Postmortem on my Comment Challenge

post by adamShimi · 2020-12-04T14:15:41.679Z · LW · GW · 12 comments

Contents

  Good days can be worse than bad days
  A single chunck is too much for me
  Not enough commenting guidelines
  Start more discussions
  Conclusion
None
12 comments

On the 11th of November, I challenged myself [LW(p) · GW(p)] to comment every frontpage post for the rest of November. The result? Well, I failed [LW(p) · GW(p)] a lot. But I also commented far more than I usually did. And I think from both my successes and my failures, I found some nuggets interesting enough to share with you.

Good days can be worse than bad days

If you had asked me before hand when I expected to fail, my first answer would have been on bad days -- days where I'm overworked, or too busy, or depressed. And indeed, there's probably a couple of such days in my failures. But good days, days where I just relaxed with friends and my girlfriend, are also time sinks, and I usually realized too late that I didn't read the LW posts you wanted to read.

So any commenting habit should probably be safeguarded from both good and bad days, by being either short enough or having a specific scheduled time for it.

A single chunck is too much for me

I began this challenge by choosing an hour each day for reading and commenting every frontpage post. But it was just too big a time commitment to me. Worse, I noticed that because it was that big, it felt tiring, and I pushed it further and further, until it was the last hour of the day and I was not in top form.

My solution was obvious: read and comment one or two posts at a time. I don't forbid myself to go on a LW binge, but it's not a requirement either.

Not enough commenting guidelines

When introducing my challenge [LW · GW], I wrote:

Just a "I really enjoyed this post, even if I had nothing else to say" is useful feedback. So is "I'm not sure why, but I had trouble connecting with the arguments of the post".

That's certainly true for me. But while commenting on other posts, I sometimes felt like this type of feedback was not enough for the post in question. It usually happened with technical posts and AF posts.

The authors of these posts probably appreciate such feedback, if this post [LW(p) · GW(p)] querying LW authors is any indication. But this nagging feeling remains. And I'm pretty sure it's correct in some instances.

What's the solution? Ideally, everyone would explain what kind of feedback they're interested in. And we already have a mechanism for that: moderation guidelines! Maybe it's misguided, but I believe that having a sentence or two clarifying the kind of feedback wanted in your post will tremendously help people to comment. If it becomes a community trend, we will all know where to look for this information, and it will be available for almost every author.

I'm starting it myself, in hope that more will follow. You can find my guideline when you comment, but here it is anyway:

Any feedback (positive or negative) is appreciated, as long as it's respectful and contains at least one sentence about the post itself.

Start more discussions

My reason for challenging myself was that I feel like my posts don't have enough discussion going on in them -- not even discussion telling me how misguided I am. Yet it took me more than a week to realize I was part of the problem: my comments always gave closed feedback instead of starting a discussion. Or put another way, I saw each post as a finite product to judge, instead of a jumping board for discussions.

This brings me to my new commenting habit, replacing this challenge: I will write a comment starting a discussion every day. I don't have to comment everything that way, because I still value pure feedback, and it would make me comment less. But still, I want to make LW a place where I discuss important ideas with smart people, and it's mostly my fault if I don't feel it's completely the case now.

Conclusion

I'm glad I challenged myself that way. I read new posts, found new ideas, and understood a bit more about the difficulties of commenting on LW. I'm interested in any feedback you have, be it on my takeaways or on my flurry of comments this month (Did they annoy you? Were they helpful?)

12 comments

Comments sorted by top scores.

comment by abramdemski · 2020-12-05T01:31:33.584Z · LW(p) · GW(p)

I somehow missed your original challenge post! Great idea.

That's certainly true for me. But while commenting on other posts, I sometimes felt like this type of feedback was not enough for the post in question. It usually happened with technical posts and AF posts.

The authors of these posts probably appreciate such feedback, if this post [LW(p) · GW(p)] querying LW authors is any indication. But this nagging feeling remains. And I'm pretty sure it's correct in some instances.

