Has SIAI/​FHI considered putting up prizes for contributions to important problems?

[jsalvatier]

Because of my interest in reducing LessWrong’s public goods problem, I’ve been reading about prizes. Prizes seem like they could be very useful for shifting the perception of the scientific and mathematical public on the importance of existential risk.

Prizes have a couple of attractive properties

1. Shift public perception about what is important and possible

2. Draw submissions from non-traditional sources

3. Mobilize existing resources to work on your problem (e.g. people use grants to try to win prizes)

It seems to me that there are several open problems related to existential-risk that might be amenable to prizes. Three prizes seem immediately attractive:

1. Reflective Decision Theory - This is an important unsolved problem, and the criteria seem like they could be fairly objective (a set of problems the decision theory must handle ‘well’, plus some judging). Of course, perhaps SIAI/​FHI do not want reflective decision theory to be public; in that case a prize would be a bad idea. This prize should probably be called The Good Prize after I.J. Good.

2. Contributions towards a Theory of Friendliness—Shifting public opinion on this issue is very important, so I think a prize is attractive. The criteria for strong contributions are fuzzier, so you would need especially respected and trusted judges.

3. Contributions towards understanding existential risks—Same as above.

Is my enthusiasm for this approach unwarranted? Are there other related topics amenable to prizes?

Disclosure: I have only recently read up on prizes and I think I am in fanboy mode, so take my excitement with a grain of salt.

[Morendil]

Maybe you ought to read something contrary. I suggest Alfie Kohn’s Punished by Rewards which is pretty scathing on the whole topic of extrinsic motivation, citing plenty of research in evidence. A drawback is that it’s pretty old, the research is from the 1980′s.

[jsalvatier]

For others interested, the google scholar keywords ‘extrinsic motivation review’ seem to work well (link).

[timtyler]

There are a few prizes to do with machine intelligence listed here.

One attempt to do something similar is here. I think that attempt was pretty farcical, though: it is pretty obvious that noboby will win that money.

[jsalvatier]

Thanks, this reduced my unjustified enthusiasm. I wonder if more specific prizes (such as a reflective decision theory prize) would do a better job of attracting effort and shifting perception.

[timtyler]

FWIW, personally, I think that sensibly-designed prizes are a great way to pay for getting things done. I note that SIAI has tried some prizes—including one last year.

[jsalvatier]

Interesting. Do you know they make a decent spectacle of awarding that prize? Publicity (pre and post prize) seems to be a key element to effective prizes.

[malthrin]

It seems you’re not the only one who recently read up on prizes. Google just launched Prizes.

[jsalvatier]

Interesting; thanks :)

[lukeprog]

It seems that InnoCentive might be the place to do this.