User talk:M kuhner

Mordey Article cleanup

Re William Morris Mordey

Hello @M kuhner I couldn't leave it so I've gone though systematically and removed 'bulk' info I don't think intrinsically relevant. it was hard to do this probably because I've got so interested in the subject matter I think its all relevant.

Ive removed the power station 'lists' entirely and opted for a short paragraph. I've removed some sections entirely and shrunk others down or removed entrances that do nothing but pad. As you recommend this is my first pass. next will be to reword If needed how the article and words flow.

Thanks again Mordzy (talk) 19:00, 29 June 2026 (UTC)

I am having a work crisis so may not get to this for a few days--just wanted to let you know. M kuhner (talk) 14:38, 3 July 2026 (UTC)

Roger that. I have some time this afternoon to re-read and change the wording to flow better. I have left it for a week so I can return to it with fresh eyes. Mordzy (talk) 14:03, 5 July 2026 (UTC)

Hi, Have you had chance to review? on a side note I have had alot of my research published by the Institute of Engineering and Technology so could be used as additional citations. Mordzy (talk) 12:01, 3 August 2026 (UTC)

Whoops, sorry! I will look at it today. M kuhner (talk) 14:39, 3 August 2026 (UTC)

Hello @M kuhner have you managed to review. Mordzy (talk) 15:29, 26 August 2026 (UTC)

I left a bunch of comments on the Talk page on August 4, have you seen those? M kuhner (talk) 16:00, 26 August 2026 (UTC)

@M kuhner I have now, thanks. Will review and update you. Mordzy (talk) 21:16, 3 September 2026 (UTC)

Teahouse closures

Hi M kuhner. I discussed it a bit in an edit summary, but I figured a fuller explanation would be better. I would generally recommend against closing discussions on the Teahouse, especially when they are so close to archiving on their own (as functionally all this does is prolong their lifespan on the main Teahouse page). The Teahouse isn't like ANI or other conduct noticeboards, where closures are needed to state the result of a large thread. Teahouse threads are generally smaller and more manageable. IMO, there is broadly speaking one reason to close (or hat) a thread at the Teahouse, and that is to pressure others to stop replying at it. This can be the case if a conversation isn't productive (though I would urge you to be cautious about applying this rationale) or threads created by socks. Closing threads as resolved doesn't make much sense IMO since it denies the ability for the OP or other new editors to ask questions using that thread. Best, 45dogs (they/them) (talk page) (contributions) 02:38, 7 July 2026 (UTC)

Okay, I'm sorry for any disruption, and will refrain from doing this in the future. My reasoning was that the page was getting rather huge and directing attention to open questions might be helpful, but your argument makes sense. Thanks for letting me know. M kuhner (talk) 02:40, 7 July 2026 (UTC)

Letting threads naturally archive is generally quicker to reduce size. Since closing produces a signature, it resets the archiving clock on that thread to 72h. If a thread is unproductive though and keeps getting replied to (preventing it from naturally ending via automatic archiving), then a closure can be helpful. 45dogs (they/them) (talk page) (contributions) 02:43, 7 July 2026 (UTC)

Thanks, that's clear.

While I have your attention, are you able to use TAIV to see if ~2026-36771-82 is likely to be LTA/CB885? I am finding the combination of claiming to be brand new, claiming to be young, and filing at AN to be surprising, and "I'm new" is something this LTA apparently says a lot. M kuhner (talk) 02:53, 7 July 2026 (UTC)

Perhaps Wikipedia:Sockpuppet investigations/ChronicleBooks885#07 July 2026 will answer that. 45dogs (they/them) (talk page) (contributions) 03:11, 7 July 2026 (UTC)

I can't say much of anything in regards to specifics that would be helpful due to policy, but ~2026-36771-82's IP holds several similarities to previous TAs used by CB885. 45dogs (they/them) (talk page) (contributions) 03:23, 7 July 2026 (UTC)

Thank you.

I'm sorry for not hunting that down myself, but I find SPI peculiarly hard to search. Anyway, glad to know it's being appropriately dealt with. M kuhner (talk) 03:27, 7 July 2026 (UTC)

Sorry I vanished. Our cat got diagnosed with diabetes and that is consuming my free time. I'll let you you know when we finally settle into a routine. Austinious (talk) 20:26, 18 July 2026 (UTC)

Oh dear. Medicating cats is so hard! Anyway, thanks for letting me know. I have a draft of the E. atrica article started but haven't had much time either. Will look forward to hearing from you when things settle down. M kuhner (talk) 20:30, 18 July 2026 (UTC)

Draftify/userfy

Hi Mary, great work at AINB (you're a force of nature ngl), just FYI it's better imo to recommend that editors userfy articles instead of draftify, I keep coming across articles that were draftified and then much later moved back to mainspace by someone unaware of the problems/context (eg. ) Kowal2701 (talk, contribs) 19:04, 16 July 2026 (UTC)

Thanks, that's good to know.

For external reasons (I'm an American, OMG) I have a strong desire to get my teeth into someone or something, which I'm taking out here: hopefully not biting anyone who doesn't deserve it. M kuhner (talk) 19:59, 16 July 2026 (UTC)

you're doing great lol, don't worry about it Kowal2701 (talk, contribs) 20:48, 16 July 2026 (UTC)

I have to say, I no more understand how you can do that many LLMPRODs that quickly than I understand how top Starcraft 2 players can hit 500-600 actions per minute. You've explained as best you can, I'm not asking for more explanation: just marveling. It takes me more time to batch-update the trackers for your LLMPRODs than it seems to take you to do them! M kuhner (talk) 21:08, 16 July 2026 (UTC)

lol, have to be a bot to defeat the bots! Kowal2701 (talk, contribs) 22:03, 16 July 2026 (UTC)

In Starcraft 2 there's a bot scene where people make bots and match them against each other. One of my favorite bugs: the Zerg faction can burrow its units into the ground, where they regenerate and are difficult to attack. Bots favor this as they can do it unit by unit and extremely rapidly. But sometimes they seem to forget that the goal is to unburrow, attack, and reburrow, and just pop up and down at a rate too fast for any attacks to happen. If both players are Zergs, you get two armies sitting next to each other, fizzing like a shaken soda, but not actually fighting.... M kuhner (talk) 02:29, 17 July 2026 (UTC)

haha, I'd much rather we all did that tbh lol, the encyclopedia might actually be in better shape! Kowal2701 (talk, contribs) 09:56, 17 July 2026 (UTC)

Btw, idk if you're already aware, but mw:WWT is great for WP:LLMPRVing those old edits, User:Nardog/DiffUndo also looks helpful for restoring the pre-LLM content without having to completely restore a previous version (though I haven't gotten it to work yet). I'm gonna do a resources page for AI cleanup soon, will run it by you when possible Kowal2701 (talk, contribs) 20:31, 26 July 2026 (UTC)

I don't know what I'd do without WWT. Give up on the harder ones, probably. It's an incredible tool; the one big deficiency is that it won't do text in tables (which showed up in a Global Donald air accident article where 90% of the text was in a huge table of crashes).

Looking forward to the resources page.

