starball
I'm a hungry analyst, and an industrious janitor.
As an active site curator (since 2022) and OG member of the AI-detection working group, modship would enable me to do data analysis to support the AI content ban more deeply and accurately, and to more swiftly perform the dozens of curation actions I do daily.
You can find public examples of my analysis work on MSE and on MSO. Behind the scenes over the past year, I've done some brutal data analysis work enabling the recent approval of a highly desired AI-detection heuristic.
Some stats:
- 1,000+ helpful AI-generated-content (AIGC) flags, with 2,000+ still pending handling.
- 44,000+ helpful flags, ~99.6% helpful flag rate, 20k+ up/downvotes, 9k+ close votes, 2k+ delete votes.
Here's my Meta Stack Exchange profile.
Obligatory: I'm human. I've made and will make mistakes. I try to learn each time. Feedback welcome.
- How would you deal with a user who produced a steady stream of valuable answers, but tends to generate a large number of arguments/flags from comments?
If the comments are actually constructive and just not being received well, then I'd use my judgement. That could involve comment cleanup, temporarily locking comments, or closure, etc.
If the flags are warranted and they haven't already been warned, I'd give a warning, being clear and specific about what is and isn't okay (what behaviour should change), point to the Code of Conduct, and exemplify an appropriate comment. If later needed, I'd escalate action. I'd consult more experienced mods if concerned that I would be out of step.
For example: "It's fine to be frustrated and to express frustration, but calling someone stupid or an idiot on this platform is not okay (read the code of conduct). There are also kinder and more constructive ways to give feedback. You could instead say "That's a really bad idea. {reason(s)}. You should read {educative resource}". Alternatively, you can disengage."
- How would you handle a situation where another mod closed/deleted/etc. a question that you feel shouldn’t have been?
If it seems like it was probably an unintentional mistake, I'd open a chatroom and ask them if it was intended or not.
Otherwise, I'd (probably) check mod-guidance texts to make sure I'm not missing something more obvious.
If that doesn't help (or if that mod likes teaching), I'd discuss it with them in a chat room. I'd presume that they have a good reason (and take it as a potential learning opportunity), but also explain my own reasoning. Ex. "I noticed that you {action}ed [this question]({link}). I'm confused because {reasons}. I think I would have done {action} instead. Could you help me understand why you did this?".
If I couldn't come to understand and support the action after that, I'd ask other mods for their thoughts, and submit to their consensus.
I'd potentially open discussion on meta if other mods agree that that would be useful.
- [...] you will not be allowed to handle [most] pending AI flags by deleting the post, because the post doesn’t meet company-approved heuristics for deletion. [...] Why do you still want to become a moderator specifically under these conditions? How will you deal with the frustration [...]?
I'll continue to devote my energy to doing the R&D necessary to improve the pool of approved detection heuristics. I've been doing this in my spare time for over a year now. (I started even before the AI ban policy was put in place!)
I hope that by signing the mod agreement (esp. the privacy policy bits), the company can trust me with access to simple tools that would help me do that R&D much more efficiently. Frustratingly, I already have access to the info I need- just in an extremely difficult form to work with, and Slate said it would be difficult to give me a more workable form for legal/privacy reasons. I'm still discussing this with them).
Also- FYI- if you're concerned about me declining AIGC flags solely because they don't satisfy currently approved heuristics, you should know that over half of the 4,000 pending flags on AI-generated answer posts were raised by me- my flag reasons often referring to heuristics that are still pending approval. I.e. I dual-wield: I both flag AIGC using heuristics that I want to see approved, and contribute to the work that makes those flags processable by mods. Yes, I have the audacity to do that.
- Given recent developments on the site [...] in the past year, [...] What motivates you to serve this community, when many users openly object to moderation and the company's decisions sometimes seem at odds with moderator and community consensus?
Stack Overflow's model, community, and content matter. I believe the same core reasons they've been valuable to the world since day one still hold and will continue to long-term. It is not a panacea, but it holds a very important place in the world. I'm part of it, and I want to develop and protect it.
Yes, many people find moderation unpalatable. But I hold an appreciation for the SO model (including voting, editing, duplicate-closure, etc.) as a pillar of SO's value. It took time, effort, and humility for that appreciation to develop in me. To all the folks who've given up on SO (many finding something more palatable in ChatGPT, which feeds on SO in its training data and web searches), I hope they can somehow come to a similar appreciation for SO. I hope that whenever the machine fails them, they consider coming here. And SO is made by people; I hope my story in finding a place here as a contributor can light some inspiration to another generation.
