User talk:Schiste/what-now

Just tricking you to come here

Yes I know that was sneaky, but it's just a way to make that damn link blue :)

Let's talk and move forward! schiste (talk) 23:24, 9 January 2026 (UTC)

Thanks for writing this, quite thought provoking and needed essay.

Coincidentally, your point about the site being optimised for desktop users is true to this essay, as well: the talk button does not show up on mobile (if you are not logged in) and if you use apple private relay you get blocked…Bence (talk) 00:55, 10 January 2026 (UTC)

+1 - This is a really important essay and I could not agree more with your points about expanding Wikipedia Enterprise and acting with urgency. It is going to be a even more critical part of how we fund the movement in the coming years. The good news is that it is growing, but there is so much more we can do as you say. Small point: Fundraising banner revenue is about 1/3 of our revenue, not most. FUndraising revenue is about 85-90%. However, I completely agree that pageviews are important because they are a critical way for us to bring in new donors. It is also a big unknown if donors will continue to give through any channel if they only encounter Wikipedia content though an LLM and stop visiting wikipedia.org. We are not the only source cited on LLMs and people are already paying $20 a month for the service. We are not alone in this. Lots of publishers are struggling with it and we are talking with them. I do think Enterprise is going to be more important going forward and I would love to talk more about your ideas for it. Lgruwell-WMF (talk) 01:23, 10 January 2026 (UTC)

@Schiste: +1 Consider me tricked, Christophe. :) Thanks for taking the time to share a broad and long-term view of how the movement has been developing and for your perspectives and data on what risks we collectively face. I have already gone ahead and started to share this essay with some of my colleagues in the Community Resources and Partnerships team, as I think it is worth careful reflection, especially around the kinds of impact that will be important for us to encourage and more actively support. Related to our work around funding and resource distribution in the movement, I wanted to note a few things:

• Several Wikimedia affiliates have been long engaged on some of the very challenges named in this essay. For example, when considering whose knowledge counts, the aptly-named user group Whose Knowledge? has been long engaged with this very challenge and how best to support marginalized communities and knowledge around the world to be more fully represented in the work we do together. They often and correctly point out that marginalized communities represent a majority of the communities in the world, and engage meaningfully with those communities through events and activities (such as their longstanding VisibleWikiWomen campaign) to support their representation and participation in the movement, and yet we still have a lot of work ahead of us on engaging this marginalized majority.
• It might mean, and I know this is controversial, considering whether some kinds of work should be compensated. This may be less controversial than one might think. The reality is that our funding programs have long been compensating people or certain kinds of work that relate to Wikimedia movement activities, as well as the needs to adequately run and manage affiliate organizations and their programmatic work. In general, our funding policies stipulate that we can sometimes provide compensation for work that doesn't replace volunteer activity in a given community. This means we habitually do not fund what most community members would consider paid editing even if it is consistent with local project policies, but do fund other tasks that volunteers are not doing as a matter of practice, often due to the specific expertise or time commitment needed for that work. These include Wikimedian-in-Residence-type roles and as well as other tasks related to project management, graphic design, certain kinds of translation and interpretation work, digitization efforts, and training of volunteers, among other examples. However, is there is a suggestion here that the Foundation consider supporting paid editing more broadly? If so, what might that look like in practice, and how might the well-known issues around paid editing be addressed?
• Movement funds are being more equitably distributed to projects and affiliate organizations across regions compared to 2021, when our team adopted a qualitatively different strategy around funding partly focused on this need. From that report: in FY 2020-2021, before the new strategy began, 60% of funds were distributed to North America (NA) and Northern & Western Europe (NWE). While funding has increased for each region, funds distributed in NA and NWE have grown more slowly over the last four years. Funding to Latin America and the Caribbean, Sub-Saharan Africa, East Asia and the Pacific, and Central and Eastern Europe & Central Asia made up 52% of all General Support and Rapid Funding in 2024-2025, up from 36% in 2020-2021 but similar to 2023-2024 (51%). This is not to say that work is complete, but that our progress around budgeting and working with Regional Fund Committees shows good signs of moving in a consistent, productive direction around funds distribution.

Thanks again for a thoughtful essay. Take care, I JethroBT (WMF) (talk) 05:17, 10 January 2026 (UTC)

Thanks! And thanks for all the feedback. And Whose Knowledge <3 <3 <3.

On the rest, I think that yes we have some good initiatives but we are not as a movement committed supportive enough of them. schiste (talk) 09:12, 10 January 2026 (UTC)

very interesting essay! it's put a couple of my Wikipedia anxieties into some elegant (if a bit dramatic) prose. i guess i'd like to give my two cents on some things that i feel the essay overlooks. i'm pretty bad at putting down my head static into organized thoughts, so my apologies if this reads awkwardly :Þ

re: decline in recruitment--i agree that this is a problem, but i feel like one possible cause is overlooked here. Wikipedia (or at least en-wiki) now covers pretty much every notable major topic to an appreciable extent, and now much of the work left to do in terms of topic coverage is increasingly niche and technical. the low-hanging fruit has been picked years ago, so to speak, and now adding new content often requires background knowledge and access to resources that a lot of people don't have! i overwhelmingly edit in STEM topics, so much of what i source from are journal articles that are very dense to read through and unfortunately often are closed-access (which i am able to access from my university).

assuming that this is indeed a contributor to the decline in recruitment, i think it would serve us well to try and target people who do have the skills and resources to chip in--college students, maybe even researchers. i know that en-wiki already has an education program, but bluntly it doesn't seem to function very well. the work done by students often leaves a lot to be cleaned up, and students don't stick around. i think that "marketing" editing as a way to pick up additional skills is something that could help. editing on Wikipedia has genuinely taught me quite a bit about reference finding and has strengthened my skills in reading and transcribing dense, technical information. those skills could be attractive particularly to STEM students and students in journalism.

i'd argue another major contributor is how difficult it is for someone to learn how to edit Wikipedia. editing is, bluntly, an unwieldy and taxing process. whilst discussing with a friend, they said that they "know quite a few folks personally who have gotten the impression you need to be some kind of tech wizard to edit wikipedia". there's only so much we can do to help that in terms of UI and tools--after all, editing is in a sense website-building. but resources? honestly, the resources for new editors on en-wiki are just,,, bad. guides and how-to's are often information-dense yet at times unconsolidated and scattered about, and it seems like they're having difficulty teaching Wikipedia's principles effectively. the latter point in particular is problematic, because new editors who aren't used to Wikipedia's principles will become entangled in disputes down the line, which can become heated and result in the editor leaving or becoming blocked. we need to figure out a way to make resources for new editors easier to understand and more convenient to access, somehow. ArkHyena (talk) 13:48, 10 January 2026 (UTC)

The person me above shared this essay with me :)

I don't have much time and energy to comment on every issue mentioned in this essay, but I'll drop my thoughts on one specific set of issues.

Firstly, is the fact that the Wikipedia policies and guidelines are really hidden. In my experience when I started, the page for Wikipedia:Policies and guidelines (and other guides related to editing Wikipedia) was never shown to you at all, so new editors are left completely clueless about what contributions would be acceptable. I don't know if there's any plausible way for a new, unfamiliar editor to find these policies in present-day Wikipedia (I doubt most people are aware of Wikipedia-space articles); I think there should at least be some big notification explicitly laying out these policies and asking the user to read them. Experienced but lazy editors like me often revert these newbie edits without a offering a detailed explanation (writing that up would be too much of a hassle to spend my time on, frankly), so I can see how newbies end up very frustrated and confused. Having dealt with this kind of situation many times in the past year, it's really telling that more should be done into getting new users familiar with Wikipedia policies as immediately and effectively as possible.

Secondly, adding on to my first point, is that the Wikipedia policy, guideline and manual of style pages are WAAAAY TOO DENSE. Even as an experienced user, looking through the monstrosity that is the Wikipedia:Manual of Style is daunting because of its sheer size. Doing Ctrl+F is the only way to navigate through this and find whatever you're looking. Regarding the newbies that I dealt with last year, many of them are young (school age) and from primarily non-English-speaking countries, so my attempts at directing them to guideline pages obviously did not work out as they simply did not understand or bother reading them thoroughly. I don't know if there are simplified versions of the Manual of Style and other Wikipedia guideline pages, but I really think that they should exist.

Thirdly, I want to mention that some WikiProjects have their own guidelines/policies which are not mentioned in the main hub of Wikipedia guidelines as I described above. For example, I primarily work in WikiProject Astronomy and it has its own guidelines/policies on notability, naming conventions, etc. There's no centralized "hub" of policies for that specific WikiProject, and I don't think other WikiProjects like WikiProject Elements have that either (or at least, I can't find them readily). I feel like WikiProjects should be enforced to lay out their policies to their front page and make them as accessible/readily seen as possible, though I'm not sure if that's a proper way of handling this issue. Nrco0e (talk) 18:58, 10 January 2026 (UTC)

For simplified Manual of Style, go to the search box at the top of the page and type WP:simplified manual of style. You can also search for the various welcome templates. If the welcome templates are not right for your students, you may be able to request something different. Caffeinated Chihuahua (talk) 03:37, 11 January 2026 (UTC)

If you are not familiar with WikiEd, you may be interested in their training modules. https://outreachdashboard.wmflabs.org/training/editing-wikipedia There is also an events dashboard that can be useful for tracking group outcomes. https://outreachdashboard.wmflabs.org/ Caffeinated Chihuahua (talk) 06:11, 11 January 2026 (UTC)

Here from Stephen Harrison's WIRED article.

I've contributed a little (probably less than 1K edits since 2004, under a few accounts (previous accounts registered with now-dead email addresses).

I have disabilities (cognitive/social/physical function deficits), don't conform to dominant "female" gender/sexual norms, and have found it difficult to navigate community/communication/documentation on Wikimedia project sites that interest me (wikipedia in the past - wikisource, most recently).

My big frustration on Wikipedia has been that the topics that motivate me to edit are ones about which I'm enthusiastic... but it seems that my passion creates a bias that I can't remedy in order to contribute on those topics.

So I switched to proofreading texts that I feel are important to the times we live in (eg Lawrence Lessig's "CODE 2.0" and other books discussing how social/political structures can more equitably represent the inhabitants of those structures; Japanese/Chinese/Korean language books on cultural or textile history). But "help" documents on challenges (eg navigation excessively complicated endnote appendices; lack of tools to facilitate automated translation between languages) have consistently been unavailable or so outdated as to be unusable and overwhelmingly confusing... the remaining editor/proofreading community is as welcoming as they can, but there are no volunteers who consistently focus on training up new volunteers... so I go through sporadic episodes of slogging my way through documentation, trying to understand, and then finding that my proofreading has errors that I must go back through and reverse the changes I've made so I don't end up making more work for others in the future (if ever anyone actually cares about the same books as I do).