I guess I'll just throw in here, since it has no obvious home but I want to get it off my chest: Global Donald got archived, but I was within 20 articles of a complete pass (not a complete cleanup, but every article addressed somehow). Will try to finish. There are about 2-3 questionable dePRODs in there and quite a few that we could do nothing better than tag. I've gotten a few thanks from local editor communities who knew very well that something was wrong, but not what. It does feel like an accomplishment of sorts. M kuhner (talk) 20:41, 26 July 2026 (UTC)

I've made phab:T433208 btw. Sorry about the Global Donald thing, the page was too big and was having difficulty loading so I archived some of the large sections (like 100+ comments) that I thought were just going to end up in the backlog anyway (ofc I didn't account for you!). Feel free to unarchive it if there's more input you'd like from others (generally feel free to revert any archiving), otherwise you can just edit the archive to say cleanup is complete (I've been doing this for LLMPRODs, even though you're not meant to edit archives). Yeah, it's always nice to see efforts appreciated (often not the case for those doing the patrolling lol). I've had the same feeling after doing a large batch of LLMPRODs, though they're much easier than what you're doing. Godspeed, Kowal2701 (talk, contribs) 20:59, 26 July 2026 (UTC)

New guide/resources page is at [1], feel free to boldly edit it Kowal2701 (talk, contribs) 00:36, 4 August 2026 (UTC)

I've started making tracking pages for cases in the archives (sorry), for ones where I struggled with the start date I'm probably going to make a list in the backlog bonanza pages Kowal2701 (talk, contribs) 09:59, 5 August 2026 (UTC)

(just FYI, the UpdatingUK case you closed has one article listed as ongoing which needs stubbifying) Kowal2701 (talk, contribs) 12:14, 6 August 2026 (UTC)

Ugh! Okay, will do. I mark tagged but not fixed as "ongoing" but often just in the fond hope that someone else will fix it.... M kuhner (talk) 16:53, 6 August 2026 (UTC)

lol don't worry, I can do this one, but yeah, for cases where there's a few like that it might be worth trying WikiProjects Kowal2701 (talk, contribs) 18:32, 6 August 2026 (UTC)

WP:BLAR'd it because it didn't come close to meeting GNG :) Kowal2701 (talk, contribs) 18:38, 6 August 2026 (UTC)

Thanks! Biographies are not really my thing (I've been mostly dodging them in the backlog drive unless very obvious). M kuhner (talk) 19:04, 6 August 2026 (UTC)

(I'm gonna try to avoid archiving cases you're working on or ones close to completion, if I do feel free to revert) Kowal2701 (talk, contribs) 23:28, 29 August 2026 (UTC)

btw MSK has created User:MSK/rdp.js which automatically cleans up deprecated infobox parameters Kowal2701 (talk, contribs) 19:57, 31 August 2026 (UTC)

Woohoo! I will give it a try. I have been painstakingly hand-fixing them for fear of another visit from the Infobox Gnome. (Though a funny note on that: I think I got non-probationary AfC reviewer status, which I wasn't expecting given how green I am, because the person who granted it looked at my interactions with Zachmann and said "shows accountability.") M kuhner (talk) 20:28, 31 August 2026 (UTC)

Lol, careful now, that's how people get roped into becoming admins! I prefer to maintain a strategic whiff of incompetence personally Kowal2701 (talk, contribs) 20:31, 31 August 2026 (UTC)

I think I could defend myself from such threats by noting that my articles created stands at 2, good or featured at 0. I have already destroyed more articles than I will create in the rest of my life. (I can only imagine how many you have!) This seems to be a problem at RfA. (And now you know something unsavory about me, which is that I do think about that.)

M kuhner (talk) 20:46, 31 August 2026 (UTC)

Unfortunately, you only need one or two GAs to show you know what content policies are, and you do tons of article reviews already :( Kowal2701 (talk, contribs) 20:48, 31 August 2026 (UTC)

I write SF novels as a hobby, and my protagonist was just told by a higher-up in her spy organization that if she didn't stop being so damned precocious, she'd either have to be promoted or disposed of.

Always amused how things echo each other. M kuhner (talk) 20:52, 31 August 2026 (UTC)

might find WP:ADEADVICE interesting (long-term ofc, 2 years of editing is usually the bar AFAIK) Kowal2701 (talk, contribs) 20:51, 31 August 2026 (UTC)

would it be possible to add a bolded comment/note to trackers where the start date is yet to be determined? (btw there's discussion re a new system at , since you're like the most knowledgeable/affected by this lol would be good to hear any thoughts you may have, cheers) Kowal2701 (talk, contribs) 17:49, 20 September 2026 (UTC)

also, your work ethic is incredible, closing out so many monstrous cases recently! Kowal2701 (talk, contribs) 17:50, 20 September 2026 (UTC)

Frustrated with my job right now--we are about to apply to the National Science Foundation for money, but it's not clear that they have enough political independence to be relied on. I am deeply appalled by the fact that I have to weigh how my state voted in the last election, and will vote in the next one, in a decision whether to apply for funding. So I take out my rage on hapless LLM users.

Also the improved tools help IMMENSELY, you don't know how much you need something until it magically appears. When your 100+ NShehan LLMPRODs complete they can be updated in two clicks and then I can find which ones you skipped without looking at all of them!

Do you think you could look at 1-2 Orokbest1 edits at the bottom of that tracker? I am feeling like they're not LLM down there, but I fear I'm prejudiced by wanting to be done....

M kuhner (talk) 18:24, 20 September 2026 (UTC)

Yeah, that really does sound infuriating. I guess things have to get badly broken for them to get fixed, but then you look at the people who seem like they'll be the ones tasked with fixing them and... If it's any consolation, it looks like the rest of the West is heading the same way lol. If you need to take a break for whatever reason, do say. I burnt out a couple days ago and have taking it easy since, should be back to full 'fitness' next week.

I know, I didn't even ask for it lol. LLMPRODs used to take me like 25-30 seconds each, now only 12. Currently in talks with some devs, one of the things I'm gonna ask for is a tool that makes reverting easier. Will copy you in if that goes anywhere.

Just had a look at David Chartrand the added ref in Special:Diff/1323365057 didn't even mention his brother lol (it had already been reverted). The ref in Special:Diff/1323073225 for Alice West Fleet is about a project for a school, doesn't appear to even mention the person. Kowal2701 (talk, contribs) 19:46, 20 September 2026 (UTC)

Kowal2701 (talk, contribs) 22:30, 20 September 2026 (UTC)

is it alright if I revert the close of WP:AINB#User:NShehan? There's some entries that are normal edits and not llmprods, I'll try to do some of them now Kowal2701 (talk, contribs) 16:37, 22 September 2026 (UTC)

Whoops, sorry! I thought you had tagged them. Yes, please. I think it'd be more efficient, though, to wait on the tagging until the LLMPRODs go through, mass-resolve those, and look at what's left. M kuhner (talk) 16:40, 22 September 2026 (UTC)

Normally agreed, but I can see the non-LLMPRODs cos they're not purple-linked lol (does that make sense) Kowal2701 (talk, contribs) 16:45, 22 September 2026 (UTC)

Ah, right! Usually doesn't work because articles not LLMPRODed had to be clicked into to verify they couldn't be, but in this case, should be fine.

This is a sad one, good faith editor but hasn't come back despite saying they would help. We've got to inform people sooner that they can't do this. (They really can't; stubifying the ones I did convinced me of it. Wretched bots can't even write a basic article from database without completely botching it.) M kuhner (talk) 16:52, 22 September 2026 (UTC)

Good news:

• There's an LLM version of Paste Check coming out 24 September (Thursday): phab:T420258, which doesn't block an edit but shows a notice informing about WP:NOLLM (we can choose what the message is), this is flagged for other editors
• Devs are currently looking at a similar thing for chatbot-derived links (phab:T434810) and content watermarking (phab:T423545, which I hope in the mid- to long-term we can use to block edits from being published)
• I'm going to ask whether it's feasible to make a tool specialised for reverting a user's edits to an article (i.e. better than rollback, potentially even reinstate content that was 'overwritten' by the LLM user)

Bad-good news:

• This'll likely lead to a much higher rate of AINB cases, especially in the short term, and we haven't even started moving to the new AINB system yet. I also don't know how to do any of it.

Kowal2701 (talk, contribs) 18:07, 22 September 2026 (UTC)

(Noticeboard is gonna be a bit messy for a couple days while we're moving cases, you can see the table of subpages at Wikipedia:AI noticeboard/AINB/Cases for now, closed ones are at Wikipedia:AI noticeboard/AINB/Closed, lmk if anything's confusing, and do pester @NicheSports if you've ideas for improvements) (btw templates like don't work in edit summaries, has to be a normal wikilink, so like to ping you'd use [[User:whatever]] instead of {{u|whatever}} or {{ping|whatever}}) Kowal2701 (talk, contribs) 16:59, 13 October 2026 (UTC)

The one thing I'm really noticing--let me know who would be the right person to say this to--is that having all notifications say "Kowal2701 posted to Discussion" is not very useful compared to the previous "Kowal2701 posted to User Blah-blah." Hard to find a particular notification when they all have the same name. M kuhner (talk) 20:21, 13 October 2026 (UTC)

I said the same thing! What I had in mind was what's at Wikipedia:AI noticeboard/Sanchoof, thoughts? Kowal2701 (talk, contribs) 21:00, 13 October 2026 (UTC)

That is precisely what I was going to propose. M kuhner (talk) 23:12, 13 October 2026 (UTC)

July 2026

Thank you for your contributions to Wikipedia. Regarding your recent edits to Yosemite Valley when you modified the page, you introduced unknown parameters. Just because you specify |some_param=some_variable does not always mean that variable will display. The |some_param= must be defined in the template. You can look at the documentation for the template you are using but it is also helpful to use the preview button before you save your edit; this helps you find any errors you have made and ensure that the values you have added are displaying correctly. Below the edit box is a Show preview button. Pressing this will show you what the page will look like without actually saving it. It is strongly recommended that you use this before saving. Note I have likely fixed the error by now so check the history of the page to see how it was fixed. If you have any questions, contact the help desk for assistance. Thank you. You cannot just revert multiple edits because ONE of them is LLM. Zackmann (^{Talk to me}/_{What I been doing}) 23:52, 17 July 2026 (UTC)

I apologize for not spotting the issue using Preview, and will be more conscientious about that in future.

