Brief comment on featured

post by Ben Pace (Benito) · 2017-10-29T17:12:51.320Z · score: 33 (12 votes) · LW · GW · 1 comments

Someone asked me a meta question in the object level of this site (i.e. the frontpage), so I'm copying my reply here instead of continuing discussion there.

Me (on Chapter 1 of Inadequate Equilibria):

I've promoted this to Featured because it's a great piece of writing, and also because this book seems to be about some of the fundamental epistemological disagreements in our broad communities. I really hope this is the place we can communicate clearly and successfully together on these topics.

Lucretious:

What are the grounds for relying on individual judgment to promote posts to Featured? Isn't it more reasonable to rely on community opinion, as expressed in the epistocratic karma system?

(I'm aware that this is a tangent, but the point seemed germane given that the reasoning behind this decision may itself reflect the "fundamental epistemological disagreements" which you refer to.)

Me:

Yup, it's a tangent and best suited for Meta, but seems like a fine question, and I'll respond to it here for now [edit: well, now I've moved it to Meta].

My current plan is to have karma do a lot of work in sorting, but for the moderators (Sunshine Regiment) to also have a lot of say on what goes to Featured. I often think of karma as the System 1 of the site, and the mods as the System 2 - while S1 makes most decisions and sets up basic incentive gradients, S2 tries to give useful input on important decisions, make plans and route around obstacles. I expect the S1 to be imperfect - as an example, I imagine some posts will be undervalued that dont' contain many new ideas but summarise a great deal of previous discussion well, so that people can read a single post instead of 5 posts and 3 long comment threads. Recently I Featured one or two posts in part for being no longer than they needed to be, which I want authors to know is great, but I also don't know that the karma system is incentivising as much.

A related key point is that only the S2 can be easily queried for its reasons - if you ask the karma system why it promoted something, it can't tell you, and so authors have to speculate on why their posts got into Featured. I'm very pro transparency.

In general while I think the karma system is very useful and I take guidance from it a lot, I want the key decisions on the site not to be made by an algorithm whose parameters we do not know and cannot easily query (I mean, I can imagine having everyone on the site vote on certain questions to get answers about particular decisions, but this is very costly and can't be done very regularly).

(Also, at the minute the moderation team is quite quiet, and what's been promoted to Featured has largely been a function of my thoughts and the people I talk to most - Ray and Oliver. This is mostly because we haven't given the mods most of their tools yet, but I expect when the team is more active and a wider section of the community is adding posts to Featured, this will allow many theories of what is valuable to flourish and coexist - like the karma system itself, only more transparent.)

Added: I like people to know the norms so I'll leave this here for now, but if folks would like to discuss this Meta topic more, totally make a Meta post about it and discuss in that section of the site.

1 comments

Comments sorted by top scores.

comment by ChristianKl · 2018-01-10T01:22:09.247Z · score: 5 (1 votes) · LW · GW

Practically I don't think it's the case that opinions of the karma system are completely opage. Discussions about karma scores were frequently a part of post and do allow to come to shared opinions.