Building Mod-Bot: An experimental community moderation tool
Introducing the mod-bot community moderation tool. Comment for a share of the rewards about whether you'd use the tool, and any ideas you have.
Hello everyone! Over the course of the last month, things have been hectic. I had a family emergency to address, and didn't have a lot of time to work on things Steem related. But my dad (@remlaps) and I have come up with a new idea which we have now implemented in his Popular Stem community. That idea originally was for an account which monitors promoted posts in a community, and pins them either based on order, or based on weighted pooling. Little did I know when @remlaps suggested that original idea, how much this tool could potentially do for the community experience on Steem. So I will lay out where the project currently is, and some of the ideas I have had about where we can take this project.
Mod-Bot: An experimental community moderation tool
What do we have so far?
Right now, I have mod-bot looking at several different parameters (in regards to pinned posts) for the communities it monitors (which right now is just the popular stem community). The most important of these are promoted-ordered, promoted-pooling, burn-ordered, and burn-pooling. For each of these 4 parameters, a community owner can put a number, and that many posts will be selected for pinning. I.E. If a community owner wants the top two promoted posts pinned and 2 promoted posts pinned based on random selection of a pool (which is weighted in proportion to promotion amount), he would say promoted-ordered:2 and promoted-pooling:2. When the bot checks the promoted posts, the 2 posts that are promoted the most will always be pinned, and the other 2 slots will be rotated randomly based on the other promoted posts. The concept is the same for burned posts.
The cool thing about this tool is that it will pair nicely with the browser extension @remlaps is working on which colors in burned and promoted posts.
In creating this, I made it so the bot checks if it was the original pinner of a post before unpinning it. This means that posts pinned by other community moderators will not be interacted with during the pinning process. This also means that the number of pinned posts will be in addition to whatever posts are already pinned by other accounts.
There are other parameters a community can use which are:
- A minimum and maximum amount of words required to pin a post
- A minimum and maximum promoted amount
- A minimum and maximum reputation for an author of a pinned post
- A required language for a post to be in to be pinned
- Whether a post has to be powered up 100 percent to be pinned
- Possibly more in the future
My father has not used any of them, but I'm sure other community owners would want to. For instance, I figure a community like the New Comer's community might not want to pin posts from accounts above a certain reputation. And a Spanish community may not want to allow for non-spanish posts to be pinned.
I've also added a system which monitors the account's mentions and follows commands if the commenter is a moderator or higher in the community. I.E. Blacklisting posts from being pinned and blacklisting authors from being pinned. I will keep the format for commands secret until we are ready for use by other community owners. I don't want a million people spamming the bot.
What do I plan to do?
I plan to add features for muting posts. Most communities have certain criteria pre-established. I.E. This is a community for posts in Spanish, or this community requires posts with a minimum of x words. Etc.
This system will be designed to have specific parameters for posts to be muted. Such as if a post isn't this many words, it will be muted. Or if posts aren't in Spanish mute them. This will eliminate the posts that don't meet the community's established standards which are easily monitorable for a bot. This concept also raises some other cool ideas:
something like a members only community. I.E. The community owner has a list of people who are allowed to post in the community, and anyone else who posts will be muted. This would be cool for something like a hypothetical Steem witnesses community. Anyone can comment on their posts, but all of the original posts in the community come from the witnesses. So if you want to look for official posts from that team (or any other team on Steem), you just look in the community. No one would have to sit there muting posts not from those authors, and it would create a good place to go to see those people's posts.
Similarly, this would be a cool concept for a community of only burned posts. While the burnsteem25 tag exists, people who aren't actually burning steem at 25 percent could still post with that tag, and you have to look into it to see if they are actually burning at 25 percent. With this tool, a community could be set up where all of the posts shared that are unmuted have to be burning at least 25 percent of their rewards.
Another possibility that excites me is the potential to make it so posts in a community require a community account to receive a certain amount beneficiary awards. That could really make it much more rewarding to run a successful community. So I'm definitely considering running that for communities that are interested.
There are many other concepts that this tool can be used for, and I'm sure that as community owners discover this tool, they will suggest even more concepts that @remlaps and I can't think of.
How can you get this tool for your community?
For right now, @remlaps and I are going to finish getting the service up and running for his community. Eventually we will likely run a beta period where the service will be free for those who want it while we work out the kinks and get the service working on a larger scale, and eventually we will make a post discussing how to sign up for the service. For right now, feel free to comment down below with any ideas, and if you'd be interested in using this service for your community. I will be making it so that 20 percent of the article's rewards go towards fueling that discussion using @penny4thoughts. Remember that in order to receive a share, your comment has to be meaningful and relevant.
Anyway, have a nice weekend everyone!
Now I have taken the time to read your post a third time and I think I have understood everything (except the technical implementation of course)... ;-)
I admire your idea and how many details you think about. If such a bot works: Kudos!