However, I am working through edits by one LLM using editor to 600 pages. When there are only a small number of subsequent changes and they appear to be cleanup rather than substantive additions to the page, it seems best to revert. In many cases the editor added tens of thousands of bytes and their work is full of source/text mismatches and false statements. The consensus at AINB seems to be that these should be reverted even at some risk of damage, because letting them sit is worse. If the added text forms separate sections I can sometimes revert by hand, but for this particular editor it often doesn't.

If you believe I'm wrong about this, I suggest an RFC, because I'm not the only editor operating on this basis and it would be best to clear it up for everyone. M kuhner (talk) 00:03, 18 July 2026 (UTC)

You are responsible for the edits you make. When you revert without checking your edits you remove valid content from the page. Zackmann (^{Talk to me}/_{What I been doing}) 00:10, 18 July 2026 (UTC)

I will pledge to use Preview every time and look for errors. You're perfectly right about that.

I'd be interested to hear what you think should be done about 5K bytes of LLM text with several small copyedits sitting on top, though. M kuhner (talk) 00:13, 18 July 2026 (UTC)

I want to be clear, I am NO fan of LLM and I also fully support removing LLM material. My point is that when you remove it, you need to make sure you are removing just the LLM and not other material. When you reverted, your edit also remove the image, the caption and the map from the Infobox. Obviously I know that wasn't your intent but this is why it is so important to WP:PREVIEW. Zackmann (^{Talk to me}/_{What I been doing}) 02:00, 18 July 2026 (UTC)

Will do, and I'm sorry for the trouble caused. That was a useful wakeup call and I appreciate it. M kuhner (talk) 02:01, 18 July 2026 (UTC)

Appreciate you being so receptive to the feedback. Keep doing your thing and fighting the good fight! Zackmann (^{Talk to me}/_{What I been doing}) 02:02, 18 July 2026 (UTC)

I found one article where there were two massive LLM edits, followed by an infobox-correcting edit by you. I ended up reverting the LLM and then hand-reinstating your edits. Is there a tool that would make this easier in future? M kuhner (talk) 18:28, 18 July 2026 (UTC)

Unfortunately not really... This is one of the many problems with these LLM vandals. You can certainly look into using JWB/AWB but these really only help when you are making the same change to multiple pages. Appreciate you taking the care to fix the Infoboxes. Zackmann (^{Talk to me}/_{What I been doing}) 18:38, 18 July 2026 (UTC)

Thank you for your contributions to Wikipedia. Regarding your recent edits to Union Glacier Blue-Ice Runway when you modified the page, you introduced unknown parameters. Just because you specify |some_param=some_variable does not always mean that variable will display. The |some_param= must be defined in the template. You can look at the documentation for the template you are using but it is also helpful to use the preview button before you save your edit; this helps you find any errors you have made and ensure that the values you have added are displaying correctly. Below the edit box is a Show preview button. Pressing this will show you what the page will look like without actually saving it. It is strongly recommended that you use this before saving. Note I have likely fixed the error by now so check the history of the page to see how it was fixed. If you have any questions, contact the help desk for assistance. Thank you. Zackmann (^{Talk to me}/_{What I been doing}) 05:14, 25 July 2026 (UTC)

Apologies. The difference between - and _ seems to have eluded me.

Is there something I should be doing to see these errors in particular? I did preview and saw nothing wrong. M kuhner (talk) 05:18, 25 July 2026 (UTC)

Unless you are using the visual editor (which inexplicably doesn't support preview warnings, you should have seen a large number of error messages. If you go to your revision: Special:Permalink/1365911771 and preview that revision, you should see a number of errors... Zackmann (^{Talk to me}/_{What I been doing}) 05:21, 25 July 2026 (UTC)

Don't know how I missed that, then. My apologies, will strive to do better next time. M kuhner (talk) 05:24, 25 July 2026 (UTC)

All good! These things happen. Appreciate you being receptive to the feedback. Keep up the good work fighting the LLM vandals. Zackmann (^{Talk to me}/_{What I been doing}) 15:43, 25 July 2026 (UTC)

I figured out what happened: I used "restore this version" and that does not allow preview. I looked at the article afterwards and it looked fine, but I didn't preview, which was a mistake. I'll go back and preview in future.

"Restore this version" seems like a bit of a hack, honestly. Itty bitty box for edit summary, no preview. Also kind of surprising that, while rollback is a user right, this is not: seems equally destructive. M kuhner (talk) 15:47, 25 July 2026 (UTC)

A bowl of strawberries for you!

Your contributions to Wikipedia are legendary; many appreciate all the work you've put in. Thank you for all that you've done. Benahue (talk) 18:07, 18 July 2026 (UTC)

I would not have expected that response to...arguing to delete your article. But thank you! M kuhner (talk) 18:46, 18 July 2026 (UTC)

Hah, well, debates are always healthy. I replied to your comments on the GMass page and had some questions for you. I really think that page belongs on Wikipedia and am curious to hear more thoughts from you. Benahue (talk) 19:28, 18 July 2026 (UTC)

A cupcake for you!

Another day, another cupcake. Everyone loves cupcakes. Benahue (talk) 23:08, 20 July 2026 (UTC)

About the GMass article

Hey, M kuhner, I added three more citations on GMass' page in an effort to improve the article. These sources are all interviews that the founder of the company, Ajay Goel, did with places like GrowthHacker.tv. In the interviews, Goel discusses how his company reached $1.5 million in annual revenue and so on. I think that I've significantly improved the page and demonstrated that GMass is notable. Would you consider changing your vote to "Keep"? Benahue (talk) 23:21, 20 July 2026 (UTC)

I'm sorry, no. Reasons are at AfD.

Business articles are a hard topic to get through AfD because a lot of editors feel that businesses are constantly pressuring Wikipedia to host free publicity for them. Business articles have a ferocious rate of undisclosed conflict of interest, paid editing, and outright spam. So you're running into a particularly prickly reception. It might be easier to make a first article on something else, and come back to this one later with more experience.

Thank you for the cupcake! Good thing they are virtual as I'm not allowed to eat them in real life. M kuhner (talk) 23:36, 20 July 2026 (UTC)

Well, I planned to make an article about a Tom Petty song, but since getting an article like that to stick is relatively easy (I assume), I doubt I'll learn much from it. Benahue (talk) 21:15, 21 July 2026 (UTC)

a cookie

~2026-38678-64 has given you a cookie! Cookies promote WikiLove and hopefully this one has made your day better. Spread the WikiLove by giving someone else a cookie, whether it be someone you have had disagreements with in the past or a good friend.

for dealing with numerous fires at ANI ~2026-38678-64 (talk) 23:30, 20 July 2026 (UTC)

Thank you for the cookie, though right now I want something that BLEEDS when I bite it.... (I have to rewrite a long science paper that my boss has edited into an incoherent mess.) M kuhner (talk) 23:37, 20 July 2026 (UTC)

Just have AI rewrite it for you (just kidding). Good luck with that. Benahue (talk) 22:52, 21 July 2026 (UTC)

OMG I just spent lunch trying to talk him out of writing an important report to the Dean that way! (But he did pay for lunch. That's something.) M kuhner (talk) 22:54, 21 July 2026 (UTC)

Icinga article

Hi M kuhner, thank you for your AI cleanup efforts! I've left a note on the Talk:Icinga page about some technical accuracy issues in the current version of the article. As you made the most recent edit, I wanted you to know directly. I've disclosed my COI and also included official sources for each point. I appreciate, if you have a look at it when you have time. Let me know, if you have any comments. Thank you! So1289 (talk) 10:41, 21 July 2026 (UTC)

I will look into it. The problem with recent LLM is that the sourcing issues tend to be subtle, such as attributing statements from Source A to Source B: these are hard to spot. But we can certainly update the information. Regrettably, the volume of LLM cases is so high that each article gets only the best quick-fix we can come up with, often not a very good one. (There were over 600 articles touched by this particular editor.)

