Proposal: Systematic Search for Useful Ideas

by Nic_Smith2 min read1st Jun 20115 comments


Personal Blog

LessWrong is a font of good ideas, but the topics and interests usually expressed and explored here tend to cluster over few areas. As such, high-value topics may still be present for the community in other fields which can be systematically explored, rather than waiting for a random encounter. Additionally, there seems to be interest here in examining a wider variety of topics. In order to do this, I suggest creating a community list of areas to look into (besides the usual AI, Cog Sci, Comp Sci, Econ, Math, Philosophy, Psych, Statistics, etc.) and then reading a bit on the basics of these fields. In additional to potentially uncovering useful ideas per se, this also might offer the opportunity to populate the textbooks resource list and engage in not-random acts of scholarship.

Everyone Split Up, There’s a A Lot of Ideosphere to Cover

A rough sketch of how I think the project will work follows. I’ll be proceeding with this and tackling at least one or two subjects as long as there’s at least a few other people interested in working on it too.

Step 1, Community Evaluation: Using All Our Ideas or similar, generate a list of fields to investigate.
Step 2, Sign-Up: People have the best sense of what they already know and their abilities, so at this point anyone that wants to can pick a subject that’s best for them to look into.
Step 3, Study: I imagine this will mostly involve self-directed reading of a handful of texts, watching some online videos, and maybe calling up one or two people -- in other words, nothing too dramatic. If a vein of something interesting is found, it’s probably better that it’s “marked” for further follow-up rather than further examined alone.
Step 4, Post: Some these investigations will not reveal anything -- that’s actually a good thing (explained below); for these, a short “Looked into it, nothing here” sort of comment should suffice. Subjects with bigger findings should get bigger, more detailed comments/posts.

Evaluation of Proposal

As a first step, I’ll use a variation of the Heilmeier questions which is an (admittedly idiosyncratic) mix of the original version and gregv’s enhanced version.

  • What are you trying to do? Articulate your objectives using absolutely no jargon.
    Produce comments or posts providing very brief overviews of fields of knowledge, not previously discussed here, with notes pertaining to Less Wrong topics and interests.
  • Who cares? How many people will benefit?
    This post is partially an attempt to determine that, but there seems to be at least some interest in more variety on the site (see above). Additionally, the posts should be a good general resource for anyone that stumbles across them, and might even make good content for search purposes.
  • Why hasn't someone already solved this problem? What makes you think what stopped them won't stop you?
    The idea is roughly book club meets Wikipedia, but with an emphasis on creating a small evaluative body of knowledge rather than a massive descriptive encyclopedia, and with a LessWrong twist. The sharper focus should make the results more useful to go through than just hitting “random page” in yon encyclopedia.
  • How much have projects like this cost (time equivalent)?
    Some have the ability to take on “whole fields of knowledge in mere weeks” but that’s not typical -- investigating a subject in this case is roughly comparable in complexity to taking an introductory class or two, which people without any previous training normally accomplish over a period of about three to four months at a pace which is not especially strenuous, and with fairly light monetary costs beyond tuition/fees (which aren't applicable here).
  • What are the midterm and final "exams" to check for success?
    For each individual investigation, a good “midterm” check would be for the person looking into a field to have an list of resources or texts they’re working on. The final “exam” is a posting indicating if anything useful or interesting was found, and if so, what.
  • If y [this community search] fails to solve x [uncover useful knowledge in fields previously under-examined on LessWrong], what would that teach you that you (hopefully) didn't know at the beginning?
    Quite possibly, this could be a good thing -- it indicates that the mix of topics on LessWrong is approximately right, and things can continue on. In this case, we’d end up seeing a bunch of short “nothing interesting here” comments, and can rest more or less assured that further investigation into even more minute detail in unnecessarily. This is conditional on not-terrible scholarship and a reasonably good priority list from step 1.
Personal Blog