Hm, maybe the tool is useful for some communities (better: their admins). It always depends on how you use the AI and how much personal energy you put into looking after and controlling it.
Here is my personal answer to your question whether someone would use this bot as admin of a community.
As fascinating as I find the possibilities and technical implementation as well as your competence - BUT, no, I would not use the bot.
Certain automatisms are okay and useful for me. But not for the evaluation of content via votes. For example, I would only follow a trail if I knew that the starting account really reads carefully and votes in my favour. To be honest, various - even very established - trails have already been eliminated for me. Chiller's "Follow Voter Function" is much better if you don't have time to vote. Here you can call up the votes of a user you trust and simply copy them (you can also omit individual votes if you don't agree with them).
Of course, your tool checks a lot of aspects, but in my opinion only a person can decide about the real quality (of course, this is often subjective) of a content.
I even think it's a bit "dangerous" to bring this automatism into connection with "promoted". To end up in this category, a contribution should be really good (after all, it should also be advertising for the Steem and set itself apart from all the "plagiarism and boring diaries"). Promising free slots on the basis of automatically read-out parameters seems to me like a lottery in which the cleverest or richest tricker can win again... :-(
The topic of automation is a whole big conversation, so I'll maybe write my thoughts on that in a future post. Short version, my opinion is that all tools can be used well or used badly, and that automation is just a tool that will get better with time. I agree that people will always seek ways to game the current iteration, but to me the solution to that problem is to make the next version better.
I'm not trying to change your mind, but just for clarity, mod-bot isn't actually voting for any posts. For now, it's just pinning and unpinning them - in keeping with my original idea to use post promotion for audience building. Also, there's a way to prevent pinning if a moderator finds that a post is plagiarized or spammy or whatever. The next phase is probably muting posts that don't meet a community's rules (although @cmp2020 is back at school now, so things might be slow). In the end, the moderator will have ultimate (manual) control over both pinning and muting.
Of course. With a lot of components to consider. I have already read many posts or comments from you regarding the automation of various processes on the Steem. Our opinions often diverge somewhat, but never completely. I don't demonise automation at all (especially since it simplifies a lot). But it has to be constantly maintained by real people.
E.g. the former plagiarism bot @cheetah. I liked him a lot... ;-) But at some point he started to strike as soon as he discovered a repetitive footnote. When I pointed this out, its operators did not remove the automatic downvotes. Worse still, the users ended up on a blacklist in case of repetition and had no chance of rehabilitation because the people responsible no longer cared about the bot.
I know @cheetah doesn't really fit the topic now, I just wanted to get away from the vote bot.... ;-)
Yes, I understood that. By pinning the articles (and thus making them important), he influences other voters indirectly at best (as your son already said). That's what promotion is for, advertising.
And here, I think, the selection has to be careful. Go beyond parameters and bring human considerations into play. But I'm certainly not accusing you/your son of not doing the latter!
That should be out of the question anyway. But for that, the mod has to control the articles. In that case, he could just pin them by hand... ;-)
This mod bot is going to be great and I think it's super cool that you're programming something like this! Try it out, evaluate it, declare it perfect or improve it. The best prerequisites... :-))
I reserve the right to decide against its use in my community at the moment - that's what @cmp2020 asked about.
I have a few "time problems" regarding my activity on the Steem. But I promise to keep this comment open and answer it later this week!
First of all, this project does not use any AI technology (it's basically just checking to see what the community's parameters either for pinning or posting are, and if the post fits them). Just clarifying that. Second of all, either way you will need to put energy into controlling something. If you are putting energy into controlling the bot, the bot will be saving you energy from having to filter the baseline violations in your community that are easy to automate.
The first thing I would point out in response to this: as @remlaps already addressed, this service would not be about voting. It's not a curation trail, and will not have much directly to do with curation. It may be indirectly related to curation because it may allow for curators to see posts that fit certain criteria that the community owner sets up which is appealing for the curator, but this project is not directly looking to curate.
At it's core, the pinning part of this project is about incentivizing post promotion and burning rewards. Both of these things are ways to improve the price of steem/sbd by decreasing the amount of currency that is being added to the market. Both of these things reflect value. Burning SBD for a post means you think that post will make more than the money you are burning. Burning a portion of the rewards shows that you value reducing the amount of steem being produced (and therefore improving the price of steem) more than you value the amount of steem you are burning.
To end up in the promoted category, one must promote a post. That is done at the blockchain level. So this product is completely unrelated to whether or not posts get promoted. What this product does is cycle through the posts that have been promoted to give them visibility. Like I said, my father and I thought this is useful because to promote a post, one effectively has to burn SBD. Similarly, burning a post means to burn the steem that post would have earned in rewards. Both of these things will improve the price of Steem.