It may take me a couple of days, apologies, due to work pressure. M kuhner (talk) 14:25, 21 July 2026 (UTC)

Than you, that's much appreciated. If there are any sources, that are unclear or if you need references to additional documentation for a specific piece of content, just let me know. Thanks again for taking the time. ~2026-40989-19 (talk) 06:45, 22 July 2026 (UTC)

Sorry, that was me! Wikipedia logged me out. So1289 (talk) 06:47, 22 July 2026 (UTC)

Hi! I have posted replacement sources for the five points you mentioned on Talk:Icinga, including the issue of archived versions for the dead links. I just wanted to flag that the sources are ready whenever you find the time. Thanks again for taking this on and let me know if you need more sources. So1289 (talk) 09:54, 7 August 2026 (UTC)

Hi M kuhner! I totally understand you're dealing with a large backlog of AI cleanup, so no pressure at all. Since the article still reflects the outdated version with deprecated features, I'm going to flag it at WikiProject Computing to get some additional eyes on the suggestions I posted on Talk:Icinga. Thanks again for your help so far and let me know, if I can support in any other way. So1289 (talk) 13:00, 31 August 2026 (UTC)

Oh crap. I started working on this, found it very hard, and bogged down. I'm sorry. I think your idea is a good one. I'll also see if I can pry loose some time. M kuhner (talk) 13:10, 31 August 2026 (UTC)

I completely understand that the article can't be corrected right away and that it takes effort. I'm sorry to pass the burden on to the community, but I'll try my best to at least provide sources. So1289 (talk) 05:34, 1 September 2026 (UTC)

A barnstar for you!

	The Barnstar of Diligence
	Thank you for rolling up your sleeves and cleaning out the slop! You're making a huge difference. —ClaudineChionh (she/her · talk · email) 23:11, 22 July 2026 (UTC)

Article: 1969 Vnukovo Airport Il-18 crash

I've just stubiffied the article with sources, feel free to revert if you think this is unsuitable. Ohok12 (talk) 16:04, 28 July 2026 (UTC)

Thanks for doing this. It looks fine to me: hopefully it can be built out with human text and sources in future. M kuhner (talk) 16:12, 28 July 2026 (UTC)

Massive Appreciation

Just stopping by to thank you 🙌🏿. Thanks so much for taking the time to look at the articles I am cleaning up for AI issues @M kuhner. Your kindness is inspiring. 💐💐💐 SyzygyRebel (talk) 19:44, 31 July 2026 (UTC)

Beauty Pie AfC review

Hello @M kuhner, Thank you for reviewing Draft:Beauty Pie, and particularly for pointing me to WP:CORPTRIV. I went back through that section of NCORP and reviewed the sources in the draft again. I understand the concern regarding several of them, particularly those centred on a funding round, a retail opening, or other routine business developments. However, I am having more difficulty understanding how the 2016 Allure article and the 2020 New York Times article fit within CORPTRIV. Both contain fairly extensive discussion of Beauty Pie itself and its business model, rather than simply reporting a product launch or other routine event. The NYT article in particular also includes independent commentary on the limitations of the membership model. I may of course be applying NCORP incorrectly here, so I would appreciate some clarification before deciding how to proceed. Do you consider those two sources themselves insufficient to establish significant coverage under NCORP? If so, could you clarify what you see as the issue with them? Or is your concern rather that, even if one or both qualify individually, the draft does not contain multiple sufficiently strong sources overall? Thanks again for taking the time to review the draft. Toometa (talk) 10:45, 7 August 2026 (UTC)

I think I dismissed the Allure source because it is so extremely positive, and so promotional, that I assumed it was paid quasi-advertising. However now that I look at the magazine its editorial policies and reputation seem good enough to make it a reliable source. The NYT source does have a couple of lines of critical analysis. The fact that they are bylined is positive for both. Is there any chance of finding one more like that?

NYT says: Some industry observers have pointed out that for a Beauty Pie membership to make financial sense in terms of overall savings, the volume of purchases has to be well into the hundreds of dollars. Others have noted that at a time when the sector is heaving with innovation and choice, not all beauty aficionados will want to commit their beauty budget in one place, and that the multiple levels of membership could be confusing to some shoppers. Unfortunately it does not name these sources, but they must be out there. If you could find one of those I think that would tip it over for sourcing and provide needed balance. M kuhner (talk) 15:40, 7 August 2026 (UTC)

Thank you, that's very helpful. I went back through the sources with your comments in mind, and I think I may have found what you are looking for. I had initially discounted this 2022 Guardian/Observer article because it is framed primarily as a profile of and interview with Marcia Kilgore. Looking at it again more closely, however, there is a substantial amount of independently written material specifically about Beauty Pie and its business model, separate from Kilgore's quoted comments. In particular, the article discusses the membership model, the company's growth and financial position, and raises the question of how the model might fare during the cost-of-living crisis and in the context of consumers cutting back on subscriptions. Would you consider this the kind of additional source you had in mind? I also wondered whether reference 9 in the current draft, the 2021 Fast Company article, might be worth another look in light of your comments. Although it includes product reviews, it also discusses the membership model in some detail and identifies limitations of the pricing and spending-limit structure rather than simply describing the company's products. Thanks again for taking another look at this. Toometa (talk) 16:55, 7 August 2026 (UTC)

I think if you foreground those sources a bit, and tone down promotional material, it should be acceptable. In the article as submitted there may actually have been enough sourcing, but it wasn't sufficiently reflected in the text. You'll want to be clear which era of the product plans the criticisms relate to--I think that will also strengthen the article by putting the changes described in the last paragraph into context.

Once you are satisfied with the changes, just put it back in the queue. I don't know if I'm allowed to re-review: I'll check and find out.

Best of luck with your article! M kuhner (talk) 18:02, 7 August 2026 (UTC)

It looks like while re-reviews by the same editor aren't forbidden, they're not considered best practice. There's currently a backlog clearing drive going on, so reviews should be relatively fast right now. M kuhner (talk) 18:07, 7 August 2026 (UTC)

Thank you very much for taking the time to look into this further and for the additional guidance. Much appreciated. I'll revise the draft accordingly and put it back in the queue once I'm satisfied with the changes. Thanks again! Toometa (talk) 20:13, 7 August 2026 (UTC)

Lock Upp 2

Hi, thanks for closing that discussion. I appreciate it :) Tenshi Uisu (talk) 22:26, 8 August 2026 (UTC)

Zeedah Meierhofer-Mangeli Redone

Hello, @M kuhner, I have taken your advice and largely rewrote Zeedah Meierhofer-Mangeli. I wish I can say it will be error free, but I don't have that confidence🤞🏾. Please, when you have a minute, kindly check it out. Thanks for always. 💜 SyzygyRebel (talk) 05:55, 12 August 2026 (UTC)

Your submission at Articles for creation: Eratigena atrica has been accepted

Eratigena atrica, which you submitted to Articles for creation, has been accepted!

Congratulations, and thank you for helping expand the scope of Wikipedia! We hope you will continue making quality contributions. Feel free to further improve your accepted article.

Future articles

• Since you have made at least 10 edits over more than four days, you can now create articles yourself without submitting a draft for review. However, you may continue submitting drafts to Articles for creation if you prefer.

• Once you have made at least 10 edits and had an account for at least four days, you will have the option to create articles yourself without posting a request to Articles for creation.

Next steps

• Find other Wikipedia articles related to your topic and add links pointing to the new article. This helps readers find it.
• Group the article with similar subjects by adding relevant categories.
• You may consider nominating a fact from the article within the next 7 days to appear on the Main Page's Did you know section.
• Wikipedia is a work in progress. You can continue to expand and improve the new article.
• The new article will become eligible to be indexed on search engines once patrolled or after 90 days; search engines may take time to reflect this.
• For friendly peer support regarding editing, sourcing, or policies, visit the Teahouse, a question and answer hub for new editors.