Thank you for this powerful and succinct call to action. I hope that the WMF and other power structures here heed it. But I feel pretty discouraged about the possibility of things changing, especially structures to help encourage new volunteers to stay involved and improve their skills... when most of the educational materials in the one WM project where I can make edits and it's clear that my edits aren't "biased" are so desperately old and unrelated to current tools/processes/standards. Grayautumnday (talk) 18:02, 3 February 2026 (UTC)

Note: I forgot to mention that I have laptops, but don't use them. I edit/proof/validate using my phone or occasionally my tablet. The interfaces are consistently ABYSMAL for mobile users. Grayautumnday (talk) 18:06, 3 February 2026 (UTC)

Imagine if I, a reader, could thank you for writing this piece, in one simple click. Instead, I had to try to figure out how to write with these weird colons. I don't even know if I did it right :D And thank you, again. :

Brand vitality

I do actually tell people proudly that I edit Wikipedia. It was a bit awkward at first but it's pretty natural at this point. It helps that it's not a much of a secret anyway. Obviously I'm only one person, but I do think I've made some sort of difference when it comes to visibility. There's a little bit of a ripple effect, I hope. Clovermoss (talk) 05:00, 10 January 2026 (UTC)

Thanks!

That’s the whole point though! “I edit Wikipedia” should be something people brag about, not something you have to work up to saying or feel awkward about.

Like being a badge of honor that makes people go “wait, you can do that? how do I get in?”

We had that energy once. Editing felt like being a pioneer, we were talking about it and promoting it all the time. Because we were the challengers.

The world see us as a commodity, but we behave the part too mostly.

It’s a discussion I think I have late at night at Wikimania Katowice (with Anusha Alikhan I think) with AI content everywhere brand trust and coolness are so important :) schiste (talk) 08:24, 10 January 2026 (UTC)

I think this has a lot of variation around the world. In Sweden, we never had that energy from people who didn't understand how Wikipedia works (which takes some time to grasp), even though I think as soon as you identify in the in-group, people were proud about it. However, as far as I can tell in conversations and public discourse, the trend is going towards it, not away from it, and is increasing the more slop is produced elsewhere. ♥Ainali ^talk_{contributions} 08:48, 10 January 2026 (UTC)

I totally agree with you, this is a highly localized topic, I generalized. And do think we still have a window of opportunity to seize. The more AI content is going to mess up people's digital reality, the bigger the opportunity of being the go to trustable source. But currently, we are not part of that game I think (and again it may change depending on locations and cultures). But what I see, and this is not data, is people going more and more easily to AI Chatbots for verification rather than Wikipedia. Opportunity and threat at the same time ^^10:52, 10 January 2026 (UTC)

seconding this--i've felt far more proud than ashamed to be a contributor to Wikipedia, and reception from folks i've told has actually been pretty positive! i'd certainly like to engage more in encouraging people to join in the project, though; even a lot of my nerdier friends haven't edited. anecdotally, i suspect that the broader increase in burnout and stress plays a role, which only amplifies Wikipedia's learning curve problem. ArkHyena (talk) 12:55, 10 January 2026 (UTC)

It wasn't that I didn't think it was cool/didn't want to brag about it, but because I've often found that people dislike when I'm passionate about anything. My life hasn't been that kind to me. But the past two years or so I've been a bit more comfortable with openly being myself, even if I tend to be somewhat terrified about it at the same time. Clovermoss (talk) 13:09, 10 January 2026 (UTC)

Other avenues of discussions

Creating here a list of other places the discussion is happening :

• Wikipedia subreddit

a path forward

It's a bold idea and worth trying. In addition to that, and something that could be done even quicker, is to reinforce the weight of the strategy recommendations (and the initiatives). All applications for funding should be explicitly aligned with these, and essentially all others should be rejected. Similarly, reporting should be templated around them to reiterate the importance of what needs to be done. Thanks to the thorough process, we already have community buy-in on them and this shouldn't feel a surprise for anyone. ♥Ainali ^talk_{contributions} 08:55, 10 January 2026 (UTC)

What about Wikifunctions?

Hi @Schiste:

How does Wikifunctions (and Abstract Wikipedia) play into this?

Cheers, VIGNERON * ^discut. 12:17, 10 January 2026 (UTC)

Honestly? We’re not ambitious and committed enough.

But there are real capacity and resource constraints. Which is exactly why this needs to be treated as a moonshot, with moonshot-level support.

So to answer your question, this is how I see it: Wikidata + Abstract Wikipedia + a “data confidence model” built on 25 years of edit history and community judgment = a verification infrastructure for generative AI.

We have something no one else can build: not just content (everyone scraped that already), but a quarter century of humans debating what’s “true”, what’s contested, what’s stable, what changed. That’s gold. That’s the trust layer the AI world desperately needs.

And here’s the fun part: this isn’t charity. Frontier labs would probably fund this. OpenAI, Anthropic, Google: they all have the same problem (hallucinations, trust, grounding) and none of them can solve it alone. We can. It serves the mission AND its sustainable revenue.

But, and this is critical, none of it works without massive investment in editor tooling, recruitment, and retention.

The humans are even more core to the product than today. Lose them, lose everything.

Anyway, you asked the right question if you wanted me to talk all night :) You know firsthand I can go on forever. Happy to dig deeper anytime. (Next week?) schiste (talk) 13:15, 10 January 2026 (UTC)

Oh and the verification thing is just an example of how to make it a product that could foster mission and get revenues.

But there are so much more things, and you know that much better than I do, we could do even “just” making data confidence and maintenance scalable. schiste (talk) 13:19, 10 January 2026 (UTC)

Every day communities are still dealing with various backlogs of new software changes from years ago. For years, software changes have been rolled out without providing practical support or simplifying the work for communities. In practice, this means a massive drain on community resources spent on cleanup as a result of those changes. If we really want to go forward, start with solving the older issues, before creating another round what probably would result in extra community work, for what we have no resources. Romaine (talk) 18:13, 10 January 2026 (UTC)

Interesting, thanks Schiste.

The commitment could be bigger but it's already quite high (there is a team of 15 employees for instance).

The main goal is to generate content without AI to avoid all the problems AI have, so I never though about it to verify AI. Not sure how it could work but this is definitely something worth exploring (DVrandecic (WMF) DVrandecic (WMF)).

I fully agree with the human part, this is definitely something here we should invest more (the Wikifunctions community is still quite small, less than 100 people). It is also related to what Romaine is saying, Wikifunctions could help a lot to treat the backlog but it will not "automagically" happen.

Cheers, VIGNERON * ^discut. 13:38, 11 January 2026 (UTC)

Happy to talk about it. From a conceptual standpoint it would be easy, from a technical standpoint I think, but it needs more thinking with more people more competent than me ^^. In my "mind" the way I see some parts of it is Wikidata, Abstract and a data confidence scoring system baked on top of Wikipedia. And that confidence scoring computed from editorship signals (that's a topic that we tried to tackle many times but never succeeded but it's not impossible, we just never had a reason to absolutely crack it). And that way you can build a product other AI products can use as an automated "fact checker". "Chat GPT spewed facts A, B and C WikiScore tell me what's your confidence scoring" "Fact A is Paris is the Capital of France, I'm 100% sure of me. Fact B is that there are 120 millions French in Paris, I'm 0% confident" etc. I'm simplifying, but if we get to a more technical level it could be used as a temperature corrector for LLMs to help discard horseshit states. Not sure I'm clear, even in French I'm not always using the exact precise word for each concept. What is clear in my head is not in writing. schiste (talk) 09:25, 12 January 2026 (UTC)

Photos/photographers as a growth area?

Chiming in as a relatively new contributor on the organizing side, mainly through WikiPortraits, my main observation is that photography/photographers are a huge wikigrowth opportunity both for contributors and for reader engagement. It is also a very international way for people to have impact across Wikis, and reflective of the visual direction the Internet is editing. And it’s a way for contributors to feel broadly impactful, as some photos are used in upwards of 50 different language wikis.

The number of people who find WikiPortraits through word of mouth and by stumbling across our media mentions in BBC 404media, Die Ziet and Radio France (or Depths of Wikipedia’s Instagram posts) still startles us (and we have to scramble every time we get a surge of signups). The vast majority of people who contribute to WikiPortraits are new to the Wikiverse, and while the BBC article calls them “amateur photographers,” a good number of them are professional photographers who do everything from red carpets to fashion photography to the World Cup and the Olympics. And we end up training a number of them to edit.

WikiPortraits started almost exactly two years ago, and at this point more than 1 out of every 200 pageviews on Wikipedia globally is illustrated with one or more WikiPortraits image. That’s in large part because of entertainment related photos, though our WikiPopetraits images of Leo XIV (form a new contributor) have received hundreds of millions of views. Though I’m not entirely sure how filtering for out bot traffic (as in @Schiste’ citing of the 177 billion views per year) would change that ratio (I *think* it improves it since bots go for long tail content, and we skew high traffic pages).

We have approaching 200 WikiPortraits active contributors (about 1/3 previous existing Wiki username, 2/3 new contributors. Not sure how that ratio fares once you weight by number of photos uploaded. But our new contributors have uploaded at least 20,000 photos and created 508 articles)

What I appreciate also is how international the photography community is (we have WhatsApp groups in English, Spanish and French). There are a lot of active wikiphotographers in Africa, and we got a modest grant to buy them better camera equipment!

But photos have a tough time in the wiki ecosystem, which is striking given how much the Internet is shifting towards visual (Instagram, Tiktok). I remember once struggling to editing our galleries on-wiki (tiny photos and hard to edit manually), and thinking “Oh wow, mediawiki really hates photos.” (This is one of the reasons we applied to WMCH for a grant for Commons Gallery, so we can send portfolios when we apply for credentials).

There was a presentation from a WMF staffer at WCNA (originally a session that was canceled because of the gunman incident, and became a lightning talk?) that showed research that readers found text with photos on wiki much more engaging (if anyone can find that for me in the hackpad, I’d love to connect to her!)

Jenny8lee (talk) 14:07, 10 January 2026 (UTC)

Badly written essay

I'm reading a lot of hasty conclusions, shaky arguments, skewed comparisons, and the omission of important factors. Just about every platform experiences a decline in user numbers after peaking. Far too simplistic. I really feel the urge to go through that text with a red pen and highlight all the flaws. That doesn't mean there isn't plenty of room for improvement (please!), but concrete ideas for doing so aren't being offered. Romaine (talk) 18:06, 10 January 2026 (UTC)

It’s always lovely to have such constructive comments.

Instead of being agressive, please take your «  red pen and highlight all the flaws » and correct me, at least it will contribute to the discussion. schiste (talk) 22:10, 10 January 2026 (UTC)

The text you have written is poorly constructed, and when that criticism is expressed, you call that aggressive. That you call it that way shows that you have no interest in learning where the problems lie in the text. Criticism is in my opinion the cornerstone of Wikipedia and Wikimedia, as that can lead to improve the situation. Would action be welcome? In my opinion yes. At the same time, the past 20 years in Wikimedia have learned me that initiatives that start with a factual groundwork to be missing/inaccurate, always have resulted in failure.

