Forum:IP reader restrictions

Forums: Index > Village Dump > IP reader restrictions

Note: This topic has been unedited for 14 days. It is considered archived - the discussion is over. Do not add to unless it really needs a response.

Recently a sysadmin, Ciencia Al Poder from Spanish Uncyc, implemented a restriction on IP readers in order to stop AI bots from slowing down Uncyclopedia. Some AI bots read through as many pages on the Internet as possible in order to "learn" or whatever it is they do. This has essentially increased traffic on our site, causing it to run slower, according to some people. I personally did not experience this slowing of the site, and I was active during the time it occurred, but maybe some others did.

The restriction simply stops IP readers if they try to click on a non main space page, such as clicking "View history", giving them an error message instead of loading the page. While this appears to have solved the problem of AI bots increasing traffic, it has made navigating our site virtually impossible for IP readers and editors, and it is certainly a turn off to potential new Uncyclopedians. If someone new to our site wants to explore and see how everything works before creating an account, like I did, they will be stopped by this restriction.

Furthermore, this restriction was implemented without any discussion with the Uncyclopedia community. As far as I can tell, it was just a few brief messages from just a few users in the help section of the Uncyc Discord. This was a huge change to the function of our site that should have required at least a discussion before implementing. I believe we should remove this restriction and only implement it again if we see a more significant slowing of the site, and I would like to hear what everyone else thinks of this restriction. Thank you. Cheers. MrX blow me 22:46, 29 May 2025 (UTC)

TO BE CLEAR the error messages I'm talking about are those "429 Too Many Requests" some of you have been getting when browsing Uncyc. They were implemented INTENTIONALLY. MrX blow me 14:06, 30 May 2025 (UTC)

Could those bots be where all those recent 429 errors are coming from? Oᑭöᔑᔑᙀᙏ (vandalize my talk page) 22:56, 29 May 2025 (UTC)

I believe that is possible. I hadn't gotten any in a long time, but to be fair, maybe others did. MrX blow me 22:57, 29 May 2025 (UTC)

@OPOSSUM: Are you getting those 429 errors while logged in? Those are the very error messages I'm talking about. Those were implemented intentionally. MrX blow me 13:58, 30 May 2025 (UTC)

Ciencia said that there wouldn't be a problem if the bots only crawled article pages since they're cached. I don't know much about backend web development but would it not be possible to serve IP users cached version of non-mainspace pages? It sounds like that could work well based on what I heard from Ciencia (and if my intuition about how websites work isn't completely wrong.) SystemPhantom (talk) 23:07, 29 May 2025 (UTC)

Maybe that could work. In Discord, I've been asking about alternative solutions to the AI bot problem, but Ciencia doesn't seem to think anything else will work. MrX blow me 13:53, 30 May 2025 (UTC)

I've seen around some folks using Anubis (a proof-of-work challenge) to take the scrapers away, wouldn't this be viable although it require server-side configuration? Coco (talk) 23:10, 29 May 2025 (UTC)

Maybe. What all would Anubis entail? The link only offers a little information, unless I'm missing something. MrX blow me 13:54, 30 May 2025 (UTC)

Perhaps we could force all IP users to do a captcha or smth like that so that we don't have to kneecap IP users? Data Devourer (talk) 23:22, 29 May 2025 (UTC)

Yeah maybe. Captcha would be less than ideal but better than what we currently have. I still think we should remove the restriction entirely, then implement a captcha if the site slows down too much. MrX blow me 13:56, 30 May 2025 (UTC)

Just let my ballz Ip run around for an hour or two. if it creates garbage articles then it is most certainly a robot.[[File:Lawrence.gif.gif|70px|link=User:ColonelKurtz]] ([[User talk:ColonelKurtz|talk]]) (talk) 23:24, 29 May 2025 (UTC)

ColonelKurtz is a robot, got it 23:40, 29 May 2025 (UTC)

nnonononono if someone accuses someone else of being a robot the accuser is psychologically projecting[[File:Lawrence.gif.gif|70px|link=User:ColonelKurtz]] ([[User talk:ColonelKurtz|talk]]) (talk) 23:59, 29 May 2025 (UTC)

