Disclaimer: Please read the full post before commenting. This post will be refined overtime, so I highly encourage to check back if there are any updates!
Hey everyone!
My name is Francis and I am a DEV user for 6 months. There is something I want to reveal...
I am a Trusted Member AND a Tag Moderator for the #discuss (https://dev.to/t/discuss). With both, I am a DEV Community Mod! :D
My goal as a DEV Community Moderator is to ensure the community abides DEV's Mission and their Terms of Service when it comes to implementing best practices to writing articles and interacting with the community.
A guide for community interaction and getting started on DEV can be found here:
The Goal of this Post!
I want to create a thread that allow users to ask questions and get a response in a timely manner. As a DEV Mod, I want to ensure you have your questions answered from a Mod perspective.
@ben, @jess and the rest of the DEV Team are extremely busy, especially with the acquisition of Major League Hacking.
A lot is happening behind the scenes at DEV and I believe I want to take the responsibility to answers questions and concerns as we are moving forward. It will save valuable time for the DEV team!
If you have questions about:
- AI usage on DEV and ways I can improve my Moderation.
- Questions on how to use DEV.
- What a Community Mod can and cannot do.
- Other fun Icebreakers (Love to chat!).
Note: Any questions related to how DEV and MLH team operates internally can be contacted through to the DEV team email: yo@dev.to
I would love to hear from you!
Q/A
The article is based on these key factors: User activity, reading the article in its entirety, running via GPTzero. No favoritism is involve in the decision making whether the post is consider high or low quality. If you believe there is a mistake or that you notice something different when posting content on DEV (such as visibility and Sloan Messages), let me know! How do you determine the Article's Quality?
Nah. It's a bot that Community Moderators control. Note that Tag Moderators exclusively do not have this power. Is "Sloan" a real person?
No. There are others out there like me. According to @jess, I am the most active Moderator currently. Note that Moderation changes will be implemented in the future. I recommend revisiting this post for updates! Hey all, just wanted to share that we'll be updating our general moderation guidelines and tooling for the community soon. A lot goes on behind-the-scenes that's a combination of manual moderation and automated spam detection. There is a lot of nuance here that requires constant adjustment and rebalancing. I'll note that there are hundreds of mods with the ability to send messages via Sloan, so not all of it has fallen on Francis's shoulders - though we deeply appreciate his dedication to the the community, and he is by far one of our most active volunteer moderators. We'll share a more substantial update on community moderation and our perspective on AI-assisted content in the next few weeks. Are you the only Moderator?
Comment on Ask a DEV Community Mod! 🚀
Jess Lee
Jun 16
Top comments (68)
Hey all, just wanted to share that we'll be updating our general moderation guidelines and tooling for the community soon.
A lot goes on behind-the-scenes that's a combination of manual moderation and automated spam detection. There is a lot of nuance here that requires constant adjustment and rebalancing.
I'll note that there are hundreds of mods with the ability to send messages via Sloan, so not all of it has fallen on Francis's shoulders - though we deeply appreciate his dedication to the the community, and he is by far one of our most active volunteer moderators.
We'll share a more substantial update on community moderation and our perspective on AI-assisted content in the next few weeks.
This is a big priority for us going forward after lots of work to catch up with new demands of moderation in this age. I think we'll land in a really good place.
Thanks @jess and @ben! If you believe if there is something I can do to improve, do let me know in the meantime!
Thank you, as someone with next to no creative writing skills 😅 I beg for leniency when it comes to technical posts, so I dont have to spend hours writing a MD table for the benchmark results and run Grammarly to make sure it makes sense. Never in comments though, just in my posts. It would be a mess if writing a post takes me longer than developing the thing does.
Do keep up the good work, I did mention to Francis a tool I saw 4 years ago on LTT for YT comments, I looked into it a bit, last update was about a year ago, but I'll fork it and spin up the foundry to expand it into a fake-follower detector too and let y'all know when it's ready if it helps?
Hi thanks for opening this up I have a question about AI use and disclosure.
I've been writing on DEV for about 6 months My typical workflow is I use AI for about 10% of the writing mostly for structure, research, and organizing thoughts. The remaining 90% is my own voice, my experiences, my tone.
Sometimes I add a disclosure at the bottom of my posts (e.g I used AI to help structure this post. The experiences and opinions are my own.).
My question is - is disclosure required even for minimal AI use like this 10%?
I want to make sure I'm following the guidelines correctly and not risking any flags or suspension. I've seen some people posts get flagged recently and I want to stay on the right side of the rules.
Thanks for your time.
Hey Harsh! Thanks for reaching out and clarifying! It depends.
There is cases where I saw that a person will write in their own words and they use AI to fix the grammar and such. The problem becomes when they start re prompting it to make it more refine to the point it is no longer their words, which will become fully AI.
With that said, I wouldn't worry too much. For example, @sylwia-lask uses both her own words and uses AI to fix grammar and other stuff, which she did not get flagged for. It is both used as a mix of both, it is a good sign to me. If it is near 100% fully AI, it is an issue.