As for company woes, I'm also very frustrated with a lot of the decisions the company has made- those made before my arrival (essentially pre-2022), and those made while I've had front-row seats. But I'm not here because I love the company. I'm here because the model is here. The community is here. The content is here. Some community members have gone to greener pastures (everyone is entitled to their own decision-making of when/why to stay/leave, and I respect those who have left), but they also recognize that SO still (despite it all) holds the name recognition that makes many people come to it- recognition which many competing platforms don't have. The model, community, and content are movable and not wholly tied to the company. If a significant migration happens (probably very gradually), at some point, I will go with it. I am here to serve the knowledgebase and its community- not the company. Still, I hold vain hope for the company to try to have a better relationship with the community, and hold trust in Slate.
- In what way [will] being a moderator make you more effective than what you could do with the access moderator tools, trusted user, and other privileges?
Note: I've already earned all those privileges (and a dup-hammer in my main tag) and have been trying to make good use of them for some time now.
Diamond-mod tools would let me more swiftly handle the bulk of what I'm already handling, which is in large part non-answers and comments that are no-longer needed, and would also let me help out with the mod flag queue.
I'm interested in the ability to search deleted posts, which would help in gathering accurate statistics for heuristics development (in detecting AI-generated content, to enforce the AI-ban policy), which I do in my spare time. Frustratingly, I know that even mod tools are (currently) insufficient for efficient AI-detection heuristic R&D (see Q#3).
- This question was posed by me before Q#4 was suggested. I've answered this in Q#4.
- Do you have any particular philosophies on moderation or curation that might set you apart from other candidates?
I have complaints :P
Mods have a lot on their plate that I think can and should be delegated to privileged community members. Ex. Comment mods is a good idea. I've also mulled on ways to share investigation load about AIGC flags to trusted community members, but nothing has crystalized yet.
Moderation should come to need to happen much less, because there's so much room for site onboarding and JIT guidance to be improved. For example, it's absolutely ridiculous the number of questions I see posted using answer posts every day. It's very costly: For non-mods to handle a single one, at least three people need to action it, and those people are statistically very rare (who have the review/del-vote privilege, and want to spend their time being an unpaid internet janitor). The community has already developed a set of pattern-matching rules to detect when an answer post is likely not actually an answer (Natty, Dharman's bot, and more). The company should implement a feature for the answer box to pop up a hint like "note that answer posts are for answering the question" when someone is drafting an answer post that matches one of those patterns.
I have more to say, but character limits are a thing.
Do these takes/ideas set me apart from other candidates? I have no idea. I'm sure they aren't new. But all of them require the company to actually care, and do something about it (sigh).
- [...] Do you seek alignment in moderation with other moderators? If so, what will you propose to achieve alignment in moderation with other moderators?
Yes. I'll be joining a team and I'll try to hold myself to be a good team member. That includes seeking alignment and submitting to consensus.
I'm not afraid to ask for guidance when I unsure of what to do, and when someone informs me that something I'm doing is not appropriate, I seek to learn why and change course to a better way.
As for systematizing alignment, I don't have any proposals right now. As a non-mod, I don't have a good understanding of how the problem looks on their side of things. Trying to solve a problem one doesn't understand well is a recipe for a poor solution, and I don't want to enter a team as a new member and start telling people how to solve a problem that I have no experience with.
- A frequent answerer flags for moderator attention because a curating user persistently closes questions (rightfully as duplicate/off-topic) then posts comments under the answerer's posts to remind them to close close-able questions instead of posting an answer. [...] How will you handle such a situation to best serve the platform and its users?
If these are single-vote ("hammered") dup-closures and I don't have subject-matter expertise in the topic, I'd first want to make sure these are really duplicates by consulting with the community / the subject-matter experts in that topic (likely via meta+chat). I'd want at least two people with subject-matter expertise to weigh in.
If there's community/expert-agreement that these are really duplicates, then I'd communicate to the unhappy duplicate-answer poster that answering duplicates is discouraged by official guidance and frowned upon by the community, and that unless they can give a good explanation as to why the questions are not duplicates, instead of answering them, they should be closing them, and seeking to improve the canonical Q&A post(s) (perhaps by writing an answer post there instead, and/or editing to improve the Q&A's searchability).
- What personality traits and/or technical skills do you have that you think will be useful in your modship? How did you come to acquire them (if relevant)?
I have a high tolerance for mundane, rote work, and I'm able to find enjoyment and satisfaction in it. It's probably part nature (some brain wiring for particular intense interests), and part nurture (from a young age, I've been doing a lot of weekly venue setup and teardown at my church). You can probably expect me to help out especially with the more "boring" mod tasks things like comment flags.
I also really like patterns and looking for patterns (again, probably brain wiring). I enjoy querying SEDE and doing AI-detection heuristics R&D. As a programmer, I'm also comfortable writing tools to help me perform repetitive tasks. These have enabled and driven me to do R&D work for detecting AI-generated content on SO. Where relevant, you can expect me to try to weigh in in discussions in mod spaces (including discussions with the company) in a data-informed way.