While I'm not against LW wiki itself (it already exists, for starters), I'm very much against making LW “something like a wiki”, because I'm >50% confident it will fail. I flinched when I read “community-maintained wiki pages with explanations and links” in the original post, because “community-maintained wiki” are almost universally dead before reaching maturity.

Michael Snoyman wrote a small article on why people are willing to contribute to free software documentation via pull requests, but not via wiki edits. I wholeheartedly recommend everyone to read the article, but the gist is as follows.

For a wiki:

• maintainers think they are encouraging the community to write documentation
• contributors are intimidated by the wiki, because they are afraid they aren't justified in editing it
• readers rightly expect incomplete, unstructured, and messy information.

For documentation that is improved through pull requests:

• maintainers deal with documentation in atomic fashion using tools they know
• contributors don't worry about inadvertently doing harm, because their contributions are checked by the maintainers
• readers know that the information is canonical, because somebody reviewed the contribution before publishing it.

Why LW-as-a-wiki would discourage contributing (writing wiki-like articles)? Of all wikis I remember, the only successful are Wikipedia and very narrow-focused wikis (e.g. UESP for The Elder Scrolls or Ring of Brodgar for Haven and Hearth video games). In both cases they are thriving because there are very clear expectations of what a final article is supposed to look like.

LW is far away from being definitive canonical reference, which is good. Every rationality-relevant topic could be explained from different perspectives, so I would hate there to be the one definitive article on, say, control theory.

Then you'd have all the Wikipedia problems: edit wars, deletionism, constant arguing over the rules and article layout, slowly corrupting powers of wiki moderators, censorship. On a wiki everything is supposed to be canonical, so much effort will be wasted on arguing over canonical definitions and phrasings, or on referencing more and more rules and guidelines. Wiki model has bad incentives: wins the one, who is more stubborn.

LW-as-a-wiki would stagnate very quickly, as there will be huge psychological and social obstacles for people to contribute. I will go into these obstacles in greater detail in follow-up comments. For now I want to say that we should analyze what is wrong with the wiki model from cognitive psychology and science of human motivation perspective, and see how we can do better.

The most important revolutionary idea behind LW (and more specifically lukeprog era LW) is that science is a superweapon, and if diligently learn relevant science and then try to fix the problem, you can outdo your competitors by a large margin (see also: beating the averages). So maybe we should figure out psychology of motivation, incentives for contributing, that kinds of things, before patching LW codebase. Maybe LW should be a community blog, a Reddit-style site, a wiki; or maybe it should be something completely different.