Ever since I got involved in emotional well-being and mental health efforts (ERGs) in BigTech some years ago, I’ve been interested in what might be pretty niche (inter)personal challenges. And more recently, as a moderator in a Discourse forum, I’ve also come across similar questions/thoughts. Let’s call it “episodic unwellness for whatever reason”.
If some John Doe has good days and bad days, and has a habit of embarrassing themselves on bad days and then regretting it on good days, what could a good-day-John do to reduce the embarrassment they might need to suffer at the hands of bad-day-John’s actions?
A moderation team willing to deal with this can go a long way, and might be sufficient for the majority of Discourse’s users… but a moderation team might also say “we just don’t have the resources for this”.
Suppose good-day-John would like to carry as much responsibility as possible for bad-day-John’s actions, he might be interested in setting up a rule like “all posts by John should be (re)approved by John at least 24 hours later”. This, for example, could avoid problems with bad episodes that last less than a day – sampling John’s decisions on two different days could ensure a more grounded or sober good-day-John always has a say before bad-day-John gets to post anything. The cost: all of good-day-John’s postings would be delayed by a day (in this example). Good for first posts in new threads, but for replies to existing threads, badly hindering timely participation in discussions that are lively.
A related idea: giving John the option to designate some friends who are willing to approve his posts. This adds more ACL complexity, but potentially reduces the time-delay problem. (“Friends or moderators or future John could approve posts by John.”) Friends getting involved might be necessary for episodic problems that could last days or weeks.
It could be self-imposed; it could be moderation-team-imposed on users with repeated episodes. How niche would this be? I currently know of only one person who would be interested in self-imposing this kind of thing, but I’m on the lookout for more…
I’m consequently not sure this is worth the effort, but I nevertheless wanted to put it out there (e.g. in case it’s much easier to implement than I think).
That’s a very interesting topic. This kind of “write and regret” thing also happens in real life to pretty much anyone after all I guess (and can sometimes have dramatic outcomes!), and knowing how and when to put safeguards can be challenging.
A quick example of how I deal with this kind of thing in real life: I know that I might sometimes overshare or write overly (and unwanted) emotional things, whether it’s in public or private spaces.
Since I know it happens, when I want to write something that leans into this area, I (most of the time, nobody’s perfect) forbid myself to send those messages/posts from the evening to night, because I know the accumulated fatigue of the day can influence in the wrong direction what I want to communicate.
So, I write it somewhere (Google Doc, notepad…), but I don’t publish it. I wait until the next day and read it in the morning/afternoon. Then, I can decide if I post, or if I edit and then post, or if I don’t post at all.
I’d be interested in seeing a feature as you described, and believe it can be interesting to imagine how it could work, especially in health-related community spaces.
Sorry that my post doesn’t add anything tangible to what you’re asking. Take it mostly as an emphasized “like”
Time for some more thoughts: what we already have in Discourse, for post-hoc self-serve cleanup of regrettable comments, is the ability to edit comments and posts.
Replies can be deleted too, not just edited. Initial posts are trickier than replies: can initial posts in a thread be deleted? If so, the replies to a thread would be lost. Hence benefit in being able to edit the initial post, to not delete a thread others have participated in.
Time limits on delete & edit
I think the default time limit is 30 days? That should usually be adequate for more common sources of episodic disinhibitions: someone got drunk and posted bad stuff… There are different kinds of alcoholism, I don’t know how many of them end up having several days in a row of drunken disinhibition.
Then, manic episodes: they can be short, they can be long… even as long as a month or two. And a manic episode could be followed by a depressive episode, for people leaning in the bipolar direction. I could imagine that depressive episodes might also be debilitating, delaying how long it takes for someone to be able to face the consequences of a manic episode, and go clean up any mess.
My guess is, 30 days is enough for “almost everyone”; whereas significantly long manic episodes are hopefully rare enough that moderators could get involved and help out with any cleanup, if it’s also only happening once or twice (whereas they may tire of it if it keeps on happening again and again.)
I have certainly seen people lean on friends or closer connections as sounding boards about things they are considering posting first.
Within the context of Discourse, this might look like sending a draft of the post in a personal message or in a direct message on chat.
Another idea coming to mind based on your ideation here is some kind of LLM based just-in-time feedback system – “you sure you want to post this?”
I don’t know how I really feel about that, but it’s hard to avoid thinking about where AI might fit into the picture these days. Perhaps if it were opt-in, it’d address this need though?