When browsing the site, I keep encountering "429 Too Many Requests", Imposing restrictions seems like a helpless move? 09:46, 30 May 2025 (UTC)

@C780178: Are you getting those 429 errors while logged in? Those are the very error messages I'm talking about. Those were implemented intentionally. MrX blow me 13:59, 30 May 2025 (UTC)

Not so much anymore. Oᑭöᔑᔑᙀᙏ (vandalize my talk page) 13:58, 30 May 2025 (UTC)

@MrX: I have an account on the Spanish Uncyclopedia. Maybe I could ask the guy that made those restrictions if there are viable alternatives. (and yea I know the guy is on Discord too but maybe it's better to reach through the wiki idk it's just an idea I had) 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 14:12, 30 May 2025 (UTC)

That would be helpful and appreciated! I also have an account there (soy SrX) that I've made like 5 edits with and haven't used in a long time, maybe I'll chime in there too if it helps. MrX blow me 14:14, 30 May 2025 (UTC)

every time im logged out I cant go to recent changes or else my computer acts special and loada 429 error for some reason (talk) 18:13, 30 May 2025 (UTC)

Yeah, that's exactly what I'm talking about. The reason is the restriction that we have implemented. MrX blow me 18:14, 30 May 2025 (UTC)

but couldn't the bots just go through contributions on me for example and just click on every page I have ever read or edoted? (talk) 18:15, 30 May 2025 (UTC)

No, because your contributions page is a "special page" and they get rate limited on those. 18:42, 30 May 2025 (UTC)

Next step

If we want to get rid of the error messages that are preventing people from reading and navigating our website, we need to tell the sysadmins, like User:Lyrithya. If most people want to keep the error messages, um, okay then. I obviously disagree, but I will accept that. Please vote, and please feel free to continue discussing in the section above. MrX blow me 15:34, 1 June 2025 (UTC)

WTF do you mean by "remove error messages"? Do you mean remove the limit that keeps the server stable for everyone else to use? Let me repost what I said on Discord: Remember how the site was constantly crashing or very slow for everybody before this measure was put in place, because ai crawlers were consuming all available resources for the server. If we remove this particular error message, maybe everybody will start getting other error messages, not just accessing history or special pages when not logged in, but also reading normal pages. More important, Uncyclopedia (in English) is being hosted on the same server with other Uncyclopedia languages, which all will be affected, because the slowness on one wiki will impact all wikis. Maybe they have something to say about this, too. Sure, a vote on "Remove error messages" vs "Keep error messages" will get all the votes on the first option, if people have no idea about what's really going on behind the "error message" and why was done in the first place. --Ciencia Al Poder (talk) 21:27, 1 June 2025 (UTC)

And let me repost what I said on Discord and what everyone else agrees with: your solution is a problem in and of itself. “Remember how the site was constantly crashing or very slow for everybody before this measure was put in place”? Again, no, I don’t. Can we at least try removing the restriction? Maybe just for a little bit? If we “have no idea about what’s really going on”, then maybe you can do a better job explaining it to us, and you definitely need to understand that your solution isn’t working, because it creates other problems. Please try and be a little cooler too. I have been polite throughout all this, and I appreciate your help. Thanks again. Cheers. MrX blow me 00:25, 2 June 2025 (UTC)

And let me reiterate that this was a major change to the function of the site without any community discussion. MrX blow me 00:46, 2 June 2025 (UTC)

I have a GENIUS idea! Just have the error message spit out whatever the page was going to be in the first place! NO MORE ERRORS! Donald J. Trump@realDonaldTrump •  23:57, 1 June 2025 (UTC)

I for one believe that blocking access to IPs and spambots do not work since all they have to do is register goddamnit and then they'll be immune. What's next, you're going to delete log in and create account? (talk) 14:39, 2 June 2025 (UTC)