Thanks again, and happy editing!

Sulfurboy (talk) 10:01, 15 August 2026 (UTC)

Fixing failed pings

This edit won't send a ping, see WP:PINGFIX. fifteen thousand two hundred twenty four (talk) 02:16, 16 August 2026 (UTC)

Oh!

Discord has a similar behavior and it bites me there all the time. Thank you for pointing this out! I think maybe the best fix is to delete my previous edits and start clean: and get it right this time. M kuhner (talk) 02:26, 16 August 2026 (UTC)

btw there's WP:DISCORD and m:Artificial intelligence/Discord if that's your type of thing :) Kowal2701 (talk, contribs) 23:27, 29 August 2026 (UTC)

I am an assistant TD for a Starcraft 2 tournament league which runs off Discord, so I use it for that. Otherwise, nope! (Similarly with Reddit. I quite enjoy allthingszerg but hardly touch anything else.)

I am a Usenet veteran (showing my age here) and nothing since has worked as well. M kuhner (talk) 03:19, 30 August 2026 (UTC)

I'm only from the time of IRC being the dominant protocol (showing my age here too) and I quite like Matrix, though it's not comparable to Usenet (but the ever-so-terrible email is). ‑‑gurkubondinn 06:51, 20 October 2026 (UTC)

Christian Science dispute

Hi Kuhner,

I posted in the Teahouse last week about a dispute in the Christian Science article where you told me to go through the dispute resolution process and come back to you if I needed help navigating boards, and I think I'm at that point. I posted to the talk page and to their talk page letting them know about the main talk page. We talked on their page, but even after I think 2 days, they never made an objection to the revert I suggested on the Christian Science page, so I went through with it. Today they essentially undid the revert, including the parts I objected to and that Bon Courage agreed about, with no explanation. Given the past warnings that they have gotten and how they continued with the behavior, I don't really see a path forward to resolve this just between us, so I was looking to get someone else involved.

This involves POV, Primary sources (so maybe original research), and an edit war, so I don't really know which board to go to for help, and I'd appreciate you pointing me in the right direction. Thanks- Therealdoomcat (talk) 19:33, 16 August 2026 (UTC)

I think the key question is: Do you think this is a troublesome difference of opinion on a topic matter? Or do you think it is an editor that's not able to work within the collaborative process?

I am hearing that the other editor's communication (or lack of it) is a bigger problem than the topic disagreement. I browsed the admin noticeboards and was pointed at this essay: WP:Responding to a failure to discuss. Down near the bottom is a template for giving a kind of final warning. The essay then suggests either the edit-warring noticeboard or ANI. I think I'd go to ANI, myself. The edit-warring board seems more set up for fast edit wars, and this one is not fast. I do recommend the warning first: there's a strong tendency to criticize people who show up at ANI too quickly.

The four key points about an ANI filing are (a) show diffs of the specific issues--not just a verbal description; (b) be clear, concise, and unemotional; (c) clearly state your preferred outcome (warning? topic ban? block?); (d) immediately post on the other editor's talk page to let them know of the case (there's a template provided for this purpose)

When I have gone to ANI I have usually created the case report in my own sandbox, polished it to the best of my abilities, and only then posted it. Getting the diffs right can take a while. I've had reasonable success, though once in a while a case just won't get any attention, which is frustrating. You cannot try another board at the same time--this is "forum-shopping" and gets a negative response.

Sympathies on the frustration. I haven't been in an edit war yet myself--I do relatively little content editing--but I've been in a lot of fights due to anti-LLM work, and it's exasperating. M kuhner (talk) 00:23, 17 August 2026 (UTC)

Thank you so much for this, the advice is helpful and I never would've found that essay myself. I did also just see that the Christian Science article falls under Wikipedia:Contentious topics. Does that change how things will work very much? If you don't know that's alright. Therealdoomcat (talk) 00:38, 17 August 2026 (UTC)

Contentious topics give administrators permission to be quicker with bans, and to impose limits like "no more than 500 words each in this discussion", but don't otherwise change the situation much. It's worth a brief mention in any case you make ("This article falls under CTOP such-and-such").

The list of CTOPs makes for some fascinating sociology. Some are obvious--some are very much not, at least to me! M kuhner (talk) 00:44, 17 August 2026 (UTC)

Review - think tank page

Hello, you have recently rejected updated to ECIPE page citing lack of reputable sources talking about the think tank, yet a similar tink tank - Bruegel (think tank) has even less references and yet its page is published. Similarly the language and the stye of the presentation is not that different yet it's not accepted. I wonder what criteria are used given that two alike organisation receive different treatment. Nmecipe (talk) 11:04, 18 August 2026 (UTC)

With the number of articles on Wikipedia it is just about always possible to find a worse article. None the less, newly created articles are held to policy standards. This is discussed at WP:OTHERTHINGS.

Bruegel is a 2003 article. Standards were more lax in 2003. If you browse WP:Articles for deletion you can see the community slowly finding and deleting early, poorly sourced articles, but they persist until someone notices them. (There is no formal process for re-review of old, established articles; someone has to take an interest.) On the other hand, it does have some independent reviews of the think tank (not very well utilized in my opinion, but they are there).

I suggest searching for independent reviews of ECIPE; the sources you need may well exist. Don't neglect foreign language and offline sources, if you have a way to search those. You might also ask for help at a relevant WikiProject. M kuhner (talk) 14:22, 18 August 2026 (UTC)

Request to review a Draft

Hi @M kuhner I recently reworked the article Draft:Achal Agrawal. Could you kindly review it when you get time? I have checked each individual claim to ensure that my WP:COI does not lead to WP:OR. I am currently working on cleaning up the mess I created. Thanks again! Flyingphoenixchips (talk) 16:30, 18 August 2026 (UTC)

I don't think I can be the AfC reviewer as I am "involved" as an AINB reviewer. However, I am putting comments on the Talk page. M kuhner (talk) 17:19, 18 August 2026 (UTC)

Hi @M kuhner, I wanted to ask if you could also mark off the articles that I recently recreated after a clean up through AfC. You can take a look at my talk page for them. Appreciate your help, once again

Best Flyingphoenixchips (talk) 02:59, 20 August 2026 (UTC)

Do you think you could make a list? I am a bit overwhelmed right now. M kuhner (talk) 03:16, 20 August 2026 (UTC)

Hi @M kuhner, sure thing, no worries. Hope this would turn out to be helpful for all the WP:AINB reviewers. See User:Flyingphoenixchips/Recleaned Articles Reviewed via AfC Flyingphoenixchips (talk) 03:37, 20 August 2026 (UTC)

AfC August Backlog Drive: Only ten days left!

Our pending drafts!

Hi M kuhner, thanks for participating in the Articles for Creation August Backlog Drive! We've done amazing work so far, dropping the backlog by more than 3000 drafts already. We have around 1300 drafts outstanding, and we need your help to get that down to zero in the next 10 days. We can do this, but we need all hands on deck to help make it happen – if everyone who signed up to the drive reviewed just 1–2 drafts a day for the next ten days, we'd be down to zero in no time! A list of the pending drafts can be found at WP:AFCSORT, where you can select submissions in your area of interest, or you can find a random submission here. Thank you so much for your work so far, and happy reviewing! MediaWiki message delivery (talk) 00:20, 21 August 2026 (UTC)

Draft:Novomind

Hello M kuhner, I hope you're doing well. A couple of days ago you commented on the Novomind draft. I have since replied, you can find my reply here: . I would very much appreciate your thoughts on this. Best regards, Daniela ~2026-45713-02 (talk) 06:22, 27 August 2026 (UTC)

The only thing I can offer is a word of advice: When a source is not easily available, or not in English, reviewers will often go with what you cite it for in the article. If you cite a source for only routine facts, it will be assumed to contain only routine facts. See if you can bring more of the rich sourcing you describe in your comments onto the page.

On policy I don't re-review articles, but I wish you the best with future reviews. M kuhner (talk) 20:50, 31 August 2026 (UTC)

CC-BY-SA attribution

Hey there. I'd made a mental note to tell you something, but now I can't remember where I saw it. I thought that it was something I'd seen on AIN, but I can't find it there.