I can confirm. For my more technical posts, feedback that's just sort of positive or negative without engaging with my questions leaves me feeling <photo of puzzled dog with head sideways>. I mean, positivity can be encouraging, but still.

Replies from: adamShimi
comment by adamShimi · 2020-12-05T19:35:57.390Z · LW(p) · GW(p)

Thanks!

And yes, I thought about that for some of your posts indeed. I think my take on technical posts is that it's possible to engage with them without having anything new to bring to the table. But at least for me, knowing if people generally agree/disagree is still useful.

Although the prediction feature might be more appropriate for that... What do you think? Does replacing the feedback without engagement by a vote on a prediction "You agree with the content of this post" or something more detailed, seems like a good idea to you?

Replies from: abramdemski
comment by abramdemski · 2020-12-06T16:12:29.654Z · LW(p) · GW(p)

Not sure I understand. What's "a vote on a prediction 'you agree with the content of this post'"?

Replies from: adamShimi
comment by adamShimi · 2020-12-06T17:18:27.706Z · LW(p) · GW(p)

I mean the Elicit binary predictions you can embed on LW, as presented in this post [LW · GW]. So making a prediction with "You agree with my take on [topic of the post]", and then letting people predict/vote to give contentless feedback.

Replies from: abramdemski
comment by abramdemski · 2020-12-06T19:05:30.252Z · LW(p) · GW(p)

Ah, I see. I understood that you meant Elicit, but was jarred by the way this isn't really a prediction.

I think it would make sense to offer predictions on object-level issues in a post, rather than the indirect "predict whether you personally agree".

comment by magfrump · 2020-12-06T06:50:39.743Z · LW(p) · GW(p)

Thanks for doing this, and for writing about your experience!

comment by Hazard · 2020-12-05T00:38:57.030Z · LW(p) · GW(p)

Love that you did this and learned something about some of the reasons discussions don't actually get started. I notice that I have often don't comment in a discussion conducing way because I don't enjoy trying to discuss with the time lag normally involved in lw comments. On twitter, I'm very quick to start convos, especially ones that are more speculative. That's partially because if we quickly strike a dead end (it was a bad question, I assumed something incorrect) it feels like no big deal. I'd be more frustrated having a garden path convo like that in LW comments.

Replies from: adamShimi
comment by adamShimi · 2020-12-05T19:48:41.179Z · LW(p) · GW(p)

Agreed that the lag in discussion can be frustrating. Which is why I think we should consider LW conversation closer to an exchange of letters (or emails) than instant messaging. This means having longer messages, with as little need for clarification as possible.

(Not that I'm always doing that, but I'm trying).

comment by Maxwell Peterson (maxwell-peterson) · 2020-12-05T00:01:14.422Z · LW(p) · GW(p)

I liked your comments on my posts.

Replies from: adamShimi
comment by adamShimi · 2020-12-05T19:37:10.326Z · LW(p) · GW(p)

I'm glad! It was actually pretty fun to look for the measure definitions and examples for your post [LW · GW] on the subject. ^^

comment by supposedlyfun · 2020-12-04T22:36:23.241Z · LW(p) · GW(p)

I liked this series of posts for its attempt to identify a community problem, develop a potential solution, and describe one attempt.  I would be interested in more details about the various failure modes because I suspect they are relevant to anyone who sets any kind of productivity/focus goal.

Replies from: adamShimi
comment by adamShimi · 2020-12-05T19:51:45.275Z · LW(p) · GW(p)

Thanks!

Not sure how relevant my failure modes are, but here's what comes to mind after rereading my list [LW(p) · GW(p)] of failures:

  • Not being interested enough in the topic of the post to leave an honest and useful comment
  • Not being competent enough in the topic of the post to leave an honest and useful comment
  • Feeling too tired
  • Having too many posts to comment in one day.