@ColonelKurtz: I don't think these bots can log in, tho? if they could, they'd be raiding this wiki with accounts just to do as you say, yet few recently logged users are spambots, and those that are probably have nothing to do with the bots that are raiding Uncyclopedia (which are bots made to feed data to generative AIs). 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 14:44, 2 June 2025 (UTC)

ye well AI is gettin' more advanced day by day and so the 429 errors does nothing. all the person has to do is register an account for the spambots and let it run free (talk) 14:46, 2 June 2025 (UTC)

They don't care about this website specifically, they are scraping any and every website on the entire internet. — L10nM4st3r ^{ROAR at me!} / _{What my paws scratched} 10:37, 3 June 2025 (UTC)

Remove error messages

Score: 6

1.  For. MrX blow me 15:34, 1 June 2025 (UTC)
2.  half for they're pretty annoying, but I'm unsure if the alternative would be even worse, so I'm voting only half for. maybe I'll change it in the future if that Ciencia al Poder guy explains these things better. 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 15:48, 1 June 2025 (UTC)
3.  For. I support imposing our will on the six other wikis hosted on the server! The American™ Uncyclopedia should impose its will on all others! Donald J. Trump@realDonaldTrump •  16:28, 1 June 2025 (UTC)
4. WTF? (neutral) How the hell did you make the 429 error messages appear intentionally? Oᑭöᔑᔑᙀᙏ (vandalize my talk page) 16:46, 1 June 2025 (UTC)
5.  Penis Why would you make 429 messages appear? That doesn't dissuade the bots at all and it just makes IPs annoyed. (talk) 17:45, 1 June 2025 (UTC)
6.  FORE!!! *hits golf ball* ❄️SystemPhantom, with an 🍊orange for you! (talk)❄️ 17:53, 1 June 2025 (UTC)
7. {{For|What's it called when you decide something because everyone else chose it? Half for, maybe there is a better solution... Like more page caching. — L10nM4st3r ^{ROAR at me!} / _{What my paws scratched} 10:40, 3 June 2025 (UTC)
8.  For. Not convinced of the need for such a restriction. 00:15, 30 June 2025 (UTC)

Keep error messages

Score: 1

 For. fuck ai (until it becomes sapient, then no because they'll be our overlords) Schism (talk) 01:54, 2 June 2025 (UTC)

Update

The error message is now a little more informative, so I guess that is a very, very small step in the right direction. @Ciencia Al Poder: thank you for your help. I hope you can please understand that this is still a problem, not a solution. I hope you can soon enlighten us as to what the plan for solving this problem is. Thank you. MrX blow me 13:48, 10 June 2025 (UTC)

The plan is to involve a second sysadmin. If another sysadmin agrees into removing the restrictions I'll remove it, but I'm not gonna lift it based on the results of this vote, as I already said on Discord. I'm not the only sysadmin managing the server. --Ciencia Al Poder (talk) 16:43, 10 June 2025 (UTC)

Ok, thank you, please keep us updated. MrX blow me 12:49, 11 June 2025 (UTC)

Maybe I haven't expressed it correctly... since I'm not interested in the change, you should find another sysadmin interested in the change, not me. --Ciencia Al Poder (talk) 20:56, 11 June 2025 (UTC)

Ok, who are the other sysadmins and how can I contact them? MrX blow me 21:22, 11 June 2025 (UTC)

@MrX: from the Discord server whose roles are "sysadmins", there's User:Lyrithya and User:Bizzeebeever. apparently there are other three more in this wiki but they seem to be from the old age of Uncyclopedia and I doubt they're around here anymore.
there are probably more sysops on uncyclomedia.org that aren't in here. 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 21:50, 11 June 2025 (UTC)

okay so I'm searching for more Uncyclomedia sysops and so far I've only found this Brazillian user and this user from Flanders who apparently each of them have [[User:Rhubella Marie|their own user account on this wiki. They don't seem to be very active here tho so I doubt they'll come to this forum page. 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 22:06, 11 June 2025 (UTC)