Within Wikimedia there is (too often) the habit to ignore the criticism and also that has resulted in failure and frustrations. The most successful organisations/people in the world have organised criticism around them and enable that criticism to be present in early stages of plans, writings, etc. Hé, but sure, go ahead and frame my comment as "unconstructive", failure then becomes imminent. Romaine (talk) 08:05, 11 January 2026 (UTC)

I love criticism but you have not given a single reason you are disagreeing. Not one argument whatsoever. If I missed point me to it. schiste (talk) 08:40, 11 January 2026 (UTC)

Just some highlights:

• Comparison internet users growth and Wikimedia page views is too simplified. Large platforms that had huge growths have used sneaky algorithms to attract and keep people on their website.
• "Two forms of production" -> Qualitative work (factual, accurate, neutral, etc.) has human input as basis, and I am sure we will find ways to get more machine productivity. But at this moment community members are still working on solving backlogs from 15 years ago and since, because of lacking of good tools/infrastructure to solve those easily. That forms a drain on our community, time and energy that can't be spend on actual content.
• "and we kept optimizing for desktop editors in the Global North" -> Vector 2022 is a failure with many bugs, WMF has not been optimizing for desktop users, but downgrading it for them. At the same time the mobile experience has insufficiently taken care of and is even for the "Global North" horrible.
• "The Commodity: Falling Behind" -> It is an illusion that we can compete with large corporations who use sneaky algorithms to attract and keep people on their platforms. Wikipedia is not in good shape, but making a comparison to them is simply weird.
• "We called it success. It was the beginning of the gap." -> Only people in ivory towers called it a success. Communities have raised issues over and over and hasn't been listened to.
• "2001-2007, 2008-2015, 2016-Now" -> This devide is an extreme simplification and misses substantiation in reality.
• "while the internet nearly doubled in size, Wikimedia's share of global attention was cut in half." -> Skewed comparison and logical fallacy.
• "We went from being essential infrastructure of the web to being one option among many" -> False! For the world Wikipedia serves still as a stable backbone. Also the suggestion that Wikipedia has become only an option, is false too, as this always has been the case.
• Also the 1 to 1 comparison to internet growth with new registrations is simply mindblowing that you use that also as (false!) argumentation. This is not how the world works. Making a comparison of apples with pears makes more sense. If you have every believed that in any way with a growing amount of internet users a similar growth is possible, such is a fairy tale that has nothing to do with reality. Each platform experiences after a growth a decline, and the goal then is to make sure it stabilizes. The expectation of a long-lasting growth is fiction.
• "We didn't beat Britannica by being better at print." -> We weren't competing at all, but were democratizing knowledge, making that knowledge easily available to everyone, and meeting the need by offering knowledge at an accessible level.
• "We won by "understanding" that distribution had fundamentally changed." -> We did not play a game to win or loose. Also the idea that Wikipedia/Wikimedia understood how distribution had changed is a fairy tale.
• "The communities that tried to ban Wikipedia, that installed filters, that punished students for citing it." -> Wikipedia isn't intended to be cited, but intended to be used as starting point for finding information.
• "ChatGPT" (...) "We competed with Britannica over fifteen years. We have maybe two years to figure out our relationship with AI before the window closes." -> If we think that Wikipedia can compete with AI, then Wikipedia already has lost. It is an illusion to think that a non-AI can compete with AI. Should we find a way to relate to AI, sure, but thinking that we can compete is fiction.
• "We need to be in the fight, not standing on the sidelines." -> By fighting, we throw away our core principles, and then definitely we will "loose".
• "The attention economy" -> The attention economy attracts users with algorithms and manipulates them in various ways. Wikipedia does not do that and serves as a stable island on the internet where every day raging storms pass by. Huge improvements can be made in our mobile experience, but let us not make the mistake to throw away our values.
• It is certainly true that organisations that didn't move with their time, have a risk to become extinct. At the same time, in the age of the newspapers log ago, people said that with the arrival of the radio that newspapers would become obsolete and extinct. That did not happen. When tv came people said that radio would become extinct and obsolete, which did not happen. When the internet came a common good, it was said tv would become obsolete, which did not happen. And now AI has come, you are saying Wikipedia will become extinct if we don't compete. Sorry no. With every human evolvement organisations need to find a way to adapt themselves to the new situation, because yes, then (!) if no adaptations are made, their is a risk of extinction. We need to find a way, that suits with us (!), to go with the time and evolve ourselves.
• "Our mental model is: people visit Wikipedia → people donate → people edit → cycle continues." -> I don't know who's model that is, but in 25 years I haven't heard users thinking such. The logic sequence that people donate and then they edit is false. The idea that it is a cycle I do not see either.
• "AI answers → no one visits [Wikipedia]" -> Sure there is a huge decrease, but thinking black-white like this is far from helpful. We live in a world were technology (AI) is making it easier for people to find (online) information, which results in a mental drain. A Dutch professor and leading expert in brain studies phrased it recently as "I call AI declining intelligence". In simpler words: The more humanity uses AI (as it does now), the more stupid humanity become. (And I strongly want to emphasize here that such is not the case for all AI use.) So far I know I haven't seen any solutions to that situation. We have to face the situation that less people are willing to look up information themselves. But that does not mean, that "no one visits" Wikipedia, let's stick to the facts and not exaggerate it.
• "rather than pretending we're still competing for readers" -> We should have never thought that we are competing for readers, it is a false description of the world. We are offering something, we have a purpose that with knowledge, we want to make the world a better place.
• "It's whether we'll shape how our content is used or be shaped by others' decisions." -> We have never shaped how our content is used and we have always (on purpose with our goal in mind) avoided as much as possible to follow other's decisions.
• "In 2018, I argued we should think of ourselves as "Knowledge as a Service"." -> I think we should not. We have almost zero service to offer. What we have to offer is knowledge. If we focus on service instead of knowledge, we move away from our core. Organisations who tried such a thing have a much higher risk of getting extinct and there is almost zero benefit coming from it. In challenging times, the worst thing as organisation can do is moving away from its core, and here you propose such a thing.
• "And yet. English Wikipedia has 6.8 million articles. Hindi, with over 600 million speakers when including second-language users, has 160,000. The ratio is 42:1." -> You compare the lingua franca of the whole world with a language mainly only spoken in one country. Such as an extreme false comparison. Sure, the Wikimedia infrastructure is for many languages and their speakers too limited, but that does not justify such a terrible comparison.
• "We optimized for the editors we had rather than the editors we needed." -> I am sorry, but no, WMF did not optimize MediaWiki for the editors we have.
• "We built the most sophisticated vandalism-fighting tools in the world, and those same tools systematically reject good-faith newcomers, especially those who don't already know the unwritten rules." -> For certainly 20 years the community is asking for a better interaction with newcomers so it becomes clear what is expected from them. We got the visual editor that makes editing easier, but still the Wikipedia infrastructure expects from newcomers that they already know which rules there are, while they don't.
• "Research shows that newcomers from underrepresented groups are reverted faster and given less benefit of the doubt. We've known this for over a decade. We've studied it, published papers about it, created working groups. The trends continued." -> We live in a world where people are biased and prejudiced. What initiatives has WMF taken to make/enable the interactions (to become) more neutral, less biased and less prejudiced?
• "Here's the uncomfortable truth: the Global North built Wikipedia, and the Global North still controls it." -> You present the situation as Group A versus Group B. That is problematic thinking and a false dilemma. About 9 of 10 people live in the northern hemisphere. Almost everywhere in the world where there is a larger and a smaller amount of people, the smaller group is always less represented in people and in knowledge. Quebec is less described than France. Austria is less described than Germany. And many other differences in number of people in society have their influences too. As long as there is in the world a difference in numbers between two "groups", there will be a gap in the representation on Wikipedia. But the whole thinking in groups (like north versus south) is problematic and enlarges the problems, instead of reducing them. Also here the question is what initiatives WMF has taken to reduce the gaps? This should have been much more.
• "The future of the internet is not in Berlin or San Francisco. It's in Lagos, Jakarta, São Paulo, Dhaka. If we're not there, we're nowhere." -> False! The future of the internet is being present everywhere. The future of the internet is also in Berlin and San Francisco, but as well as in any other place in the world. About ten years I proposed the idea of having a more global representation, instead of a few hotspots. This was rejected by WMF.

  It must be noted that the forming of organised groups is largely the result of the enthusiasm and initiatives by local volunteers who had the power to do so. The process of forming organised groups is often depending on just one person who is able to get others in action with enthusiasm, but the reality is also that local circumstances (geographical, cultural, economical, etc.) in countries vary a lot and have huge impact on the possibilities and challenges local people face. But here I must stress that the number one common challenge local groups have is dealing with WMF, getting frustrated by bad decisions, disturbing actions and the lack of support by WMF. And when you have a group, the huge ammount of paperwor for WMF is almost killing volunteers. Also groups need to have a perspective to grow. This seems for the larger chapters possible, for smaller chapters and user groups there are a lot less possibilities and opportunities for already a decade.

• "It means accepting that English Wikipedia's dominance might need to diminish for the movement to survive." -> We should avoid having any dominance.
• "By Wikipedia's 26th birthday, we need to have made fundamental decisions about revenue models, AI integration, knowledge equity, and contributor recruitment." -> This reads to me the same as "We identify a problem. We form a committee. We draft a process. We debate the process. We modify the process. We debate the modifications. Years pass." -> I have seen the coming of Vector, Mediaviewer, VisualEditor, Vector 2022, dark modus, and so on, and so on. it is always presented as something great, fantastic new software, and even today we are still solving the backlogs from 15 years ago, also as result of that, still forming a burden for local communities to deal with. Let me make a prediction: if this essay leads to anything, it likely ends up that the software is changed and that local community members have to clean up the mess that has been designed by some people in the ivory tower. (I hope I am wrong.)

Romaine (talk) 11:24, 11 January 2026 (UTC)

Agree with the point about EN Wikipedia being largest but not being dominant, just look at your chances of getting a grant to Wikimania if you are a regular on EN Wikipedia as opposed to pretty much any other language. Other languages are already long out of EN Wikipedia control if they were ever in it. The Hindi community get to set their rules on notability. German and Arabic communities have different standards as to what parts of the human anatomy can be shown on the mainpage. Each language community gets to decide whether they follow the English route of welcoming multiple spelling and dialect versions, or like Portuguese standardising on one version of their language. Where the WMF could be useful is in commissioning research to see which language model works better, and maybe invest in more multi dialect Wikis on the Chinese model so Portuguese speakers can click on a Brazilian or Portuguese flag icon and change the default view of the Portuguese language Wikipedia from the one that accords with their geolocation to the one they prefer. Similarly with Quebecois speakers on the French Wikipedia. WereSpielChequers (talk) 08:30, 12 January 2026 (UTC)

This original comment, and the tone of the successive message, is unfortunately not as constructive as I would hope it could be. An essay is by its very nature subjective - it is not designed to be an academic article that justifies every assumption. Honestly, I think that Christophe has done an excellent job given what seems to have been the purpose of his piece: bring across his concerns, lay out the primary reasons for it, and start a conversation beyond just celebrating. I'm afraid if it had been convincing to the point that Romaine would be on board with it, many others would have been unable to read and digest it -- you need to strike a balance.

