contests with large prizes. That way, the prize doesn’t have such a big ‘gravity well’ that it sucks in a lot of ‘might as well try’ folks who were already doing useful stuff.
Another fix, which I like most, is more retroactive prizes/funding. The Forum Prize was a good start, although it’s been retired with a promise of finding better (e.g. more democratic) alternatives. Note that some people have recently beenwriting about and trialling retroactive funding. There is likely more than a post’s worth to discuss here, but I think one key idea is that there are potential EA projects where (i) only a couple people would really be suited to doing, such that (ii) a competitive prize wouldn’t make sense; (iii) it’s really hard to track all those ideas and promise rewards for them in advance; but (iv) it would be great if someone did them and were rewarded. I’d be interested to see more discussion of what scalable retroactive funding could look like in EA. In general, retroactive funding avoids this distortionary worry — as long as you trust the evaluators to judge what was most impactful, you can’t beat just trying to do actually impactful things.
Anyway, bottom line is to notice that incentives can go wrong a bit more easily in an altruistic context, and you need to consider the kind of work that a well-meaning contest is getting people to replace. In some circumstances, the impact of a contest with a very large prize but questionable impact might be therefore be unclear, even imagining that money is free.
I think the forum prize should have focused on EAs not at orgs b/c those EAs are already sufficiently incentivised to do good work and when the prizes are dominated by people already at orgs this dilutes the ability of the forum prizes to highlight and encourage new talent.
Broadly agreed with this, but I’m a bit worried that contests with large prizes can have distortionary effects. That is, they might pull EAs towards using their time in ways which are not altruistically/impartially best. This would happen when an EA switches her marginal time to some contest with a big prize, where she otherwise would have been doing something expected to be more impactful (e.g. because she’s a better fit for it), but which doesn’t stand to win her as much money or acclaim.
For instance, I think the creative writing prize was a really great idea, but my own experience was one of feeling like I really ought to at least ‘buy a raffle ticket’, sinking a few hours into trying to write something halfway decent, then missing the deadling and feeling a bit deflated. I don’t mean this in an arrogant way, but in retrospect I think it’s likely I could have done something better with my time, impact-wise.
One fix could be leaning towards more and narrower contests, which require specialised skills, rather than fewer, highly general contests with large prizes. That way, the prize doesn’t have such a big ‘gravity well’ that it sucks in a lot of ‘might as well try’ folks who were already doing useful stuff.
Another fix, which I like most, is more retroactive prizes/funding. The Forum Prize was a good start, although it’s been retired with a promise of finding better (e.g. more democratic) alternatives. Note that some people have recently been writing about and trialling retroactive funding. There is likely more than a post’s worth to discuss here, but I think one key idea is that there are potential EA projects where (i) only a couple people would really be suited to doing, such that (ii) a competitive prize wouldn’t make sense; (iii) it’s really hard to track all those ideas and promise rewards for them in advance; but (iv) it would be great if someone did them and were rewarded. I’d be interested to see more discussion of what scalable retroactive funding could look like in EA. In general, retroactive funding avoids this distortionary worry — as long as you trust the evaluators to judge what was most impactful, you can’t beat just trying to do actually impactful things.
Anyway, bottom line is to notice that incentives can go wrong a bit more easily in an altruistic context, and you need to consider the kind of work that a well-meaning contest is getting people to replace. In some circumstances, the impact of a contest with a very large prize but questionable impact might be therefore be unclear, even imagining that money is free.
I think the forum prize should have focused on EAs not at orgs b/c those EAs are already sufficiently incentivised to do good work and when the prizes are dominated by people already at orgs this dilutes the ability of the forum prizes to highlight and encourage new talent.