Hope this makes sense Harsh! If I made an error on your posts, do let me know and I can rereview it! I try my best and even I make mistakes, which is why I made this post as transparency and being able to reach out easier than waiting for Jess and Ben to respond since they are busy. :D
Thanks for the detailed response this is super helpful I appreciate the clarity on the 10% vs 100% distinction, and the reminder that disclosure is still the safest approach I Will keep doing what I've been doing use AI lightly for structure and research, keep my own voice for the substance, and always add a disclosure at the bottom.
And thanks for mentioning Sylwia as an example good to know that mixing AI assistance with your own voice is both acceptable and common.
Really appreciate you taking the time to explain this. 🙌
To add on, either way, it is best practice to do so just in case. I have to do citations on using AI when I code at my University. Absurd? Probably, but it made sense since it is the same as looking up sources to write your paper. Same thing different context.
I feel like I got it entirely wrong 😅. I honestly thought Sloan was an actual person. When I got a Sloan message about a formatting issue on one of my posts, I assumed you had asked Sloan to comment under my post. Since you mentioned in the P.S. of your comment that you had sent a Sloan message, I completely misread that part. My bad 😅
I didn't realize Sloan was more of a moderation system/persona and that the moderator was actually the one sending the message. Thanks for clearing that up!
Yea for sure! I knew Sloan was a bot that Mod can control. It would be crazy if Sloan just say "Hi" and go from there. I would be scared for my life lmao.
I do appreciate it. It seems that people are open after the Sloan message, which is good!
MAN, I THOUGHT SLOAN WAS A PERSON TOO!!! 🤯🤯
B r u h
Based on the Moderation page (dev.to/community-moderation), 'Community moderator' is a term used for multiple types of moderators (ike tag moderators).
However i've been a tag moderator for years and I feel like your 'Community Moderator' role cover way more responsibilities. Is it right?
Hey Gérôme! Hope you are well. That is correct. I want to note that the role "Trusted Member" covers more base as a moderator than Tag Moderation. I say Community Moderator because I cover both bases as a Trusted Member and Tag Moderator. Let me know if anything else :D
Ok I see, just discovered Sloan was a specific Trusted Member feature, I felt like I missed something.
That is interesting to hear. I would assume if you get one or the other, you will get Sloan. Was wondering what Tag Moderation you are covering?
the 'angular' one. Sloan is mentioned only for the Trusted Member role in the moderation docs.
I'll ask the first question. How many Sloan warnings does it take before an account gets flagged? And will there be a notification when that happens?
Great first question! and appreciate the communication we had on @dannwaneri's post for this concern.
In my opinion this depends on the user's activity and how they are using DEV. For example, let's say a person has never been caught but their profile shows the following:
If this was the case, it will be either a single warning message before flag or flagged the account entirely.
However, if it was a genuine mistake and did not know before hand, I usually go with a few warnings before hand. Of course, I make mistakes and it could vary.
There are other deciding factors based on how long they are on DEV and other things, but those are the most common.
I never been Flagged from Sloan before, but I would assume it would be like any other comments (though sometimes dev.to email notification is slow. One time I got an email notification that someone commented on my post 3 days ago lol). Can you clarify to me?
Thanks :D
One more thing — my article "I Spent 3 Months Training An AI. My VP 'Reallocated' It. Then I Got Two Calls At 1 AM." (dev.to/xulingfeng/i-spent-3-months...) still can't be found in search. Could you take a look? The latest article is fine now.🤣
Yea I did and it wasn't there...
I would recommend re posting it. Nothing on the article is flagged or anything I have done on my end. Sometimes, articles doesn't popup. Happens to me as well and not sure why.
Feel free to repost and let me know if anything!
Got it, I'll republish and test it. I'll let you know how it goes either way.
Update — it worked! I didn't delete the article. I just edited it — literally removed one empty line and saved — and now it shows up in both the feed and search. So a simple re-save re-triggers the search index. Thanks for the help! 🙌@francistrdev
Yea no problem. Good to note in the future for me lol. Any moderation issues let me know! Do also disclose the use of AI if you are using AI (I don't want to come back and run them Flags again lol)
Don't worry — I've been putting disclosure on everything since our last discussion. We're good! 🤣
Sounds good!
Email notifications from DEV have always been prompt for me. But I never received any notification about my account being flagged.
My suggestion: add a system-level notification in the community's notification center to inform users if there's anything unusual with their account and how to improve it — if that's possible.
@ben and @jess ☝️
Two legends in a one frame
lol
Appreciate you writing this up Francis. And glad to see Jess and Ben weighing in . looking forward to the updated guidelines...
Yea no problem Daniel! I hope the Sloan messages weren't affecting you in a negative way. My only request is to implement best practices in writing and sourcing when necessary. Let me know if anything else!
Great insights! It's always interesting to hear about the experiences and challenges that come with being a community moderator. Thanks for sharing your perspective.
One question: What's the most common mistake new community members make, and how can they avoid it?