Romaine: You know, or should know, that I love the work that you do, with all my heart. Your thoroughness and eye for detail allow you to perform incredible amounts of labor and reliably so. But please, don't let us get stuck in the past, because of some previous disappointments with how your proposals were appreciated, or some negative side effects.

In fact, I am afraid I would have to warn you: most of the changes that will address the broad concern is decreasing editorship, readership and that will make meaningful improvements to our mission, will likely have some negative side effects to your experience. Not only because our current system is highly optimized to your experience already (even if you disagree - happy to have that conversation separately someday), but also because of simple statistics: even to maintain our status quo, we will need to make many changes to our workflow. The ones that don't hurt your experience, you may not even notice. But you will always notice the ones that do affect your workflow, and you may experience as having to 'clean up the mess'. Thank you for helping clean up things over the years!

I am still not a fan of some of the user interface changes that were introduced some years ago - but I also have to acknowledge that many things have improved without me even realizing it, until I had to explain it to a newcomer.

Regardless: whatever you feel about the content of the essay, I hope you can appreciate the conversation happening, and that we all have a deep concern for our shared mission, and project. Effeietsanders (talk) 10:57, 18 January 2026 (UTC)

@Effeietsanders: This essay of Christophe should be fully rejected as it is damaging the core values of our movement. If an essay is written with doing harm to facts and as well a long list of logic fallacies, if we continue on that, we take decisions on false data. And also it is cherry picking because it is very selective and many important factors involved are left out. It concerns me a lot that in a movement where we try to focus on the facts and unbiased writing, so many hasty conclusions, shaky arguments, and skewed comparisons are presented as evidence. I was hoping that as movement we could avoid such a populistic writing, that we could stick to facts and not comparing apples with pears. That differentiates us from the chaos we see in many places in the world.

 

I agree that it is time to act, but I also think we need to base ourselves on facts, that we need to make sure we make fair comparisons, that we keep our values in sight, and let us be aware where our strengths are. Wikipedia is one of the backbones of the internet and thousands of people rely on our platform every day, directly or indirectly.

 

With the rise of AI we are facing challenges, and we certainly need to find ways to deal with them. But in drop in growth number of users/etc. is part of a common live cycle that platforms have. Endless growth is fiction (unless we want to use tricky algorithms that manipulate people; no thank you). A life cycle typically start slowly, then there is growth, then there is a peak (the hausse), followed by a decline, and then, then we need to make sure the situation becomes/stays stable. From that new stability a growth can be build up again. Please let us not forget, that every single day Wikipedia is still growing in knowledge. Wikipedia is a cumulative platform, each new addition is building on the previous additions. That gives is a lot more stability than platforms where it only counts what happens at that moment.

 

With every large change in society, people are screaming that it is the end near for ... . When the radio was invented, people said that newspapers would become obsolete. When tv came a mass medium, people said the radio would become obsolete. When internet became massively used, people said tv would become obsolete. And now I read that because AI Wikipedia can become obsolete if we do not act.

 

Organisations that in the past have tried to deal with different circumstances in their respective fields can help us learn what we may do and what we absolutely should not do. In general, organisations that loose their values and what they stand for, have an extremely high risk of becoming obsolete and getting extinct. While organisations that stuck to their principles, ideals and values survived turbulent times that have occurred many times in the past centuries.

 

Of course we have to debate the situation we are in and I certainly think we need to work on improvements, but if we loose what we stand for, facts, a balanced review of information, unbiased writing, our neutrality, then we have already lost today.

 

The data analysis in this essay is very week and problematic. If our starting point of a discussion is not based on facts and fair comparisons (also in essays), we face the enormous risk that we loose what we stand for and then failure is likely to become imminent. Romaine (talk) 21:03, 31 January 2026 (UTC)

A French version possibly?

Ce n'est pas que ton anglais ne soit pas clair :-) c'est que le mien étant sans doute moins fluide, j'aimerais si elle existe (ou est sur le point de l'être) une version que je n'aurais pas éventuellement à traduire moi-même pour la citer au risque de trahir (Traduttore, traditore, tout ça) ou simplifier outrageusement. So...? Lechat (talk) 23:03, 10 January 2026 (UTC)

Je peux préparer la page pour traduction si @l'auteur n'y voit pas d'inconvénient, même si je ne serai pas en mesure de la traduire entièrement et correctement en français moi-même. Amir E. Aharoni (talk) 02:52, 11 January 2026 (UTC)

"Younger users don't search. They scroll. They don't read articles. They consume fragments."

The text from the title appears in the essay.

In Wikipedia jargon, this begs for both "citation needed" and "clarification needed".

How are "Younger users" defined? These things could actually be said about all users.

I might by wrong, but my intuition tells me almost all the time, no one reads articles, not now and not twenty years ago. A lot of Wikipedia articles, at least in the larger languages, are dozens of pages long if you print them, and most people don't read the whole thing. I'm not sure how do most people find what they need in articles. I mostly do it using the "Find in page" function that exists in every browser, but I am in a shockingly small miority. Perhaps they just check the infobox or use the section titles ("Early life" to check if their favorite actor is Jewish, "Personal life" to check if they're gay, "Controversies" to check if they expressed a political opinion they dislike, "Filmography" to find their other movies, and so on). Or maybe they do something else, I'm not sure what; if this wasn't researched, it should be.

So basically, everyone probably consumes fragments, although different people may be doing it differently, and some (possibly many, but definitely not all) people think that it's easier to consume fragments elsewhere, for example on LLM conversation simulators or short video platforms, and not on Wikipedia. And this, by itself, is not bad; Wikipedia is supposed to be the infrastructure of knowledge and not the only place where it's consumed.

Same for the claim about not searching. People do search. Perhaps people search differently now from how they searched ten years ago, but people do search. I know that some people don't read beyond Google's LLM-generated summary, but I refuse to believe that everyone does it all the time. Amir E. Aharoni (talk) 03:18, 11 January 2026 (UTC)

Indeed, people were (mostly) also not reading full articles, but at every moment in time looking for pieces of information, as that is exactly what the purpose is of an encyclopedia: to look up fragments of information. Romaine (talk) 08:15, 11 January 2026 (UTC)

You are absolutely right. And thanks for the perspective, the way I meant it, and that's why I used consume, is the fact that people are more and more reducing the path between question and answer. And LLMs are the latest fastest way of even more reducing it.

My intent here was to also root our decline in the fact that habits evolved. Wikipedia came in and made it easier to find information than anywhere in the world. We shortened tremendously the Search effort. I think, and your citation needed is right it's my personal opinion, that we are reaching the point where knowledge gets consumed and not searched.

I am not saying it's wrong or right, I'm just saying that habits have evolved, and where once we were the one embodying ease of access now we're "hard" to access. Not sure it's clearer though. schiste (talk) 09:41, 11 January 2026 (UTC)

Wikipedia is no longer a teenager

Does Wikipedia really need to be in a popularity contest with the rest of the internet? Why not just "build it and they will come"? Caffeinated Chihuahua (talk) 03:31, 11 January 2026 (UTC)

Good point! In my opinion, the world needs a stable resource for factual information, not a platform that floats with every new wind direction. So please no, let's not join the popularity contest with the rest of the internet. Romaine (talk) 08:12, 11 January 2026 (UTC)

It's not about popularity but, to me, about mission and survival.

What use is there for Wikimedia projects if they are not used by anyone?

What use is there for Wikimedia projects if they become mostly static photos of knowledge at one point in time?

I know some volunteers are in it for being part of it, and they do not care about those aspects. But I do care. I engageed in Wikimedia because I genuinely believe the world would be a better place with free and neutral knowledge freely accessible and accessed.

And so in that setting, not being used is against the mission AND a danger (less readers == less editors). schiste (talk) 09:44, 11 January 2026 (UTC)

"We've become the encyclopedia your teacher warned you about, not the movement you want to join"

the encyclopedia your teacher warned you about is a very old idea. It's not a problem. The opposite is the problem: in the big languages, the encyclopedia is so good, that for most people, fixing it feels unnecessary, and when people do want to join, it's too fucking difficult. Visual editor kind of improved this for 2012, but it's outdated for now, and there are way too few serious improvements for the editing process at the moment. To clarify, there are people who work on editing improvements, and they are doing excellent work, but there are way too few of them, and the projects they are told to work on are not even close to being ambitious enough to solve the important issues.

the movement you want to join - that is indeed a problem, but what movement was it ever anyway? Much of the early movement worked on the excitement of the Free Software movement: Linux, Mozilla, Perl, Apache, Slashdot, Creative Commons, etc. Those brands, wonderful as they are, are getting old, new ones aren't being born, Free Software appears to be kind of mainstream (although Embrace, extend, and extinguish is still very much a problem), and there is no new reason for excitement. The only thing I can think of that can generate new excitement is reinvigorating the movement around uncompromising insistence on human and not LLM writing, and if done right, it may actually work. But other ideas are welcome, too. Amir E. Aharoni (talk) 03:41, 11 January 2026 (UTC)

Perhaps I picked the wrong image, but "the encyclopedia your teacher warned you about" is not about the risk (don't use it there are mistakes" but how Teachers were presenting it by comparing it to a static object in a library. It's really about the projection / perception. It wasn't painted as something people could join, embrace, contribute to, built.

That section is really about how we're perceived and I fear that we are perceived as "just" an encyclopedia. And you don't join an encyclopedia.

As for your last part, uncompromising insistence on human and not LLM writing, is to me a complex topic. What do you consider LLM writing? How do you flag it? How do you enforce it?

On that I tend to have the same approach (controversial I agree) as with paid editing. If we enforce all of the writing and citation rules, you don't really care if it comes from a LLM or a paid editor. The issue with paid editing is how can you monitor efficiently articles on all companies, and same goes with LLM.

I would rather deeply invest in tooling, including AI based tools (and we already started that years ago), to make it much much easier for editing communities to monitor everything more efficiently at scale. Fighting LLM writing seems borderline impossible to me, I would double down on what works (people and clear guidelines). schiste (talk) 09:51, 11 January 2026 (UTC)

"If you're an editor: your work powers not just Wikipedia, but an entire ecosystem of AI systems that need verified information"

This is one of the best ideas in the essay, but it's structurally problematic to implement because of the nature of what we currently call "AI", which is actually LLMs.

In Free Software and in wikis, it's easy to see the propagation of your contribution. Some people indeed find it motivating; I certainly do. Wikipedia is both a Free Software project and a wiki, so it works.

