Judging by the number of upvotes, Brandon Reinhart's analysis of SIAI's financial filings is valuable to quite a few people. Similar analysis' of Alcor and the Cryonics Institute would be quite valuable. There has been talk of more work on condensing LW content and placing it on the wiki. I'm sure lots of people would like to know about the literature on low dose asprin. People seem to want a front page more accessible to newcomers. Will these projects get accomplished? Some of them, but probably fewer than optimal. I think we can do better.
I would like to look for ways to channel group willingness to contribute to a project into focused individual willingness to work on a project.
Observations about the problem space
The following is based on discussions at the Seattle Less Wrong meetup.
Many people would get a moderate amount of benefit from such projects, but only a small number would end up putting in the hard work to make them happen.
The people most enthusiastic about a given project may not be the best people to work on the project. Perhaps they have very time consuming jobs or have a hard time being objective about the topic (e.g. someone who gets especially emotional about Cryonics) or have too many other projects already or perhaps they are intellectually motivated but not emotionally motivated by the project which might make it difficult to Get Things Done.
Trying to generalize too early is a risk here. Going out and building fancy tools or otherwise trying something elaborate is probably not a good idea at first. Better to try some concrete trials first and learn from those experiences.
Sources of motivation
There are three major potential sources of motivation: Money (the unit of caring), social status (Karma, kind words etc.), things (pizza, books, cookies, pony pictures).
- Transfers of money (the unit of caring) are often much more efficient than transfers of other goods.
- Extrinsic rewards (especially money) can reduce intrinsic motivation.
- Large monetary rewards can also make relationship between the project contributors and the project sponsors less social.
- Many Less Wrong people are high paid
- Less likely to be motivated by small monetary rewards
- Have more money to contribute to projects.
- Not all Less Wrong people are high paid.
- There are services for collecting donations (link).
- Social rewards
- Social status
- Pizza, books, cookies, pony pictures
- Social pressure
- progress monitoring
Different motivators may work better for different kinds of projects. For example, money might be a counterproductive motivator for social projects but a great motivator for setting up a website.
How have others tackled this?
This is a problem others face as well. How do other similar groups and communities ameliorate it?
- Intrinsic motivation
- Conferring social status on those who do valuable work
- Sprints: several people get together in a single place and work together on a project for a couple of days.
- Main draw seems to be Fun
- Frequently used by Python projects
- Competition/bounties (McKinsey survey of prize literature)
- Provides social and/or material rewards
- Sometimes used on LW (link 1, link 2, link 3).
- Work seem well for some larger open source software projects (link 1, link 2, link 3), though some fail to get off the ground at all.
- Poorly arranged prizes can induce wasted effort
- Judging quality can be a serious issue especially when monetary rewards are involved
- potential for social conflict
- some people are better at dealing with social conflicts than others
- pre-designated arbiters more likely to be trusted than others
- Working groups or otherwise close contact sometimes increase people's motivations via peer pressure.
- Personally requesting someone work on a project can increase their motivation to do so.
- With certain kinds of motivation you often get people agreeing to work on a project and then getting slightly stuck and delaying it indefinitely. (Patri Friedman has given one reason why this might happen)
- Different incentives might work better/worse for different kinds of projects.
- Monitoring project progress could help motivation (it might also have other benefits, such as knowing when to rethink the project or to find another person to work on it).
- Splitting up a project into a number of small clear tasks that individuals can pick up and complete decreases the costs of working on projects. The very fact of announcing, specifying and taskifying a project can induce interest.
- Open projects (Wikipedia, open source projects) are often primarily worked on by a small group of highly dedicated contributors.
- Want to encourage quality
- sometimes something is better than nothing
- sometimes drafts and large output volume is useful for future work
- People most interested in the results of a project are not always the people best suited to do the project.
- High visibility projects
- Increase interest in working on projects
- Completed projects give social rewards to completors
- Completed projects serve as templates for future related projects
- Quantifying aggregate interest (both in terms of number and intensity) is useful for deciding what projects are most important
- Aggregating what skills potential project contributors have is useful for determining what projects are possible
In the interest of Holding Off On Proposing Solutions, please take a moment to try to identify features of the problem space that I have not mentioned before reading the comments. Please mention any features you notice as well as any potential solutions or parts of solutions in the comments. I have some ideas, and I will propose them in the comments.
I might be willing to offer pony pictures as a motivator for projects I deem valuable, if that was intended as a serious possible reward and not a silly one.
It is both silly and serious. Silly in that I find ponies silly and serious in that people would be motivated by it (I definitely want one at some point).
There are four people signed up to be in my next batch when I get around to that; you could get in on that while there are spots left.
Giving away ponies for free lowers their value, if they are intended to become a motivating force.
Ah, but after this next batch is full up, I have no further plans to distribute free ponies.
The lesson I draw from these observations is that we should just try some things. In particular, I am interested in seeing how well monetary bounties work.
what I'm going to try
I'm going to start off small. I intend to start with a monetary bounty for a survey post on Spaced Repetition research. I'd like to see how ChipIn works for collecting additional donations.
I will start a projects wiki page to keep track of both completed and open projects.
I'm curious as to what others think of the idea of using bitcoin bounties. In one sense it is fun, but in another it might pose a trivial inconvenience.
I definitely encourage experimentation. You could always have both.
In the interest of a bit more discussion:
There may be skills helpful to contributing (writing, for an example skill) that people don't know now, but which can be taught quickly enough to justify putting effort into teaching them. Finding some of these might yield benefits.
I find collaboration much easier in person, or at least with real-time communication.
There's no particular mechanism for splitting karma between collaborators.
Also relevant is how often you talk to a group of people. When I have lots of contact, I have lots of motivation and work faster, when I don't, I work more slowly.
What kind of projects are we talking about here?
I don't think too too many people can collaborate on single articles, but there probably are some inefficiencies in which people know things but aren't too good at writing, and vice versa.
It seems like other things are more scalable, but I don't know what other things people want. Research?
I'd write for money, and one pony picture.
You've primarily looked at motivation. Skill and time are also important inputs, which you've mentioned once each:
Skills are analog instead of binary. The overall equation looks something like Productivity = Motivation Skill Time. A zero input will destroy the output, and the inputs can partially substitute for each other.
For my part, I haven't seen any potential projects where being a C++ Whisperer would be useful. If there were such projects, I could probably be persuaded to scrounge up some time for them.
I'm also confused what projects to actually work on. I'm currently collecting information to actually get my Anki article written, but I don't know what else might be desired. Is there a Wiki page or something already? A bounty page would be awesome. (And imo ponies > status.)
Yes, having a bounty page is something I would like to try.
This has a few things people want.
Is that supposed to be an arrow, or a greater-than sign?
A > B iff A is more effective at incentivizing work than B is.
I am very interested in reading your Anki article once it's written and I expect there are others like me; please update your probabilities accordingly. Also, I'll send you a small box of homemade caramels when you finish writing it, if you like.
Awesome! Priority of Anki article increased. I'm currently collecting some scientific evidence so it doesn't all boil down to "I stole that from dedicated autodidacts and it always worked for me".
It's just weird that giving high status for an activity counts as an intrinsic reward.
Upvoted. Raises some good points. I like this style of mapping out concepts as a set of bullet points with minimal introductory matter, which seems to emanate from meetup discussions. Also like the fact that it utilizes Holding Off On Proposing Solutions.