this search is driving me towards a wiki rabbit hole I didn't even know about. apparently there aren't one but two different Uncyclomedia websites, the latter of which doesn't seem to be much active or useful. also there's a third general Uncyclopedia called uncyclopedia.info, plus a fourth website for Uncyclomedia, Uncyclomedia Commons (which I get is like a storage for files but it also has other pages for some reason). why so many different wikis? most of them are unused, so wouldn't it be better to just have one for all the stuff? 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 22:23, 11 June 2025 (UTC)

okay so I've found a vandal on Uncyclomedia Commons from 2022 whose edits haven't been deleted yet. there's also some user who for some reason uploaded unfunny, copyrighted images to the site. someone should probably do something about it. 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 22:27, 11 June 2025 (UTC)

also there are a lot of spam in uncyclopedia.info. why does that even let you have articles in the first place? it's not like there is much needed from that website beyond the main page... 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 22:46, 11 June 2025 (UTC)

the uncyclomedia.co website doesn't seem to have much vandalism and/or spam, but there are just five three pages: global tools, main page, the main page stylesheet, and like two userpages. this website seems like a waste of space, considering uncyclopedia.info already does what uncyclomedia.co does. 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 22:49, 11 June 2025 (UTC)

you've ended up in the weird rabbit hole that is the Carlb wikis, which are maintained (very poorly, i may add) by Carlb. all of them (except uncyclomedia.co) are owned by him and used to be managed by admin teams consisting of basically the two users you mentioned. 23:27, 11 June 2025 (UTC)

btw, i'm pretty sure the guy behind the Rhubella Marie accounts is dead, so i doubt you're gonna get any responses from there if you ever need any 23:28, 11 June 2025 (UTC)

@alula: hold on, you mean like, really dead? not living anymore???
also, uncyclomedia.co seems kinda useless when uncyclomedia.org already exists, and considering that Carlb guy isn't mantaining that website, then why still have it? 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 23:38, 11 June 2025 (UTC)

yes, i do really mean dead. and also, you should ask Carlb why he keeps it up instead of me, because i really don't know 23:40, 11 June 2025 (UTC)

@alula: how is it that they're dead, tho? I doubt that guy was that old...
also, how do I ask that "Carlb" guy? according to this that guy seems to have been vanished for a long time ago... 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 23:44, 11 June 2025 (UTC)

probably due to some health issue. and i guess you could email Carlb? his email address should be lying around somewhere, i just can't be bothered to find out what it is
also, the part about his wikis being very poorly maintained is that he more or less slapped a bandaid to fix an issue after not noticing it for almost a year and has basically not been there to fix literally any other issues, including the ones caused by the bandaid he forgot about... 23:50, 11 June 2025 (UTC)

┌─────────────────────────────────┘
Damn that's a lot of stuff. Thanks Dani and Alula. I'll probably start with Lyrithya. MrX blow me 16:25, 12 June 2025 (UTC)

Another sysadmin's take

Of course the restrictions won't be going away, because we the people sysadmins actually value our human editors and readers. It's annoying, but not quite as bad as the cancer of the Internet.

You can thank this guy, the people wishing to be that guy, and these organizations for the absolute atrocity of a timeline that we live in. (Of course core MediaWiki developers also deserve a tiny bit of the blame, as MediaWiki by default has some very expensive special pages which are accessible to all users, including anons, by default, with no real way to easily restrict access to them. But this wasn't a problem before misbehaving AI crawlers started to crawl all of the Internet with no respect to standards like robots.txt.)

Also, "AI" itself is a misleading term, as "Artificial Intelligence" is only artificial without any intelligence to it. The term LLM (Large Language Model) is a more accurate term to represent what it is as opposed to what the technocrats wish it'd be one day. (Personally I prefer "Bullshit Generator", though.) They're statistical models designed to more or less answer the question, "what word(s) could follow this [prompt]?", at least when talking about text-based models. The same is more or less true for image generators, they rely on human-supplied metadata and such to identify concepts like "bird", "car", "tree", etc. — these models have no internal understanding of what is the meaning of such words. If I told an AI that this is a tree, it would have no problem believing me. You would probably disagree, though.