Hey! Hope you are well :D
I believe the most common mistake is not reading carefully before commenting/posting. There were instances where people believe that "the most likes in a DEV challenge = Win", which is not the case. According to the rule, it is only use as a Tie breaker situation and not the overall result.
My tip is reading the full post before making conclusions. Right now, there is assumption that the project me and @codingwithjiro did was use to determine if the article is AI or not, which is not the case. It would been best to clarify the practices on how I determine the articles instead of assumptions. In other words, ask questions!
Thanks! :D
WHO WOULD HAVE THOUGHT?! excellent disclosure
I didn't even know I was a trusted member too, Idk how it works, but cool.
lol
how to get famous '='
I think Don't think about becoming famous. Focus on creating good content and improving your work every day. When you consistently work on quality, recognition and success will come naturally.
Hey! Similar to @technogamerz comment, I would focus on creating good content. It would come naturally and also try to meet new people and make genuine connections!
The DEV iceberg:
Level 1: Learn Markdown
Level 2: Write blogs
Level 3: Discover Sloan isn't a real person
Level 4: Realize hundreds of moderators can control Sloan 👀😂
Hey Francis, appreciate you opening this up and taking some of the load off Ben and
Jess.
I have a question that has been bugging me. I am sitting at around 2,380 followers,
but my total post views across everything I have written is only about 2,332. More
followers than total views does not really add up, and when I actually scroll
through the follower list, a lot of them look like bot or spam accounts rather than
real readers.
I saw Jess mention that automated spam detection is getting updated soon, so this
may already be on the radar. But is there anything on the user side I can do to flag
or clear out fake followers? Honestly I would rather have a smaller number that
reflects the real people reading the work than a bigger one padded with bots. I want
to follow and engage with the actual humans here, not inflate a count.
Thanks for the work you put into this.
I cant recall the name, but didnt some guy build an anti-bot/spam system for youtube comments a few years ago (think like 3-4), I remember LTT covered it and it worked surprisingly well, might be worth looking in to, as it was a free tool?
Oooou now that’s is insightful and helpful if there is any link or way to find that product id love to give it a try. Thank you!
Here. Like I said, it's quite dated (4 years ago) and meant for YT comments, but it is open source and idk if anyone has been updating it since or maybe forked it for bot followers?
I am not sure if it works well in the context of followers because they are concern about bots following their account.
I'm a bit busy at the moment, but I'll fork it in a few hours and see if I cant mod it to be spam and follower filter. It was designed for YT comments, which is inherently a closed-wall system, which means cant really implement it server-side. But Dev.to isnt Google, so who knows, might work? I have quite a bit of experience dealing with WAF and specifically bypassing bot-detection, so I'll apply some of that to find a decent way to flag spammers/bots and add a 'kill' mode to automatically block them, but off by default so you can atleast test it and review the flagged parties first, so you can know it's safe.
I appreciate you taking the time while busy I was checking out that YouTube link an it gave me some refreshing ideas that could help thankfully I haven’t had much bot comments it’s mostly bot followers . But I really appreciate it thank you!
Anytime, that's what community is meant to be. From experience, the easiest way to flag a bot, is to target the tool (sorry for building one that doesnt do that), 99% of bots run Selenium, Playwright, Puppeteer and almost all of them are on Linux (Alpine containers, cuz lightweight), while Data Center IPs are a reliable filter, these days residential proxies are cheap enough that it wont filter even half of them. But when you consider most bots use AI generated responses follow alot, but never post and are generally new... I think I can write up a pretty good detection system for followers and comments, atleast to just mark them as suspects.
I just got started with digging through TheoJoe's code on the YT despammer, pretty heavily tied to YT, so will need a full rewrite. Fun task though, curious what the foundry pumps out.
A few questions though:
Any and all you can answer would be extremely helpful in building a tool you can just plug in and use, instead of needing to rip apart the entire system just to test it.
Thanks in advance @francistrdev
Hey Thanks for the question. I will get back to this around today. Just to leave a note that I see this.
No rush, whenever you get time.
Thanks for the quick response!
Hey @francistrdev, @ben, @jess I finished pass 1 through the foundry, you lot can give it a try and see how it works?
Please let me know how it goes or if anything is unclear or needs adjustment to let it fit your system better?
GitHub - DevGuard
@zep1997 Incase you want to have a look too
I am gonna delay the response till next week. A lot is going on that required my attention at the moment.
No stress, it's there when you get the time, let me know if anything needs tweaking to make integration easier
Hey! Thanks for raising this concern. As of today, there isn't anything you can do about it. It was an issue for a while and I also notice that on my end as well. The best you can do is report it, but do keep in mind that you may never know if they are bots or not.
Thanks again :D
Found it, Anti-Spam have you looked into this yet? It's quite dated, but maybe a good start?
Hey it’s definitely something I appreciate the efforts even if it doesn’t work thank you so much brothaa
lol not very reassuring but i appreciate the response, thank :D
Some comments may only be visible to logged-in visitors. Sign in to view all comments.