The problem with LLMs is that most of them are not Free Software, and even when they are, they are way too hard for most people to debug in practice, and even when it's possible to debug them, their output is not entirely predictable. In addition, the current generation of LLMs requires big retraining for every version with new data, which may take months (this may become faster in the future, but I don't know when). There are advantages and disadvantages to this; in the context of this issue, the disadvantage is that there is no direct line that goes from changing something on Wikipedia to seeing how it fixes an LLM. When I change a piece of data on Wikidata and Wikipedia articles in several languages embed that piece of data, and I can see that they all get updated immediately (of after at most a few minutes because of caching and all that), it's a direct line. With LLMs, as far I know, it's not yet possible.

I don't know how to fix this problem, but if Wikimedia's engineers could solve that, it would not just give a new kind of motivation to edit Wikipedia, but it will probably be the most important innovation in what we understand as "AI". Amir E. Aharoni (talk) 03:53, 11 January 2026 (UTC)

I used AI very purposefully here indeed.

AI is different from LLM which is different from LLM products. The "verified information part" can of course be used in training, but it doesn't really need a live feed, however it could become an amazing tool for LLM based products to fight against "creativity" (I don't like the word hallucination, LLMs don't hallucinate from a technical standpoint).

We could provide the infrastructure to help LLM products strengthen / fact check themselves. There are some technical challenges, but the biggest one is one we would love to solve, it's automating how much a data point is a reliable. To be able for any articles to assess live wether a data point is potential good or not. The "funny" part is that in my opinion to be able to crack that part means increased by a thousandfold editors tools, comfort and ease of work (in my radical mind I would stop minding the wikipedia.org readers altogether and make it only an editing platform, content distribution would be separate).

I'm feeling a bit constrained in text to explain all of that, but it's where I believe we should be going and doubling down. And to echo an earlier point, I wouldn't ban LLM, however I would over focus on improving human editing experience. schiste (talk) 10:01, 11 January 2026 (UTC)

You used "AI" purposefully, but for what actual purpose?

"AI" is indeed different from LLM, but in a different and much simpler way. LLM is a name for a fairly well-defined family of technologies (are they good or bad is a separate question). "AI" is a name for nothing at all; it's a word whose meaning changes every few years. Currently, it mostly refers to LLMs because that's what the general public sees. In ten years, it will probably refer to something else. That's why I purposefully use "LLMs" and not "AI": because LLMs are an actual thing, and we no one knows what will "AI" mean in ten years. Amir E. Aharoni (talk) 12:23, 11 January 2026 (UTC)

Is it written by an LLM?

I've reached the middle of the essay, and I'm starting to suspect that an LLM was used to write at least a part of it. I'm not saying that it was all written by an LLM; obvsiouly, human thought and analytics were invested in it. But a significant part of the language style makes me think that it was phrased with the help of an LLM.

So I'd love to know how much of it was generated by an LLM. Perhaps it's zero, in which case I apologize. But I hope that it's an understandable question in this age. Amir E. Aharoni (talk) 03:59, 11 January 2026 (UTC)

I did use Claude Opus 4.5 to help write and copyedit it. Data extraction, analytics, other sources, and the core reasoning are mine (some people have already read me rant about some of that stuff in the past).

But structuring my thoughts and binding them properly in English are not my forte. LLMs are great tools to help do that.

I did rewrite part of it and almost everything you identified as odd was what I tweaked ^^ schiste (talk) 07:46, 11 January 2026 (UTC)

LLMs are occasionally great tools to copyedit things, that is, check spelling and sentence structure. And even then you should take their suggestions carefully.

They are not great tools to write things. I dragged myself through the whole essay, but reading typical LLMese was not fun.

It looks like you really invested true effort in good data analysis, and that you are trying to say something important in this essay, so say it yourself. If it's easier for you to write in French, do it, and I'll happily translate it into English. Amir E. Aharoni (talk) 12:29, 11 January 2026 (UTC)

I'm glad the data, analysis, and reasoning were yours, but it was very clear that an LLM wrote the essay. And that made me trust it less, in an ironic distillation of some of the issues you wanted to raise. I agree with Amir. Ganesha811 (talk) 13:49, 15 January 2026 (UTC)

i agree with this. a much better idea would have been to ask another writer to collaborate with you. the fact that that’s not your first instinct in a collaborative writing and editing community suggests something about the likely destructive durability of AI in reshaping our expectations around communication. 100W bulb (talk) 18:40, 15 January 2026 (UTC)

My thoughts exactly. I came to the talk page to see if anyone else had complained. Ignacio Rodríguez (talk) 20:40, 16 January 2026 (UTC)

"The billions who came online found an encyclopedia that didn't speak their languages"

This is extremely unpleasant to write, and this should be extremely unpleasant to read, but the sad part about this is that the vast majority of these people weren't loking for an encyclopedia that speaks their languages.

Most people who are coming to Wikipedia from Congo, Indonesia, India, Ghana, etc., are used to reading and writing in French, Indonesian, and English, rather than Lingala, Balinese, Hindi, or Twi. These countries have few (or no) good schools that work in these languages, few or no textbooks, and tiny and not-so-diverse publishing industries.

For the people from these countries who do know a language in which there is a big and successful Wikipedia, like English, French, or Indonesian, the fact that there are too few articles in Lingala, Balinese, Hindi, or Twi Wikipedias is usually not an issue. Most of them don't even expect such a Wikipedia to exist.

It is, of course, a problem for those people who only know Lingala, Balinese, Hindi, or Twi, but:

1. A lot of them are illiterate, so they wouldn't be able to read a Wikipedia in this language today even if it did exist. I don't have the precise numbers, but every time I tried asking people from these countries about this, they said that if someone is literate in Hindi, Twi, etc., they are also likely to be able to read in English or French, so they usually don't need a Wikipedia in the local language.
2. The cultures in which they live teach them to think that these languages are for speaking at home and not for acquiring serious education.

This is a huge cultural, political, and economic problem that had existed long before Wikipedia, and that isn't yet resolved. It should be resolved, and Wikipedia should be a part of it.

There is a tiny number of people who are breaking this mold and who write Wikipedia in those languages because they love their languages and want to develop them even though they could get most of the information they need in a big language like English, French, etc. These people are heros, of course, and those are the people who are best suited to be in the center of solving this problem. Are they empowered as heros should be? No, they aren't. As far as I can tell, the WMF only gives them grants for events, which are fun, but not hugely impactful, and the WMF doesn't make any significant product improvements for which they have been begging for many, many years. Amir E. Aharoni (talk) 04:20, 11 January 2026 (UTC)

Point taken, and knowing your expertise on those topics I definitely listen. For context, I had in mind India's languages in mind, but your point still stands.

Which makes your last paragraph even more unpleasant to read actually. But all the more important. Thank you schiste (talk) 10:06, 11 January 2026 (UTC)

Weird paragraph

This whole paragraph is weird:

But there's a crucial difference. Wikipedia succeeded by being native to the internet, not by ignoring it. We didn't beat Britannica by being better at print. We won by "understanding" that distribution had fundamentally changed. The communities that tried to ban Wikipedia, that installed filters, that punished students for citing it. They wasted a decade they could have spent adapting.

It's hard to understand it.

Some issues that I see:

• Why is "understanding" both italic and in double quotes? I think that it needs neither.
• The sentences about the communities that tried to do those three things is kind of unfinished. Perhaps it should be somehow merged with the next sentence.

Amir E. Aharoni (talk) 04:30, 11 January 2026 (UTC)

"We"

There's this paragraph:

We're not proposing to replace donation revenue. We're proposing to supplement it. Right now, 100% of our sustainability depends on people visiting our site and seeing donation banners. That's fragile. If entities using our content at scale contributed to sustainability, we'd be more resilient, not replacing individual donors, but diversifying beyond them.

If I understand correctly, this paragraph uses "we" to refer to the document's author. But a few paragraphs earlier, it says:

Let me be precise about what this means, because I know some will hear "toll booth" and recoil.

So is it "me" or "we"? It should be consistent. Amir E. Aharoni (talk) 04:38, 11 January 2026 (UTC)

One of the thing I have rewritten and didn't copyedit with LLM, I was writing how I think. We is "me plus the people I hope would agree my views and push for that change together after they have read my essay because it can't be just me". So yes we is me. schiste (talk) 10:08, 11 January 2026 (UTC)

Thanks. Can you please update it for consistency? Amir E. Aharoni (talk) 12:32, 11 January 2026 (UTC)

"Wikipedia-verified"

There's this paragraph:

What could we offer, concretely? Pre-processed training data, cleaner and cheaper than what AI companies scrape and process themselves. Confidence ratings based on our 25 years of edit history, which facts are stable versus contested, which claims have been challenged and survived scrutiny. A live verification layer that embeds Wikipedia as ground truth inside generated answers. A hybrid multimodal multilingual vectorized dataset spanning Wikipedia, Commons, Wikisource, and Wikidata. And the "Wikipedia-verified" trust mark that AI products could display to signal quality.

The thinking is not bad, but I don't quite agree with the proposed form and framing for two reasons:

1. No one should want a "Wikipedia-verified" trust mark. More precisely, no one should want it to become reified and formalized. That will inevitably create ways to game the system. The public's trust in Wikipedia is a great thing, but it must be implicit and not official. It's good that Wikipedia is informal and that its editors only summarize information and make it accessible, fact-checking it along the way. The work of creating information should be left to others, mostly universities and news organizations.
2. This framing takes the editors for granted too much. The editors' motivation to do this fact-checking is that they want a good encyclopedia. The proposition of doing it directly for free for other organizations, most of which are probably insanely rich, and within some poorly defined ecosystem doesn't sound motivating. These are human volunteers, not enterprise APIs.

A better way to harness this power is to improve Wikipedia itself by making it more convenient to edit, more structured for humans, and more machine-readable. Wikidata kind of started it, but it requires massive fixes, much bigger than what the current Wikidata strategic plan proposes. But long before Wikidata, templates were Wikipedia's structured data, and whether you like or not, they still are Wikipedia's main way to store and edit structured data, even more than Wikidata is. But templates have a lot of warts that must be addressed. If we do that, the editors will keep seeing their work result in a better encyclopedia, and other consumers will be able to better use their power for fact-checking, training, and what not. Amir E. Aharoni (talk) 05:00, 11 January 2026 (UTC)

I 100% agree with your last paragraph, it is core to be able to offer a "confidence rating" to have much much better editors tools, easier to edit, more engaging, etc. I would love to talk more about that, but I genuinely think that the editing machine we built (rules, guidelines, etc) is an amazing asset and accomplishment. BUT we now need to make it easy and fun to fully leverage it. And the "irony" is that reinforcing also is the way of building a new product for the rest of the world. schiste (talk) 10:11, 11 January 2026 (UTC)

"Our investment in people"

There's a lot to unpack here:

This might mean funding local events that bring new people into the movement. It might mean rethinking what counts as contribution. It might mean, and I know this is controversial, considering whether some kinds of work should be compensated.