You'd say somewhere that the CC-BY-SA attribution could be provided on talk pages, but that's isn't completely correct. The WP:TFOLWP attribution needs to be in the edit summary, because it is inseparably (and technically) linked to the addition of the text into the wiki. There is the {{translated page}} notice that can be placed on talk pages to supplement the attribution. It's more visible and less prone to getting lost in the page history, but it's more for informational purposes and doesn't actually satisfy the legal requirements. It is possible to give attribution afterwards in a subsequent edit summary, but as far as I understand, it is less than optimal but still satisfies the license requirements. I'm not an expert on this, so don't completely or uncritically take my words for it though.

Also thanks for your hard work over at AIN. I have a lot of respect and admiration for you and work wikiwork. ‑‑gurkubondinn 10:04, 4 September 2026 (UTC)

A barnstar for you!

	The AI Cleanup Barnstar
	M kuhner, did you know that you are currently the most prolific participant on the AI noticeboard by contribution volume? Thank you for all of the time and effort you have dedicated to protecting the integrity of content and discussions on Wikipedia by responding to reports of LLM misuse. — Newslinger talk 16:14, 10 September 2026 (UTC)

Thank you!

What would you think of having a drive for LLM cleanup similar to the recent AfC drive? We are drowning. I keep an offline log and about 2/3 of cases go to archive without being resolved. DVRTed has made some exceptionally good tools for this job, and that's nearly doubled my speed, but the problem is too big. Cases that are caught quickly may be feasible because very recent edits are quick to remove, but cases not caught for a year or so are a different story. Due to heroic efforts by Kowal2701 we generally get article creations sorted, but not edits.

Also--feel entirely free to say No here--is it possible to apply for rollbacker without the month of Recent Changes patrol? If we had jumped on Orokbest1 and GodswillSE instantly we could possibly have mass-rollbacked most of their edits--it's too late now due to bots. I feel that my AINB work shows my ability to handle reversions responsibly, and I don't want to get distracted. M kuhner (talk) 17:14, 10 September 2026 (UTC)

The cleanup drive sounds like a great idea. I think giving Category:AI noticeboard open cleanup cases more visibility, perhaps by creating a unified tracker similar to the ones in WP:CCI § Open investigations and WP:SPI § Cases currently listed at SPI, would help draw attention to where help is needed.

The month of recent changes patrol wasn't required when I applied for rollback many years ago, and I now see that the requirement was added in November 2023, but is not mentioned in the actual rollback guideline. Based on this, I don't consider the month of WP:RCP a hard requirement for this permission. If you apply at WP:PERM/R on the basis of intending to mass-rollback LLM-generated edits when there is consensus to do so on WP:AIN, which would qualify for point #5 of WP:ROLLBACKUSE, I believe there is a good chance the reviewing administrator would consider your track record on WP:AIN in lieu of RCP activity. — Newslinger talk 18:01, 10 September 2026 (UTC)

You were right! Thanks for the advice. Hopefully won't need it too often. (It's too late for the editathon cases, alas.) M kuhner (talk) 05:17, 12 September 2026 (UTC)

Rollback granted

Hi M kuhner. After reviewing your request, I have enabled rollback on your account. Please keep the following things in mind while using rollback:

• Being granted rollback is no more momentous than installing Twinkle or Ultraviolet. It just adds a [rollback] button next to a page's latest live revision. It does not grant you any additional "status" on Wikipedia, nor does it change how Wikipedia policies apply to you.
• Rollback should be used to revert clear and unambiguous cases of vandalism only. Never use rollback to revert good faith edits. For more information about when rollback is appropriate, see Wikipedia:Rollback § When to use rollback.
• Rollback should never be used to edit war, and it should never be used in a content-related dispute to restore the page to your preferred revision. If rollback is abused or used for this purpose or any other inappropriate purpose, the permission will be revoked.
• Use common sense. If you're not sure about something, ask!

I'm sure you'll do great with rollback, and feel free to leave me a message on my talk page if you run into trouble or have any questions about appropriate use of rollback. If you no longer want rollback, contact me and I'll remove it. For information on rollback, see Wikipedia:Rollback. Good luck and thanks! Staraction (talk · contribs) 04:35, 12 September 2026 (UTC)

Mass creation and spam cleanup needed

Hey @M kuhner, reaching out to get your thoughts on something. I recently nominated a batch of articles at Wikipedia:WikiProject_Deletion_sorting/Manipur, but there appears to be a lot of them.

The main issues across the board are a complete lack of notability and heavy WP:SPAM. Do take a look at strange pages like "Bibliography of Meitei culture". A single user has created over 596 of these articles, meaning there are probably hundreds more that still would need cleanup or deletion.

Since you are involved with cleanup, I wanted to let you know. Flyingphoenixchips (talk) 04:30, 13 September 2026 (UTC)

Thank you. I will have a look and see if I can find the appropriate board, assuming LLM is not involved, or go to AINB if it is. M kuhner (talk) 04:36, 13 September 2026 (UTC)

The instructions at Deletion_sorting/Manipur say that this does not actually create a deletion case--is that what you meant to do? You need WP:Articles for deletion for that, if so.

Can you point me to a particularly egregious one to start my investigation? Thanks. M kuhner (talk) 04:39, 13 September 2026 (UTC)

Thanks @M kuhner. Well I did nominate them through the standard Articles for Deletion route. I only linked to the Manipur deletion sorting page to give you a quick overview of the articles I had tagged.

The following articles might be of interest to you:

Meitei people in North America, Meitei people in South Korea, Mit (Meitei letter)

The worst article in this case is Bibliography of Meitei culture Flyingphoenixchips (talk) 04:44, 13 September 2026 (UTC)

Ah, okay, thanks! I hadn't seen them at AfD but I visit erratically. I'll have a look. M kuhner (talk) 04:59, 13 September 2026 (UTC)

This seems at least somewhat related to the § A misunderstanding discussion in my talk page archives. ‑‑gurkubondinn 13:16, 13 September 2026 (UTC)

While I was wondering what to do, someone else opened a case at AINB: WP:AINB#User:Haoreima. There are several people responding and hopefully we can reach a consensus what to do. I believe you are allowed to post to AINB and your information could be helpful in that thread. M kuhner (talk) 13:17, 13 September 2026 (UTC)

Hey thanks @M kuhner. I gave my comments in the noticeboard as well! Flyingphoenixchips (talk) 16:18, 13 September 2026 (UTC)

I have a piece of unsolicited personal advice. At ANI both the person complaining and the person being complained of will be scrutinized. This is why we recommend, as chessplayers say, "sitting on your hands" (i.e. don't touch anything). Provide no ammunition for criticism of you to derail the case. In particular, delete nothing, don't respond to the person being complained of anywhere outside ANI, and only respond on ANI to clarify facts.

(In a chess tournament a few years ago I reached out to make a move, blanched, and literally sat on my hands. My opponent, who was about eight, found this extremely funny and had trouble maintaining the silence required in a serious tournament! If I had made the move I was planning, I'd have lost instantly.) M kuhner (talk) 05:02, 14 September 2026 (UTC)

Yes def a fair point! Thanks for this, and ofc :) Flyingphoenixchips (talk) 05:14, 14 September 2026 (UTC)

I have stopped listing and commenting on those AfDs for now Flyingphoenixchips (talk) 05:15, 14 September 2026 (UTC)

My sibling is an accomplished chess player, and it does not surprise me at all that you are a serious chess player. There's lot of types of chess players, but the rounded ones seem to have a very good temperament in general. ‑‑gurkubondinn 10:52, 14 September 2026 (UTC)

August 2026 AfC backlog drive barnstars

		The Tireless Contributor Barnstar
		This is awarded to M kuhner for accumulating more than 194.5 points during the August 2026 AfC backlog drive. Thank you for sharing your time, knowledge and dedication; together, we reviewed over 12,700 drafts during the drive and brought the 3+ month backlog down to just a single week, which we couldn't have done it without you. In solidarity, nil nz 02:13, 16 September 2026 (UTC)

September 2026

Hello! Regarding your recent edits to Lieutenant Governor of Arizona when you modified the page, your revert caused an issue. While the issue of LLM on wikipedia is a serious one and your efforts are appreciated, please be careful about throwing the baby out with the bath water. In this case your revert removed valid code from the Infobox and replaced it with invalid code. This means that valid data was effectively removed from the Infobox. You are not the first, nor will you be the last to make this mistake. In the future, when you revert, please double check that you aren't reverting good edits. This can usually be achieved by previewing your revert to make sure there are no preview warnings. Note I have likely fixed the error by now so check the history of the page to see how it was fixed. If you have any questions of if I can help explain further, please let me know! Again, thank you for fighting the good fight when it comes to dealing with LLM material. Keep up the good work. Zackmann (^{Talk to me}/_{What I been doing}) 23:25, 23 September 2026 (UTC)

Sorry, caught a lot but apparently missed that one.