Keep in mind that this and the other features in regards to pinning/muting are optional, and that one could solely use the pin part of the project or the mute part. They are a way of setting a baseline standard for a community either for pinning or simply posting in a community. If I as a community owner want the community to only have posts that are longer, I can set that up. And if I as an owner want shorter posts (like in the case of the steem links community), I can set that up. But the parameters would be up to the owner and subject to change at any time. Like my father said, mod-bot has commands so the community owner can make changes quickly if something goes wrong.
Overall, I think this product could really enhance Steem. For one thing, community owners won't have to do as much work to shape the community to be the way they want it to be. That's what automation means. Of course you are right that there wouldn't be as much freedom in judging content because the owner is not judging every article directly, but abstraction is a part of the world we live in, and a necessary one. Imagine if you chose to read Steem posts at the blockchain level instead of on an interface because there is more information conveyed at the blockchain level. Yes, by putting it on an interface, some information is inevitably filtered out by that interface, but the information displayed is much more pretty and readable and presented in a way that enhances the experience.
There's perks to doing things yourself as the owner, but it's also a lot more work. And sometimes you are inevitably doing something that's easily automated, and therefore wasting your own time for the freedom to judge post's individually on a case by case basis. If that is what you are looking to do, then this product may not be useful, but if you find yourself muting the same kinds of posts, or want to ensure the same standard for all the posts in your community, this product might be useful.
I have a few "time problems" regarding my activity on the Steem. But I promise to keep this comment open and answer it later this week!
Sorry, one more little addition using #penny4thouhgts as an example.
I love your dad's idea! #penny4thoughts is brilliant! And also this fantastic tool unfortunately gives possibilities for abuse. If you look at the hashtag, you see a user who publishes there two or three times a day with cheap posts. On the one hand, he is sure to get the vote of @penny4thoughts. On the other hand, he comments and votes himself with his second account in order to receive the bonus payment. Oh, yes, he's his only commenter.... ;-)
Then there are the users who specifically search for #penny4thoughts articles and leave some inappropriate comment hoping to get a few crumbs of the cake. For me that's shabby.
Conclusion: Bought engagement is very questionable for me, the spirit of Steem as it once was is getting more and more lost through automation.
I agree that penny4thoughts is a groundbreaking and brilliant idea. I think it should be a fundamental concept to Steem that, in an ideal world, would be optionally done by the blockchain itself.
All of steem is "bought engagement" to some extent. Most people are here for author rewards at this point (Lord knows they're not here for the ground breaking user interface or recommendations systems. Two things I'd like to see improved so Steem is less bought engagement). My father just made a different way of rewarding (by redistributing rewards) a specific type of engagement: that being discussion.
I have a few "time problems" regarding my activity on the Steem. But I promise to keep this comment open and answer it later this week!
With the necessary changes to curb exploitation, I thought this suggestion was great too!
This would make many users realize that interacting on this platform is incredibly important and that it can even be a little lucrative. Then "everyone" wouldn't feel "forced" to post articles at all costs, no matter how poor their content. Maybe the latter statement is a little unfair, but that's how it often feels with the "new generation" of Steemit users, unfortunately.
Respect @chriddi for both valuable comments and for sharing your thoughts.
May I ask you what is the idea behind this project? Or direct me to some post about it? I would appreciate it
Cheers Piotr
Thank you... 😊
Remlaps developed the project some time ago (I remember it's been active before the big fork) and I was not able to find its introducing post.
Maybe this dialogue might help you a bit:
https://steemit.com/games/@chriddi/rfw32z
Lucky 5"
Curated By - @ubongudofot
Curation Team - Team 4 Curators
Dankeschön... ;-)
Hi @cmp2020
I must admit that idea of you creating such a bot does sound interesting. Im not sure if I would like to automatize things (as a community founder) as I believe, that doing those activities manually is crutial for person who moderate community.
May I ask you what is the idea behind this @penny4thoughts? Or direct me to some post about it? I would appreciate it
Cheers Piotr
Fair enough. I respect you for wanting to do things yourself. I do think this product has it's uses in making your job easier, but it's all up to you.
I never actually published anything about it because I don't want it to grow faster than I can keep up (it's been running for a long time, but I'm just a weekend warrior. ;-)
The idea behind it was I thought that reward sharing would be a way for small stakes users (with limited upvote power) to compete with larger stakeholders for engagement. So, even if I can't give a $0.25 upvote to someone who replies to my post, with quality content and reward sharing, I might be able to give a STEEM or two from my author rewards to the people who reply to my post. This comment that @chriddi already linked to has a basic description of how it works.
This post has been featured in the latest edition of Steem News...
Well , it is aimed at the communities , according to what I have read they have an order , many aspects have been taken into account , I hope that some communities are interested
woow