If I'm understanding you correctly, you don't think people who can't individually affect the equilibrium are evil? Scientists who would be outcompeted, and therefore unable to do science, if they failed to pursue a maximally impressive career seem an example of this. If they're good (in terms of both ability and alignment), there's some wiggle room in there for altruism when it's cheap, but if err too far from having an impressive career, someone else, probably someone not making small career sacrifices for social benefit, gets the attention and funds instead, thereby reducing the total amount of good being done.

I'd be interested in some concrete examples of appealing to people's conscience getting us less-bad equilibria so I can better understand what you're getting at. Is the scientists who all resigned from the board of an Elsevier-owned journal and started their own an example?

Also interested in your thoughts where we're in a suboptimal equilibrium that we're trying to get out of and there's 2+ optimal ones (in the sense that there's no other equilibrium which is a Pareto improvement), but each competes with the other. For instance, suppose it's several years ago and we agree that it's better that gay couples have the same right to legal marriage as straight couples, but the two better equilibria are (1) elimination of marriage as a legal institution and (2) extension of marriage to gay couples, which is good for gay couples who want the legal benefits of marriage and less good for those who don't want marriage but suffer from e.g. less favorable tax treatment. (I'd really like a better example for this, especially after the examples EY gave, but lack a large, responsive group of fb followers I can get to brainstorm examples for me.)

Testing effect.

(At this point, I should really know better than to trust myself to write anything at 1 in the morning.)

if you’ve read all of a sequence you get a small badge that you can choose to display right next to your username, which helps people navigate how much of the content of the page you are familiar with.

Idea: give sequence-writers the option to include quizzes because this (1) demonstrates a badgeholder actually understands what the badge indicates they understand (or, at least, are more likely to) and (2) leverages the testing effect.

I await the open beta eagerly.

I have taken the survey.

Extremely interested, would move anywhere rationalists would set one of these up.

When I first read In Fire Forged, I really liked it, but saw things I could improve. So, I left some high-quality reviews on fanfiction.net (that is, reviews that demonstrated I somewhat knew what I was talking about) and then solicited the author. From there, networking (people who you collaborated with can collaborate with you).

Back-engineering, I'd tentatively suggest just posting somewhere with reasonable visibility that selects for writers you'd like to collaborate as, and then ask anyone interested to ping you. Alternatively, you could develop a relationship working on someone else's writing and then ask them to look at your's.

You guys voted to develop Righteous Face Punching Style and add Kagome to your party. What do you need my help in decision-making for? (But, seriously, I probably shouldn't have taken the time to get caught up, much less actively participate. Fun read, though!)

Ha! I give Lighting Up the Dark—also by Velorien—last pass editing.

Thanks for the rec. It looks really good.

Do you have any examples of pieces that were written collaboratively?

In addition to In Fire Forged (in which I did first-round micro, in addition to contributing to worldbuilding), I give a last pass micro to Lighting Up the Dark (rational Naruto fanfic). I contributed a little to the Second Secular Sermon, although verse is really not my thing. I also have a partnership with Gram Stone that includes looking over each other's LW posts.

Do you keep a history of changes and discussions?

In Fire Forged has a Skype group, which keeps an archive of our discussion. Since Google Docs aren't the final publishing form, you can keep comments around, although in practice, once we've resolved an issue, the comment/suggestion usually goes away, so things don't get more cluttered. If you're interested, this is the Google Doc for this piece. But Google Docs doesn't keep a changelog, I have no desire to look back at one, nobody I've talked to has indicated any desire to look back at one, so there is no history of changes.

How do you determine the direction of the story, is there a single leader who makes the big decisions, or is it more egalitarian?

I more fully discussed this here, but the tl;dr is that experience indicates a single-leader setup usually works best, and is also the only setup I've come across. That said, it's egalitarian in the sense that the primary author doesn't give any special consideration to the words they've written or the ideas they've had; in the end, you want the best ideas expressed by the best words on the page. I can't imagine the author who would pass up improvements to their creative baby just because they weren't the ones to come up with them.

I'm sorry my title misled you.

(Since writing has trouble carrying intent: I genuinely feel bad that the title I chose caused you to believe something that wasn't true. I wish I was smart enough to have come up with a title that more precisely communicated what I was and wasn't discussing.)

This is perhaps a case of different projects being best served by different practices. There's certainly nothing stopping you from making a Google Doc where two (or more) authors have editing permission (as opposed to commenting permission).

But it's absolutely true that I'm writing from the perspective of having one primary author. This is because every piece I've worked on has had one primary author. Paul Graham writes: "Design usually has to be under the control of a single person to be any good." Indeed, almost all books of fiction I'm aware of were published by one author. A quick survey indicates that even most TV shows—which have writing staffs—usually have one author, although it's somewhat more common to have several people collaborate as equals to put together a story, which is then written up by one person. This was more or less how Buffy got written, as described by Jane Espenson.

It would certainly have been a major breakthrough if I'd discovered how to have multiple authors consistently work together to make good work. But that's above my pay grade; if a bunch of professional writers who have been in the business for decades have a strong preference for single authorship, I see that as a strong indication that I should generally prefer single authorship.

Also, if this piece comes off as having collaborators mostly making small edits, that's partly because it's true, but partly my own bias. Certainly, in In Fire Forged, we had one or two people who primarily worked with the author on macro level issues (plot, characterization, thematic consistency, etc), while I worked on the micro level. But it's also partly because it's true; outside of two fanfics (plus a poem), I mostly work on nonfiction blog posts. In these, the author knows what they want to say and have said it, and just need to say it better. They may or may not benefit from a fact check (usually not, at least for the pieces I've worked on), but beyond that, most of the room for improvement comes in the form of little changes.

Lastly, I have to thank you. This is the first thing I've actually published. An earlier draft contained a section discussing what I've just said, but I cut it because I didn't think it contained material that was useful to either author or collaborator. Obviously, I was wrong! So, now I have a slightly better sense of when cutting stuff goes too far.