The WMF and the chapters are already funding local events. It is good for PR, but not very good for bringing people into the movement. A good product that is good at inviting people to contribute directly is far more effective. Unfortunately, we are not good at understanding the usability of contribution, so we are even worse at improving it. Events, banners, and social media campaigns won't have a lasting impact without improving the discoverability and the usability of contribution. But if the discoverability and the usability of contribution is improved, then banners and social media campaigns probably won't be needed. Events will always be needed, but not so much for recruitment as for socializing.

And what exactly does "rethinking what counts as contribution" mean? Any specific propositions?

Finally, "I know this is controversial, considering whether some kinds of work should be compensated": Wikipedians very often say that other Wikipedians will probably dislike it, but they don't often say that they dislike it themselves. Yes, some kinds of work should be compensated. It must be done carefully and thoughtfully, but we should really stop starting every sentence about this from saying that this is controversial, and we should move to the stage of thinking what should be compensated and how. Amir E. Aharoni (talk) 05:10, 11 January 2026 (UTC)

"The WMF and the chapters are already funding local events." I don't think they do that any more. Before COVID they would provide coffee and water, but the few local meetups left now meet by zoom, which does not have any expense. Caffeinated Chihuahua (talk) 07:53, 11 January 2026 (UTC)

That may be true in your region and other places, but is not true universally; I know this because WMKR continues to provide offline meetup support, of which the most recent one happened last weekend. — regards, Revi 08:56, 11 January 2026 (UTC)

Our local meetups and conferences have been funded by outside sponsors, not the WMF. Unfortunately the current political situation no longer encourages editing events centering underrepresented groups like Black History Month and Women's History month, which used to be the busiest time of the year, since anything regarded as "DEI" is being removed from government websites. Also, if you look at the recent incident at a New York event, it was local volunteers, not WMF Safety & Security employees who jumped on the stage and disarmed the situation. There is also a lack of support for anonymous persons. We have tried to get some small grants through our local chapter that would help us contribute more, but it is denied. Caffeinated Chihuahua (talk) 09:41, 11 January 2026 (UTC)

I start by the end, you are right, I shouldn't have said it was controversial, this is a bad habit.

Events will always be needed, but not so much for recruitment as for socializing. > Socializing is recruiting. I once said that the best campaign we could do is to fund and organize all the pubs Sunday trivia quizzes, and even to start them around tea time. That would be a very efficient way of socializing.

To your whole point, in my Essay I say it's not a OR problem but a AND problem. We can't choose.

We need to improve the editing tools and experience.

And, we need to embrace AI and LLMs.

And, we need to double down on our rules.

And, we need to put Wikidata and Abstract front and center.

And,...

And yes there has to be an order, etc... But our future doesn't lies in improving editors experience OR events. It's AND. What I believe also is that our future does not lie in improving readers experience on wikipedia.org.

"rethinking what counts as contribution" > This is a bit personal, because I heavily edited Wikimedia projects for a few years and then focused on organizing. And this is to this day thrown at me not being a contributor. I wish we'd be a much more open movement and celebrate any bits of people contributing, giving support and not grading it. I admit this is probably wishful thinking, but maybe not. Maybe it contributes not being a very welcoming movement? I don't know I'm rambling. schiste (talk) 10:20, 11 January 2026 (UTC)

Yesterday I was at the London Meetup, I've been a regular there for more than 17 years. Of nearly 200 monthly meetups in that time I have probably been to over 100. When I first went I was probably the second oldest person present. Yesterday I was probably closer to the average, despite allegedly being 17 years older than I was in 2008. The event is self funding (enough of us eat or drink that the pub lets us take over our corner of it) and these days we have little or no connection to WMUK. I've also been involved in a lot of editathons. Where I think we could and should shift outreach is from trying to recruit newbies back into the editor retention tool that editathons were originally. I was at the original British Museum editathon in 2010. In those days an editathon was an event at a museum or similar where a bunch of editors got an invitation to visit a museum, talk to curators, skill share with other Wikipedians and improve some articles. For me that was the event that turned my newish internet hobby into something that got me behind the scenes at a museum. I think we should shift the focus of outreach back to skillshare events and welcoming newish volunteers into the community as opposed to trying to recruit completely new editors. Turning your internet hobby into something that puts you into touch with experts can be a lifechanging event. Of the ten Wikipedians who attended that British Museum event in 2010 several of us are still active, three of us were at yesterday's London meetup. WereSpielChequers (talk) 09:08, 12 January 2026 (UTC)

@WereSpielChequers I remember in 2010, the median age of Wikipedians was 22, and the most common age was 18. A common gambit of Wikipedia critics was to say, Who trusts an encyclopedia written by teenagers?

Yesterday I was saying over in de:WP that somehow, over the course of fifteen years, the average age of Wikipedians seems to have increased by thirty or forty years. Strange maths. Andreas JN466 12:50, 19 January 2026 (UTC)

Yes, and of course only part of that is that those of us who are still around are fifteen years are fifteen years older. My theory is that there is a combination of things, fifteen to twenty years ago our vandalfighters skewed young, in some cases very young. I suspect the writing part of the community skewed somewhat older than the vandalfighting part, especially after 2007 or so when we started being stricter about requiring inline cites to reliable sources. We've automated away the vast majority of that vandalfighting, and the young today are mostly smartphone users - for that generation we are more of a broadcast medium than an interactive one. At the same time I'm fairly sure we have disproportionately been recruiting retirees and empty nesters, which makes our volunteer community look more like what I would expect a volunteer community to look like, especially for a sedentary pursuit. 15 years ago, at least where I am, the 75 to 95 age group included lots of people who weren't PC literate. Those people have mostly now passed, and most currently living retirees who had white collar jobs have PCs at home and are more than capable of editing Wikipedia. One of my suggestions in the 2009 strategy process was to survey former members of the community. Sadly that only happened once, but if the WMF ever wanted to understand and conserve the volunteer community, surveying current and former members is a very basic step. I predict one thing we'd find is that our newer members include a larger proportion of retirees than the people who have been here for twenty years; With many of our teenage admins from twenty years ago now in their mid 30s and not having much time for a hobby. WereSpielChequers (talk) 15:48, 19 January 2026 (UTC)

@WereSpielChequers You pretty much nailed it, I think.

Is it your sense that middle-aged people coming in will be sufficient to maintain volunteer numbers as our generation ages and fades away? Andreas JN466 18:21, 19 January 2026 (UTC)

Well they certainly could be. Editing an encyclopaedia as a hobby is never going to be everyone's cup of tea. But there are tens of mullions of retired and semi retired people out there, in my country there are probably over a million gardeners and hill walkers who could be tempted into a hobby that they could do on rainy days when gardening is a bit miserable. If the National Trust can get over 5 million paying members and the University of the 3rd Age around 400,000, both just in the UK, then I'm pretty sure our mutual hobby could get the rather smaller numbers we need, especially if the rest of the developed world is taken into account. There are some barriers in place that need fixing: As Christophe points out, if the tech giants are allowed to hide Wikipedia content behind AI mirrors as if the BY part of CC-BY-SA didn't apply to tech giants, then ultimately how do we go on recruiting new editors? There's also the issue of mobile v PC editing. I suspect we have decades rather than a year or two before this would kill us, and in that time I believe that either the smartphone will have been superseded by something with a better editing interface, our mobile platform will have become editor friendly, or people who have used smartphones from the age of 5 to 55 will have come up with work arounds that let them edit. My own preference would be for a mobile platform that automatically added more editing features as people tried to edit; And my big regret re our lack of mobile editors is the national and generational skews that this gives the project. We still have huge numbers of people doing the odd handful of edits, as Lodjewick and others have pointed out the problem is that we are less welcoming as a community than we once were. I think that our biggest mistake in outreach was to drop the skillshare/backstage pass model of the early GLAM events and repurpose editathons into events to recruit completely new editors. What made the British Museum editathon of 2010 such a life changing event for me was that it turned my fairly new internet based hobby into one that welcomed me into museums and other places to work alongside curators and other museum staff. We could resume doing such events, I look at the London based geonotices whenever I check my watchlist, it feels like years since I have seen an invitation to an editathon, a photoshoot or a backstage pass. A newish editor who discovered their watchlist in the last five years might not yet have had such an invite.

In the long run we might find that AI moves on from its current Large Language Model of predicting the next word to something that can summarise the various reliable sources out there, including the paywalled ones and the long deleted ones. But I think we have more than a couple of years before then. We've been on the verge of having self driving cars for years, and it may well be a year or two away for many years to come. Especially as every year, and every time a reliable source goes dark, more of our content is cited to sources that are no longer available to us or an AI. WereSpielChequers (talk) 22:48, 19 January 2026 (UTC)

"accepting that English Wikipedia's dominance might need to diminish for the movement to survive"

It's a nice slogan, but let's define what "English Wikipedia's dominance" means exactly.

Its sourcing policy makes contributions about some cultures hard? No, it doesn't. Its sourcing policy is good. Giving up on it will contradict what the essay says a few paragraphs above, about Wikipedians' great fact-checking ability.

What actually happens is that incorrect application of this policy makes contributions in the English language about some cultures hard. Correct application of this policy may still feel annoying, but it prevents bad contributions.

This can be addressed in several ways:

• Identifying and preventing bad applications of this policy, such as:
  • Blocking users only because they "appear to use LLMs" to write articles or reply on talk pages, and not because of the content of the contribution.
  • Disregarding a source only because it's a Nigerian newspapers (and not because it's a bad newspaper).
• Encouraging people to contribute in their languages and not to the English (or French) Wikipedia. Yes, it's an opening for possible lower-quality contribution in other languages, but if it works, the community will grow, and sooner or later, it will come to creating a sourcing policy.

You know, it sounds righteous to say "English Wikipedia shouldn't be dominant, because what's good for the English Wikipedia is not necessarily good for other languages". Yeah, not necessarily. But a lot of things that are good for the English Wikipedia actually are also good for the other languages. A lot of editors in other languages wish they could use them, and for all kinds of technical and economic reasons they cannot. Very often, what those editors want is for stuff in the Wikipedia in their language to work like it works in the English Wikipedia. This incluedes templates, gadgets, policies, and other things.