And then there's the big money question. Nobody wants to be footing the AI organizations' bills if and when they literally are unable to play by long-standing, standardized "rules" of the Internet (and I don't mean just Rule 34 here) and they provide no value—let alone money—to us Uncyclopedians. They just want all the data and they want Someone Else to pay for it all.

I for one am grateful to Ciencia for all the tremendous behind-the-scenes ops work he does, from keeping MediaWiki updated to ensuring that Uncyclopedia remains usable by actual humans as well. Trust me, you don't want me trying to manage these things.

There's a semi-up-to-date list of sysadmins at Special:ListUsers/sysadmin and you can always contact us on Uncyclopedia's IRC channel (yes, we're old and we don't like newfangled technology, can you tell?). As of right now, 50% of the people present there are sysadmins! -- Jack Phoenix, professional killer admin (Whine?) • (Wikia ♥ ads) 23:04, 28 June 2025 (UTC)

@Jack Phoenix: k, thanks for explaining it more in detail. it's a shame how this "AI" (which is not even actual AI) technology has worsened the Internet so much in such a short spam of time. I too agree that we should try to stop these AI crawlers from blowing up Uncyclopedia's servers [although I do wish there was a way that didn't restrict IP users...]
btw, maybe you sysadmins should try moving to Uncyclopedia's Discord server, since nowadays few people use the IRC chat (I'm guessing you already know that as you guys are in there all the time, tho), and that would make you more accessible to the rest of users. even if you don't use Discord that much, at least it could be helpful to have you guys there too. 🎄🎄🎄DaniPine3 (talk)🎄🎄🎄 23:22, 28 June 2025 (UTC)

Thank you @Jack Phoenix: and everyone else who has weighed in here. I still don’t see why we can’t at least try removing the restriction, even just temporarily, especially since this is, again, a major change to the function of the site that had almost zero community discussion or input. Furthermore, again please forgive me for repeating myself, not everyone noticed this alleged slowing of the site. And we clearly have a lot of people who want to at least try removing the restriction. I appreciate what the sysadmins do, but I don’t understand this hardline stance against fixing a problem. MrX blow me 03:23, 29 June 2025 (UTC)

@Jack Phoenix: @Ciencia Al Poder: would making IP users complete a CAPTCHA if they click on a special page work? MrX blow me 03:37, 29 June 2025 (UTC)

It would work, but adding a captcha there requires an amount of development effort that's currently not possible --Ciencia Al Poder (talk) 09:42, 29 June 2025 (UTC)

@DaniPine3: No problem! I think that it's up to us to make technology and related things (like "why was this rolled out?" or "why can't we have feature X when site Y has it?" etc.) accessible and understandable — as much as I for one would wish, not everyone can be or wishes to be a programmer. ;-) I'm on the same boat when it comes being nice towards IPs (in terms of what they can/can't view); but the new reality in which we live is that the "anti-abuse" solutions will likely only get harsher (in general, not on Uncyclopedia) and more discriminatory e.g. towards users not able or willing to update their browser etc. And a lot of people, especially those who aren't programmers and interested in becoming one and just wish to run their small online business or other site in peace, will likely opt into whatever is both the cheapest and easiest to set up, with very little regard to accessibility or such things. How else do you think the cancer of the Internet got so big that it's totally justified to refer to it as "the cancer of the Internet"?
As for Discord...I'm only speaking for myself here as some of our sysadmins are already using it, but I hate it and it hates me. As a piece of software, it's simply mind-boggling to me as to how something so badly implemented has gotten so popular. Somehow the so-called "experts" at Wikipedia have failed to emphasize what I consider the most relevant points: that it's a locked-down platform for which no alternative clients exist, that it's too centralized and that it's a lazy cash grab. Now, obviously millions of people are either not aware of these issues or don't think it's relevant to them; and that's fine. For me, these are a deal-breaker. Luckily I can be reached by a variety of methods (some of which are probably considered somewhat dated by now), such as the aforementioned IRC channel, on-wiki talk pages on a wide variety of different wikis, and even email. :) That being said, the IRC channel is and has been fairly quiet, so should someone wish to build a Discord bridge bot, I for one wouldn't oppose it, but that's not something I'm too familiar with; I know that such things exist, but as for the practical implementation...yeah, no clue about that, sorry.

