I thought it was well known that the algorithms optimise for stress and anger, because it drives more engagement. I think it might even be involuntarily - the algorithms may have been set to train for enhanced engagement, and hate and anger happen to be the solution.
Definitely, the effect of algorithms gets discussed regularly - but I can't recall encountering any significant discussions about the effects of design (UI, functionality, etc) of social media platforms. As far as I can tell, the idea is almost completely off the radar.
I talk about this all the time. I find it simply insane that, out of the entire design space of possible mechanisms for computer-mediated communication, every major platform has somehow decided that typing comments into small boxes that are indented under other small boxes is the only solution. There are countless dimensions of design variables that haven't even been considered.
> out of the entire design space of possible mechanisms for computer-mediated communication, every major platform has somehow decided that typing comments into small boxes that are indented under other small boxes is the only solution
Now that you say this, I think I now recall hearing that before.
That said, "typing comments into small boxes that are indented under other small boxes" is a very useful/pragmatic feature, but a problem is that this is almost the entirety of what's out there (well, a few others things like upvoting, communities & subcategories, etc). There is so much more that could be done.
Have you put a lot of thought into specifics of what else could be done, what benefits could be realized, etc?
I've had all kinds of ideas. No claim that any particular idea is going to be an improvement, but here are a few one could try:
- Putting rate limits on commenting (perhaps taking effect only after a person has commented in the same thread a few times)
- Reading your comment aloud, back to you, before you can click to post it
- Giving you a field in your profile to explain your communication style and good intentions, which is displayed whenever someone is composing a reply to you
- Requiring you to switch to recording voice or video messages, after you've typed back and forth with someone a few times
- Automatically switching to private messages, after you've typed back and forth with someone a few times
- Simply showing you some general advice on conflict resolution, nonviolent communication, etc. while you're composing your comment
- Asking you what you intend to achieve by participating in the conversation, and then reminding you of your goal
- Asking you how much time you want to spend on the conversation, and then showing it to you
- Requiring each comment to be approved by the author of its immediate parent (so you can disagree with each other, but you have to be kind)
Interesting ideas.....I particularly like these ones:
- Giving you a field in your profile to explain your communication style and good intentions, which is displayed whenever someone is composing a reply to you
- Asking you what you intend to achieve by participating in the conversation, and then reminding you of your goal
- Requiring each comment to be approved by the author of its immediate parent (so you can disagree with each other, but you have to be kind)
A couple I've thought of:
- have a section of the platform that is highly managed by moderators - all topics must be proposed and approved, and the manner/speaking style/community goal/allowed/disallowed behaviors for the conversation are decided before launch, and enforced with an iron fist. (One of my favourite to try: "Epistemic Standard: Extreme (all statements must be truthful)". Oh my, what a learning opportunity this would be for people, including smart people).
- rather than just one up/down vote option, allow voting on numerous other metadata, possibly custom chosen by the moderators or community prior to the launch of the the topic
- custom layouts other than simple nested comments, for example: a ~collaborative "pros vs cons" layout for a given topic - this way, rather than simple disorganized arguing, you get organized arguing, and perhaps some sort of a useful end product (that could be used to feed a subsequent discussion with new goals and rules) - with a proper platform, I think it should be able to treat human minds much like compute nodes, and these nodes even have self-awareness (well, to some limited degree anyways). Ideally, people might some day even learn how to use their own minds (rather than the other way around as it tends to be now).
I've seen criticism of the karma system in general, people doing everything for likes.
Maybe the same thing as for the algorithm applies: turns out hate and anger give the most engagement and in turn the most likes, so people optimize for it.
Another thing, on Twitter it seems obvious that the way mentions and retweets draw other people into discussion facilitate mobbing, but of course they are good for viral growth.