So the English Wikipedia is not the problem; the disconnect between the English Wikipedia and the other languages is. Amir E. Aharoni (talk) 05:30, 11 January 2026 (UTC)

The English Wikipedia is not the problem; the disconnect between the English Wikipedia and the other languages i (languages and projects). Your sentence is much much better than mine. schiste (talk) 10:21, 11 January 2026 (UTC)

By most metrics the English language Wikipedia is our largest project. As a community we have a plethora of tools reports and templates. Some of the people who made them will happily replicate them on other projects if you ask them. Others have left long ago and may or may not have left some source code. But communities who don't yet have their own bot operators can still file requests at EN:Wikipedia:Bot_requests, ideally after getting consensus for that bot on their own wiki. Some other languages have passionate communities whose main motivation is in preserving their mother tongue and making material in it available on the internet. Some of our volunteers on English Wikipedia and I suspect many other languages have a secondary motivation that editing Wikipedia in a language they use or want to use professionally improves their skills in that language. We should value others for their otherness - as a community we can welcome and receive great contributions from both the person who is passionate that their native language and the person whose CV includes that their written English is good enough that they have written featured articles on Wikipedia. As well as the teenager applying for their first paid job whose references include one from their national chapter of Wikimedia re their role as a volunteer. WereSpielChequers (talk) 09:35, 12 January 2026 (UTC)

I agree with it all, but there is still a disconnect: The number of steps that someone has to make to start using hieroglyphics, math, musical score, or Commons images on any Wikimedia wiki in any language is zero (I'm not counting localization, which is very easy and uniform for all of the above); the number of steps that someone has to make to start using HotCat, Infobox person, or Cite web on their wiki is non-zero. Each of the things in the non-zero-steps group has to be separately requested, manually imported, localized (in a way that is much more difficult than doing it for the things in the zero-steps group), and after all that is done, it also has to be manually updated. The same goes for setting up page deletion procedures, administrator appointment procedures, etc. It's not the English Wikipedia's fault, and it can be fixed, but it isn't being fixed despite countless requests. Amir E. Aharoni (talk) 15:26, 12 January 2026 (UTC)

Thanks Amir, I can see how this is an overhead. But I would differentiate between functional buttons and the procedures for working out when to use them. Making the buttons work requires some IT skills and some advanced permissions, deciding who gets those buttons and when they can use them is more a matter of understanding how our sort of community can work, and I'm guessing on most new wikis in our movement that has come from multilingual people who were already active in other languages. Since English was the first Wikipedia version and I suspect is the most common second language, there is a good chance that a lot of other wikis used EN as the model to copy policies from. That's probably unhealthy because what works for a community so large that whenever you go to a community noticeboard you are mostly dealing with people you barely know, is not designed for a community closer to the magic number of 150 where everyone knows each other.

To go back to the wider point of this essay, what I think we should expect to see in a lot of our projects is communities stabilising and becoming more efficient over time as they get to learn how to do things more effectively and also to trust each other. With enough money to at least keep the projects running, and no serious competitor, we are as secure as any internet project could be. Current AI is just a new type of mirror site, using LLMs to rephrase our work, but not to replace it. We remain the foundation, even if that foundation may become hidden by other structures. But our core mission, to make the sum of the world's knowledge freely available to all we are still doing that even if in five years 95% of our users are accessing that knowledge via some sort of AI enabled mirror. I'd like us to do much better than that, but even the worst case scenario still has us as one of the sites that makes the internet not suck. WereSpielChequers (talk) 09:50, 16 January 2026 (UTC)

i stopped reading after part 1

because it was clear this "essay" was generated by an llm. ltbdl (talk) 05:39, 11 January 2026 (UTC)

As I answered above "I did use Claude Opus 4.5 to help write and copyedit it. Data extraction, analytics, other sources, and the core reasoning are mine (some people have already read me rant about some of that stuff in the past).

But structuring my thoughts and binding them properly in English are not my forte. LLMs are great tools to help do that.

I did rewrite part of it and almost everything you identified as odd was what I tweaked ^^" schiste (talk) 10:22, 11 January 2026 (UTC)

@schiste: It's insulting to read your AI-generated blog post ltbdl (talk) 03:31, 12 January 2026 (UTC)

@ltbdl Genetic fallacy schiste (talk) 08:54, 12 January 2026 (UTC)

Fundraising and page views

In your "Why page views matter", rather than seeing WMF fundraising as an exercise which may suffer as a consequence of lower pageviews, should the fundraising exercises be seen as a cause of lower pageviews?

I commented previously in 2022 and 2023 on the detriment to a casual user's experience. For a significant proportion of each year, seeking information from Wikipedia instead gives a begging-bowl splash screen that the casual user may or may not negotiate past.

This is akin to, for example, my own acquired knowledge of various worthy US news sites: I know them to publish good articles on the evolving situation there, but I also know I may see little beyond the first paragraph before a subscription appeal kicks in. So I don't even go to these sites anymore.

Most of us here are logged-in (and laptop-based) users, so may not see that others find reading Wikipedia to be a poor user experience because of the periodic fundraising impediment. An example:

• my 19 December 2022 comment was triggered by gradual recognition of a film's plot. So my partner looked up "Force Majeure" using a phone browser - and got hit by the fundraising splashscreen.
• Now go to Google and type "Has the film Force Majeure been remade?" or "Is the film Downhill an English language remake?" The AI Overview immediately delivers the sought answer. For the second query, it also shows a link to the Wikipedia article, which is good but the question has already been answered.

Yes, the AI gang have gorged on Wikipedia to serve up their answers, and they are all probably in a financial bubble, but as it stands Wikipedia looks less reliable and relevant by not immediately providing its information to the casual user. Which surely impacts user sign-up and engagement (your "Why page views matter" points 2 and 3). AllyD (talk) 12:03, 11 January 2026 (UTC)

Data source for the graphics? + some feedback

Is the data graphic "Edits per New Registered User" for all Wikimedia projects? If so, that includes Wikidata where normal changes are usually split up into many changes and which is heavily edited via semi-automated tools and cat-a-lot semi-automated categorizations on Commons etc. That doesn't have much meaning and may indicate more prevalent use of such tools. Could you please cite the precise data source for the three data graphics somewhere next to the graphic where one can see it? Also note that one ref there is a broken link: Wikimedia Statistics does not exist and Statistics is a list of links where it's unclear what is meant. @Schiste: could you please at least fix that; it's a bit surprising nobody seems to look for the datasources of the charts.

I have only glanced over the essay and think it's right on multiple important things but also flawed in many cases or making false conclusions. One of the biggest problems I have with it is that it doesn't really touch on the software development side. The development is severely neglected and when you for example write Mobile has fundamentally changed how people consume information. Our data shows the shift: mobile devices went from 62% of our traffic in 2016 to 74% in 2025 (something very important), you do not mention that categories are hidden from users on mobile. Categories are especially useful on Commons.

Another flaw is thinking high level of donations are very important to the success of Wikimedia projects. However, only a small fraction is needed to keep the servers running and again, the rest isn't used much for development of any of the countless severe bugs and important issues collecting dust for a decade. In my view, it currently doesn't matter much if donations go down substantially because Wikipedia is written by volunteers and WMF largely wastes the money or spends it on things that the projects could flourish nearly as well without or basically saves it up. Wikimedia projects aren't nearly as dependent on the donations as it made to look like here. And re And being there means more than translating English articles I think you got it wrong – it means translating Wikipedia articles (mainly English) or having them be translated since that is a readily-usable largely-untapped resource and then it's not just that but also more in regards to supporting projects, orgs, and people from these regions. But that's not done. The low-hanging fruit remains hanging. Billions of people could find millions of high-quality articles when searching the Web in their own language that they couldn't 2 years ago. They don't. That's a failure.

Lastly, in regards to the need for volunteer editors – what is needed is not complaints, or "investment in people" or 'making this a priority' etc and so forth but real-world effective action in terms of things that have impact (which usually involves software changes which gets back to the earlier point) – examples of this are making editing more fun and engaging via more feedback in the form of badges/achievements for example, or ways to make low-active users find the tasks they find interesting taking them off the shoulders from overburdened active editors, or competitions with motivating leaderboards, or some tool that helps newcomers with their editing and an easy interactive introduction to Wikipedia editing in the form of a campaign etc. Younger users don't search. They scroll. They don't read articles. They consume fragments. They scroll when not looking for something specific (but the entertainment app scrolling is not necessarily the time we can readily substitute with Wikipedia browsing except for low-hanging fruits like this+similar) and people of all ages don't read articles but consume fragments (mostly) – I think there was some data that showed that this is the most typical Wikipedia use. Note that many people prefer audio for long texts instead of reading them in entirety on their screen and many like having illustrating visuals, easy-to-understand clear short summaries as article+section intros, and lots of subheaders, all of which keep the text more interesting and easier to read. Prototyperspective (talk) 00:03, 12 January 2026 (UTC)

Distorted graphs

One of the downsides of starting a graph at 80% instead of zero is that a 9% fall looks like a 45% fall, the increase is also visually exaggerated. WereSpielChequers (talk) 09:43, 12 January 2026 (UTC)

Our use of AI

I've worked on AI created lists of possible typos. I have volunteered a lot of my time to resolve easily confused words Calvary/Cavalry discus/discuss Minster/Minister Manger/Manager and the like. I would be happy if someone uses my contributions to train an AI to do that for others. I do see the WMF as having a role in reminding AI outfits and other reusers of Wikimedia Commons, Wikipedia and our other sites of the difference between CC0 and CC-BY-SA and the implications of that on attribution and licensing. Where we currently have the edge over AI is that our content is or should be a summary of the cited source, not a statistical prediction of the next word. When AI learns how to read and reuse sources we will still have the advantage of 25 years of contributions written cited to sources long lost. We also have the advantage that some members of our community have access to paywalled sites that may not be available to AI. The WMF and the chapters have done useful work there with the Wiki library and other initiatives, but maybe that could be expanded, especially in countries/communities where library access is limited and being able to get the WMF to buy some reference books for you is a big morale boost (and maybe more useful than a free Tshirt). WereSpielChequers (talk) 10:08, 12 January 2026 (UTC)

Donors and fundraising

Yes there is a link between new donors and fundraising banners. But if we sign donors up on direct debits, then depending on the banking system we have a revenue stream that persists for decades with a gentle decline. I don't know how donor retention varies from one banking system to another - my own experience is very UK focussed. But one of the charities that appears on my bank statements has been there for over thirty years. If we continue to be the best available free online encyclopaedia and the only dictionary that exists in some languages then our donations from most existing donors will continue to flow, and if donations ever drop below the level needed to fund the project then we have an endowment that is already big enough that it could take us through the sort of fundraising crisis that would kill most non profits. This essay treats our donations as the equivalent of an annual passing the hat around to collect donations. It isn't, as many of those small donations are small annual donations from people who are likely to be with us for a long time. I'd be more concerned that our endowment and much of our global fundraising is concentrated in a country with an unpredictable regime that has values that are antithetical to ours. Now would be a good time to decentralise our fundraising and have money raised in the UK and other sizable countries handled in those countries rather than go through one global organisation. Move as many countries as possible to the Wikimedia Deutschland Model rather than the Wikimedia UK model. WereSpielChequers (talk) 12:02, 12 January 2026 (UTC)

We need better metrics

Edit count is not an ideal metric, and I'm not sure if it is getting worse or better. We have some people contributing >100 edits in an editing session of an hour, and others maybe 1. If I'm categorising or typo fixing I do far more edits per hour than if I'm writing referenced content. One solution to that would be to measure "hours volunteered". Very crudely, however many edits you save in one calendar hour you probably only donated one hour of your time. If you saved an edit that involved several paragraphs of text with multiple references then that one hour may have represented a lot more than one hour of work, so "approximate hours volunteered" would still be a crude measure of the size of our volunteer community. But I think a better measure than we have now, and to understand our problem better I think we need better measurements. WereSpielChequers (talk) 17:01, 12 January 2026 (UTC)

Agree. And that's an important subject, for example because it can be motivating and make editing more engaging. Take a look at https://github.com/geohci/edit-types/issues/73 and c:Category:Wikimedia contributions.