@MrX: You bring up some interesting and very good points! It's commonly accepted that while we're neither a bureaucracy nor a totalitarian regime, we're also not a democracy either. For that matter, neither is Wikipedia or any other site out there. It's widely understood and, at least begrudgingly accepted that sometimes sysadmins need to pretty awful things to keep the site up and running. When the site is under an active DDoS attack, there's no time nor any reason to be organizing votes and such on whether the operators should be allowed to do something about the problem.
Likewise, it's commonly understood that while things like major software version upgrades may cause brief service availability interruptions, they are for the greater good of Uncyclopedia and world domination I mean, the spreading of Uncyclopedian values across the Intertubes; that's why you haven't seen and won't see us asking, "do we have the political approval for upgrading the site to the latest software version which patches issues X, Y and Z and introduces new features A, B and C?"; the flip side of this coin is that sometimes this will result in some features being lost.

There's an inherent "cost", and not necessarily (only) a financial one, to all the features. That's why it's so neat the Wikimedia Foundation foots most of the bill for us; it allows us to focus on the things that matter as opposed to paying a bunch of nerds for things they should be giving us for free! (Trust me, I'd know.)

If we're doing our jobs well, you won't notice it. That's the dream, anyway. Reality, ehhh...

Put it simply, we like you, and we like you being and editing here. We wouldn't want you ending up on some dubious website(s) simply because Uncyclopedia fails to load due to being hit by massive swarms of AI crawlers, would we now?

As for the idea of trying to remove restrictions...while Uncyclopedia still isn't a democracy, I for one am not opposed to this, provided that we first clearly define "success" and "failure" in terms that we all can agree on. As noted on UN:NOT, we're not infallible (although administratorial infallibility totally should be a thing!) and I'm completely open to the idea that perhaps circumstances have changed and the mitigations that were once necessary are no longer so. But if we're to test this out (because who doesn't love testing in production?!), what is our criteria for this? At what point can we safely draw a conclusion — one way or the other? If the site gets hit by a bunch of Chinese IPs associated with DeepSeek five minutes after removing the restrictions, how would we go about interpreting that? Or if OpenAI crawlers hit us the next day? Or if some indie AI startup hits us with the force of over 9000 crawlers within a week of removing the restrictions?

I understand that this is a somewhat controversial matter, but I'm fairly confident that I speak for the team when I say that we just wouldn't bother making changes if we didn't see it as absolutely necessary.

As for the CAPTCHA-on-special-page idea, that's a fascinating idea to explore, actually. I don't want to get your hopes up unnecessarily, but...watch this space! -- Jack Phoenix, professional killer admin (Whine?) • (Wikia ♥ ads) 21:01, 29 June 2025 (UTC)

The criteria for success vs. failure weren't defined at all (to my knowledge) before implementing this, so any criteria would be better than what we have now. I've never said that I know for a fact that removing the restriction wouldn't lead to a significant slowdown. The number of AI bots that "attack" after removing the restriction shouldn't matter, how much slower the site becomes is what matters. It might get slower within five minutes, it might take five days, or it might not happen.

My proposal is to remove the restriction, then if a higher number of editors than last time report the site getting significantly and annoyingly slowed down, we reimplement the restriction.

And I know that Uncyc is not a democracy, at least not for some matters, that's why I've emphasized that this was implemented without any discussion. I only opened a vote because that was the best way to gauge consensus. I also know the sysadmins sometimes have to do something to prevent a catastrophe. This type of thing still should have been implemented after a discussion, because we had not reached a catastrophe, but that's water under the bridge now. This type of thing definitely could have been accompanied by some sort of notice from the sysadmins after it was implemented, something like, "Hey guys, we had to restrict some IP access to the site a little bit, shouldn't be too much of an issue, might actually deter new editors from creating an account though, but here's why we did it, here's our plan for the future, etc." MrX blow me 15:41, 30 June 2025 (UTC)