I was going to write you with a question, though: might as well ask it now. We are looking at a couple of ENORMOUS cases (too big to make a tracker = over 1200 articles). I applied for, and got, the right to use mass rollback if I have to. But I can just about guarantee breaking some infoboxes if I ever use it. Is there a way to run a script immediately behind the rollback to catch, or at least flag, when I've done so? No point in mass rollback if it has to be paired with hand examination of each article for infobox parameter errors. I can hand-fix if I know which ones to check--in recent cases it's been about 1 article in 10, so 1200 articles would generate 120 fixes, which is painful but possible. However I can by no means click into 1200 articles to see which ones broke.

(So far I have a healthy--not to say, enormous--fear of mass rollback and have only used it once. Immediately discovered that the admin who gave us the case had used it too, so mine was an anticlimax!)

Thanks for all you do. M kuhner (talk) 23:50, 23 September 2026 (UTC)

Are all these articles in the same category? I.E. all about office holders? Or all about cities? Zackmann (^{Talk to me}/_{What I been doing}) 00:00, 24 September 2026 (UTC)

In the three cases from the African editathon they were all bios, many but not all officeholders. (99% of the problems were "alma_mater".) In the current monster case, alas, it is a mix--some bios, some places, some movies. The places are the worst, often with 5-10 bad parameters.

Is there any possibility of just saying "Of this list of articles, which ones have infobox parameter errors?" My editathon cases would have gone much faster if I could just have made a fixing pass through the 10% that were broken post-reversion, rather than clicking "edit" on each and every one after reverting it.

Or is there a way to make the error messages appear when viewing the article, without needing to preview? M kuhner (talk) 00:11, 24 September 2026 (UTC)

Unfortunately these error messages are preview warnings which ONLY show up in preview... Zackmann (^{Talk to me}/_{What I been doing}) 00:17, 24 September 2026 (UTC)

wrote you a little script

User:DVRTed/check-preview-errors.js. It'll add an option "Check preview errors" to your Tools section. These "preview errors" are almost always unknown/deprecated param warnings from infoboxes. You can test it here where there are two errors: Kermit Davis Sr.. Note that it currently only checks errors in the current revision of the page, even when browsing an older diff or revision (I've exhausted myself with the AINB stuff, so I'll fix that in the future). Oh and WP:TPS, if that wasn't obvious. — DVRTed (Talk) 01:21, 24 September 2026 (UTC)

Re:2005 New Zealand general election

Dear User:M kuhner, thanks for your message on my talk page. I'll be happy to have a look! Kind regards, Anupam^Talk 18:55, 25 September 2026 (UTC)

LLM tag of Niallia tiangongensis

I just saw that you LLM tagged Niallia tiangongensis because it's on my watch list and it seems like the link you provided doesn't lead to anything. Could you tell me where that discussion is happening? Laura240406 (talk) 12:41, 27 September 2026 (UTC)

AINB is aggressively archived to keep it from getting too big. From the archives, here's the first Fuzheado case and here's the second . M kuhner (talk) 14:30, 27 September 2026 (UTC)

thanks :) Laura240406 (talk) 14:59, 27 September 2026 (UTC)

On examination, Gnomingstuff has deleted the majority of the clearly LLM text. I am going to un-tag it. I was doing a large number of these taggings to try to clear up the archived cases as much as possible, and I missed Gnomingstuff's edit. M kuhner (talk) 15:20, 27 September 2026 (UTC)

Dann Coakwell and Draft:Michele Kennedy

Hi @M kuhner:. Following up on the flagged drafts: Kowal2701 (talk · contribs) recently reviewed and approved the cleanup for Dimitrios Vassilakis, which was moved back to mainspace. I have completed a total rewrite for two more articles from the list: Dann Coakwell and Draft:Michele Kennedy.

For Dann Coakwell: I moved it back to mainspace after completely rewriting it from scratch into human-written prose, restructuring the text, and purging broken references (including an invalid NYT citation). However, a patroller re-added the AI maintenance tag. Since the article is now fully cleaned up and verified, could you please review it and remove the tag if it meets mainspace standards? For Draft:Michele Kennedy: It has been similarly rewritten from scratch, removing all rigid date-based event listings, promotional tone, and broken links, while consolidating the structure into clean, natural paragraphs. Could you please review both when you get a chance and let me know if Michele Kennedy can be moved back as well? Thanks for your time. --Lukasmeier88 (talk) 22:10, 3 October 2026 (UTC)

Hi Lukasmeier88,

Thanks for your work on these articles. I have put notes on Dann Coakwell. There are still some source/text inconsistencies and LLM-isms, and I can't support untagging at this time.

I'm going to request that you look at that article and then apply the same approach I used (looking at every claim and testing it against its sources) to Draft:Michele Kennedy. Once you do that, I'd be happy to do a review of that article too.

You have my sympathies. My first project on Wikipedia was doing "a little copy-editing" on what turned out to be someone's massively LLM-edited article: took me over a month to finish what the Newcomer Tasks told me would take 5-10 minutes. The pervasive small source/text issues with LLM are hard to believe until you've seen them. (Though my adopted article also had a big table of plumage colors which was completely made up. I had a lot of trouble believing that at first, it sounded so confident....)

As a word of warning, source/text is actually easier for an outsider. Someone who knows the topic might know that the claim is true and then miss that it's not in the source. You have to put aside what you know and doggedly ask "does the source confirm this?"

M kuhner (talk) 00:13, 4 October 2026 (UTC)

Hi M kuhner,

Thank you so much for your time, exceptional patience, and detailed guidance throughout this review. Your precise feedback was invaluable in helping me understand where the article fell short of Wikipedia's standards.Following your advice, I have thoroughly revised the draft and added several new high-quality sources to ensure full compliance:

1. New Independent Sources & Verification: I have completely removed user-generated directories (such as Operabase and Bach Cantatas) and promotional bios. To replace them and strengthen the draft, I incorporated new independent secondary sources and music journalism reviews (including The Classic Review, Seen and Heard International, Milwaukee Journal Sentinel, and The Berkshire Edge) along with Naxos.

2. Strict Sourcing for Specific Claims: Unsubstantiated details, including specific teaching tenure dates that lacked direct source support, have been removed for 100% verifiability.

3. Neutral Tone: I stripped away all promotional language and AI-generated adjectives ("unmatched", "vocal flexibility"). Critical reception is now presented strictly through concise, neutral summaries of published performance and album reviews. I hope these updates bring the article in line with what you were looking for. Whenever you have a moment, I would be very grateful for your thoughts or further feedback on this updated version.--Lukasmeier88 (talk) 17:38, 4 October 2026 (UTC)

sorry but did you use an LLM to assist you with writing this comment? It has several signs, and Pangram says it's 100% AI Kowal2701 (talk, contribs) 19:52, 4 October 2026 (UTC)

LLM discussion on AI noticeboard

Hi @M kuhner, I have come across mass revertions related to Wikipedia:AI noticeboard/2026-09-25 Vncdengc, since a few of these election related articles are on my watchlist. As I understand the concern and the context through the conversation on WP:AINB#User:Vncdengc, I wanted to approach you, so in what manner I can as a fellow editor concerned with these election related articles proceed to restore any content that purely involves formatting changes, after careful consideration, altho I am not personally sure if it generated by LLM models or not, following the Wikipedia:Presumptive removal of AI-generated content#Reversing removal of content guidelines, since this is my first encountering such a scenario where it is alleged that a user has used a LLM model for generating content on Wikipedia. Please guide me accordingly and let me know if I am missing anything or could be of any help to you. Thanks 456legendtalk 10:07, 4 October 2026 (UTC)

The general rule is that you can restore the material if you have vetted it and take personal responsibility for its correctness. "Correctness" here means that there is no LLM-generated verbiage remaining, and that all claims have been checked against their sources and all source citations are correct.

This is very difficult for articles with a lot of added words, but not so hard when the edits consisted of formatting, since correct formatting is the same whether written by LLM or by human. So I'd say that you can revert formatting-only edits once you have checked that the formatting is correct. Please watch out for any edits that added sources, wikilinks, or external links as these were often erroneous--the editor tended to wikilink to disambiguation pages.

Please use a clear edit summary explaining your restoration of the reverted material so LLM patrollers know what happened.

It's regrettable to use mass rollback, we hardly ever do, but there were over 1400 articles involved and we don't have the volunteer manpower to hand-check that. I know there were correct edits among them and would be delighted to see those salvaged. M kuhner (talk) 15:21, 4 October 2026 (UTC)

Thank you for the insights, noted them. Very much appreciated. 456legendtalk 16:49, 4 October 2026 (UTC)

Request on 16:15:25, 7 October 2026 for assistance on AfC submission by Demian Incognito

• Demian Incognito (talk · contribs)
  • Draft:MyNetDiary (edit | talk | history | links | watch | logs)

Hi! Thanks for reviewing Draft:MyNetDiary. The previous version was declined partly because it appeared to be LLM-generated. I'm now working on a revised version before resubmitting it. If you have a moment, could you point out which passages or features gave that impression, so I can avoid them further?

Demian Incognito (talk) 16:15, 7 October 2026 (UTC)

In general, watch out for empty statements that say that there was coverage in a particular venue, but nothing about what the coverage actually said. The body of the article is improved now, but the lead still has this issue. Remember the lead is "what are the most important facts about this topic?"--it is what people will see in a mouseover, or on their phone. What does a reader learn from knowing that the product was reviewed by a particular newspaper or magazine?

Don't just fix the wording, though. LLM text tends to misrepresent its sources and you will need to check every claim against the cited source. This is actually hard for the author, because you may know the statement is true and neglect to verify that it's in the source you cite (and not, as is common with recent LLM, one of the other sources in the article).

The last bit of the current draft skates uncomfortably close to violating WP:SYNTH. You include general material about this type of product, citing a source that doesn't discuss this specific one, with the intent of implying that the general material applies. (To be fair, it very likely does! But we're supposed to report what sources say, not build new arguments on that basis.) This is not an LLM issue necessarily, but it's something to think about carefully. You would be on much more solid ground with a source that actually mentions this product. (I didn't read the source--if it does mention the product, you should just make that a bit more clear and you'll be fine.)

You will also want to review the notability guidelines: WP:GNG and WP:NSOFTWARE are relevant. I did not evaluate notability, since LLM articles are never acceptable; but the next reviewer will. M kuhner (talk) 16:36, 7 October 2026 (UTC)

Draft I'm working on

I know you're one of the more involved WikiProject AI cleanup members. Any chance that extends into being interested about how AI affects the world as a whole? I'm working on Draft:Toys with LLM integration which I'll probably move it to AI toys when it's more fleshed out. Not sure yet. But the draft is still in the early stages. It's difficult to write overview articles, but I don't want to have to write several about different brands. Anyways, the more editors the merrier. :) Clovermoss🍀 (talk) 01:19, 12 October 2026 (UTC)

I'm going to put an idea or two on the Talk page. M kuhner (talk) 06:50, 12 October 2026 (UTC)

Temporary account IP viewer granted

Hello, M kuhner. Per your request, your account has been granted temporary-account-viewer rights. You are now able to reveal the IP addresses of individuals using temporary accounts that are not visible to the general public. This is very sensitive information that is only to be used to aid in anti-abuse workflows. Please take a moment to review Wikipedia:Temporary account IP viewer for more information on this user right. It is important to remember:

• You must not share IP address data with someone who does not have the same access permissions unless disclosure is permissible as per guidelines listed at Foundation:Policy:Wikimedia Access to Temporary Account IP Addresses Policy.
• Access must not be used for political control, to apply pressure on editors, or as a threat against another editor in a content dispute. There must be a valid reason to investigate a temporary user. Note that using multiple temporary accounts is not forbidden, so long as they are not used in violation of policies (for example, block or ban evasion).

It is also important to note that the following actions are logged for others to see:

• When a user accepts the preference that enables or disables IP reveal for their account.
• Revealing an IP address of a temporary account.
• Listing the temporary accounts that are associated with one or more IP addresses (using the CIDR notation format).

Remember, even if a user is violating policy, avoid revealing personal information if possible. Use temporary account usernames rather than disclosing IP addresses directly, or give information such as same network/not same network or similar. If you do not want the user right anymore then please ask me or another administrator and it will be removed for you. You may also voluntarily give up access at any time by visiting Special:Preferences. Happy editing! — rsjaffe 🗣️ 22:22, 16 October 2026 (UTC)

Regarding an article

May I know on what basis did you add AI tag to Raghuttama Tirtha article. And also Did you verify all the citations? May I What copyright violations did you find in it? - Garudindflamewing ^(talk) 05:44, 19 October 2026 (UTC)

I did not do a detailed source/text analysis of this article. We're permitted to assume that if several articles are LLM, others will be as well. But I see that you said you used LLM on some specific articles but not others, so I will do a complete analysis here. It may take me a day or two. I won't be able to check all sources as some are books I don't have access to, but I'll do the best I can.

I do feel, reading the article, that there is some text with a very different style of English than the rest. Given that you have written over 98% of the text, this is a little puzzling.

Please note that the "AI-generated" header is for general use. It does not mean that all of those things specifically have been found in the article: it is meant to list problems that may be in the article because they are common in LLM text. (If I knew that there was copyright violation, I could not just tag, I would have to involve the copyright cleanup crew.) M kuhner (talk) 05:55, 19 October 2026 (UTC)

I have three conclusions.

(1) I agree with you that most of the text of this article is clearly not LLM.

(2) There are a few passages which have very different language than the rest. I do not necessarily think those are LLM either. I think instead they may be too-close paraphrase of one of the sources. One of the most obvious is:

Few commentators in the Dvaita tradition attempt what this text sustains from beginning to end: a double exposition that simultaneously establishes the plain meaning of the Upanishad and dismantles, at every contested juncture, the readings advanced against it. This does not sound like the author of the rest of the article. Can you send me a brief quote of what the cited source says here?

Also this: Monastic chronicles and regional records state that whenever branches extend southward to cast shade over the brindavana, they wither and dry up spontaneously before casting any shadow upon the Brindavana. Can you send me a brief quote of what the cited source says here?

And this: Local agrarian and pastoral communities in the surrounding rural regions of Tirukoilur, irrespective of community or caste background, maintain an active custom of bringing cattle afflicted with illnesses or suffering from bovine epidemics directly to the shrine complex. Can you send me a brief quote of what the cited source says here?

(3) There are a large number of problems with the sources. I checked every ISBN in the "Sources" section and found the following:

Prabhanjanacharya, Vyasanakere (2013) and (2005) are claimed to have the same ISBN. But each edition of a book should have a different ISBN. Also, none of the 8 databases I tried could find that ISBN at all.

Why are both Sharma 2000 and Sharma 1961 cited? They appear to be two editions of the same book? I am not sure it is wise to use the 39 years older version.

For Bruck (2011) a title is given in English, but the ISBN goes to a book in German, titled Leben in der Kraft der Rituale: Religion und Spiritualität in Indien. It seems that the title may have been translated, which it should not be. If there were an English translation of this book it would have a different ISBN than the German one.

Sharma (1971) has a note in it "(2008 reprint)" but the ISBN appears to be for the 1971 publication. The publisher is different; the ISBN should not be the same.

The Oxford India Hinduism Reader is listed as being by Vasudha Dalmia in the article, but looking up the ISBN finds that the book is by Vasudha Dalmia and Heinrich von Stietencron.

This is as far as I can go tonight. Tomorrow I will look at the sources I can find online. But I think this is enough to show a worrying pattern especially with ISBNs that do not seem correct. M kuhner (talk) 07:11, 19 October 2026 (UTC)

A barnstar for you!

	The Editor's Barnstar
	For everything that you do, every page you have cleaned up, and every case that you have slogged through. ‑‑gurkubondinn 06:45, 20 October 2026 (UTC)