(Approximate) hours volunteered sounds like a good idea but how would one measure that? Also it will probably make a lot of people feel like they're wasting their lifetime on the projects and that without any meaningful positive feedback if any so it's not really motivating and be overall disadvantageous to the projects and contributors. The links in the two links above have some ideas. Another clue is the authorship percentage displayed in Xtools where e.g. being a top x editor of an (/y) article(s) is a clue – it also depends what the metrics are for: e.g. personal statistics for motivation & feedback etc or for tracking the activity of a project over time (as in does it go up or down) or sth else/a mix? Probably for meaningful statistics one could have a catch-it-all metric but needs to differentiate, e.g. non-reference content added to mainspace (with 10 days delay) separate to number of articles created (including redirects turned to articles) separate to other things. Also see W463: Volunteer software development recognition badges – for example volunteering on Wikipedia is not recognized as volunteering (which could bring more contributors or other advantages) similar to open source software development where there's a petition to change this. Lastly, again design and/or use of metrics is dependent what those metrics are meant for – but if done well one could use metrics used also or originally for curious personal statistics in aggregated form for tracking the health and large-order activity trends of the communities. Prototyperspective (talk) 19:12, 12 January 2026 (UTC)

(Approximate) hours volunteered sounds like a good idea but how would one measure that?
Wikiscan already does that. Choose a project, then go Tables - Time. AtUkr (talk) 15:04, 14 January 2026 (UTC)

I think the hour mentioned in Wikiscan is the total edits in each hour of the 24 hour day. That's interesting information and I assume will look very different for a language like Georgian or Thai where most editors will be in one time zone and anyone editing at 4AM is likely to be abroad. English and I expect arabic would be spread over many more timezones. What I suggest is somewhat different, how many hours are being volunteered, and the approximate calculation is to only count one edit per editor who edited in that hour. So someone who does a lot of tool assisted editing and did a thousand edits in one three hour stint? They probably spent three hours on Wikipedia that week. The chap who saved six edits, but did them in four seperate sessions, that would count as four hours. The reason I thought of this is a conversation I once had with an editor who'd been here for years and whose edits were almost all well crafted referenced paragraphs of text that may on average have represented an hour of work. My edit count was probably twenty times his, but he had probably donated more hours to the project. The reason why I think this matters is that edit count is a very crude metric, and it may give us a false sense of our community size if what is happening is that our content contributors are leaving and our established editors are getting better tools that enable them to do more minor edits in an hour. Conversely, we may find that the easy stuff is being automated away and the hours of time is actually going up. I predict that if we remeasured the 2007-2012 era of the English Wikipedia this way, we'd find that as we upped our sourcing standards, and people like me had to start citing our sources, so the hours contributed per million edits likely went up. WereSpielChequers (talk) 23:05, 14 January 2026 (UTC)

I don't know which Wikiscan page you're looking at but probably it's not the one AtUkr meant or you misinterpreted it. I think pages like this are meant and it does show estimated total time volunteered under column "Total" or e.g. "File". One could create a line chart of the annual values. However, what I meant to reply but decided earlier not to is that wikistats estimated time spent per user (in the user list) look like wild estimations and I could not find any page that explains how that estimated time is calculated and doubt it's done well. The reason why I think this matters is that edit count is a very crude metric, and it may give us a false sense… agree …of our community size I think it's more about community activity than size or at least equally about these 2-3 things. Also I think time volunteered may be slightly better than edit counts and be good to complement these but these would be / are similarly crude and not saying much about how much a contributor to a project and which types. I again recommend the two links above for some further ideas and inspirations such as this – one could slowly build stats per various types of contributions that actually have meaning and give some good insight. One would then have separate charts instead of just one showing a crude metric and that would not be a problem but be good as one can get more knowledge/meaning out of the statistics than any broad metric could show. E.g. things like 'we are getting less and less people participating in categories for discussions' or 'there's more articles being created now daily by new users than before' or 'the number of semi-automated categorization edits is going up while the number of contributors who have categorized an article (1 or more per month) is going down' or whatever. Basically nuance and differentiation is good and also what would the conclusion be if some broad potentially-flawed catch-all metric like 'overall time volunteered' would go down? There's not much insight or actionable steps to have there. Prototyperspective (talk) 00:00, 15 January 2026 (UTC)

wikistats estimated time spent per user (in the user list) look like wild estimations and I could not find any page that explains how that estimated time is calculated and doubt it's done well.

Try hovering your cursor over table headings, sometimes (not always) there's an explanation of how a metric is calculated. For time, it's this: 'Cumulated time between actions with a maximum gap of 15 minutes, the first edit count for 5 mins' AtUkr (talk) 08:59, 15 January 2026 (UTC)

From the Liaisons of the Board of Trustees

Dear Schiste,

Thank you for your valuable analysis and input.

We confirm that the Board is currently reviewing and discussing the material.

At the end of January, over 100 Wikimedians will gather in Frankfurt, Germany, for the Wikimedia Futures Lab, including representatives of the Wikimedia Foundation. Your essay is a good addition to the debates we expect to happen during the event and we hope you are able to join them.

We appreciate you taking the time to provide this input and will consider it as part of our ongoing deliberations. Should further clarification or information be required, we will reach out.

Community and Affiliates Liaisons to the Board

Victoria (talk) 19:05, 14 January 2026 (UTC)

Some feedback

An interesting essay, some comments below. Of course, this is also just my opinion.
1. Wikipedia is not and should not be a mechanism for finding "truth." Truth is sought elsewhere, not here. Moreover, it's doubtful that the mechanism is "scalable". Attempts to revive Strategy and the "knowledge as a service" concept also seem misguided. Perhaps the whole strategy was an attempt to introduce alien corporate practices into Wikipedia. A lot of time and effort was expended, and the attempt ended in failure.
2. Many methods of increasing traffic are only effective in the short term. Metrics should be treated with caution; they have no value in themselves and fixation on metrics can also have negative effects (see also Goodhart's law or Campbell's law, Chang and etc, 2024, Muller, 2018, Hayek, 1964.)
3. I would like to add that that the problems in the "West" also aren't minor. In many editions and in Western countries have problems with young people (for example, about Germany and the USA - Measuring Brand Health, Figure 8; about young readers in German Wikipedia - Readers Survey 2024). Also in the academic literature, the issue of homogeneity and diversity is debatable; it is doubtful that this is the most significant factor or that the effect is only positive, see, for example, Ostrom, 2009, p. 190 or Mokyr, 2025, p. 301.
Reductio ad absurdum: I could say that the English or Spanish Wikipedia community allegedly is more "diverse" than the Hindi Wikipedia community, that's why there are few articles in Hindi (this conclusion would be strange: the number of articles was influenced by many factors, including the level of education, the availability of reliable sources in Hindi, bilingualism, the skills and proportion of young people, the use of mobile phones, and historical reasons).
4. The problem has many causes, and there is no panacea. Oddly enough, your point of view may even be overly optimistic. With a high degree of probability, it will not be possible to achieve a boom. Investment strategies will not help. Unfortunately, we'll have to adapt to a new reality in which Wikipedia's importance will decline. In any case, such scenario can be considered a potential one (Option B). Proeksad (talk) 16:57, 15 January 2026 (UTC)

It's not absurd: It's probably correct that the English Wikipedia is more diverse than the Hindi one.

However, I never understood why do so many Wikipedians make such a big deal out of the question of the language in which sources are written. Sources for particular claims in a Wikipedia article can be in any language, as long as they are reliable. There is a certain preference for sources in the language of that Wikipedia, but even that can be overridden if there is a better source in another language.

That said, the existence of many sources in the same language in general is a relevant issue. If a community of speakers of a given language is accustomed to seeing many newspapers, textbooks, journals, websites, encyclopedias, etc., in their language, it's easier for them to work on a Wikipedia because they don't have to make a lot of effort to create a style of informational and educational writing. It's a completely fixable problem for Hindi and for any other language, but it requires effort. I wrote something related in the section "The billions who came online found an encyclopedia that didn't speak their languages". Amir E. Aharoni (talk) 19:27, 16 January 2026 (UTC)

Maybe yes, maybe no, it also depends on the meaning of the term "diversity." I was talking about a causal relationship between diversity and the number of articles in an essay. That's clearly an oversimplification; there are many factors involved, in any case. Perhaps, in general, we should not rely on naive ideas about which practices are “good” or “bad” without even analyzing their effectiveness. --Proeksad (talk) 15:59, 18 January 2026 (UTC)

Great summary, confused analysis

The summary of Wikipedia usage over time is great (other than not all the y-axes starting at 0, as mentioned above), as is the split of Wikipedia's history into three logical eras. But I find the analysis/recommendations unconvincing. Surely it's not terribly surprising if countries and cultures that have largely been on the internet since 1996 are also more interested in reading, and writing, encyclopedias than cultures that have only been online since, say, 2016? And if that is indeed the case, will making loading of Wikipedia 10% faster on mobile devices really make a difference? For that matter, is this really a fundamental problem? To take an extreme example, American football is one of the world's most lucrative sports, even though 95% of the world's population will never see a game of it. Surely Wikipedia can survive, and thrive, by just appealing to people who like encyclopedias, even if they happen to be concentrated in the first world/"Global North"? From that perspective, it's the loss of usage from the kinds of people who were previously reading it that's the more fundamental problem. Yaron Koren (talk) 00:23, 16 January 2026 (UTC)

News coverage

This essay was mentioned in Wired yesterday [1]. Clovermoss (talk) 06:41, 16 January 2026 (UTC)

Forks

With dissatisfaction with the current position of WMF towards two BRICS countries, what do you think about the forks from Chinese and Russian Wikipedias? The essay didn't mention Ruwiki.ru and Qiuwen Baike and certain areas of the English Wikipedia being forked into independent wikis. Over the last decade, I found some of Wikipedia being forked after dissatisfaction. Ahri Boy (talk) 18:58, 17 January 2026 (UTC)

I'm not so familiar with China, so I won't say anything about it, but I am quite familiar with Russia, so I will speak to that a bit.

I'm not speaking for the WMF, of course, but as far as I can tell, the WMF doesn't have any special positions towards Russia. Do correct me if I'm wrong. The government of Russia does have a special position towards freedom of speech, so it heavily censors newspapers, television, and websites. Many websites have been blocked or closed down in Russia in recent years, and laws are in place to punish spreading information with which the government disagrees (it contradicts article 29 of Russia's Constitution, but the Russian Supreme Court, unlike the United States one, always rules according to what the president wants, and not according to what the law or the constitution say).

Curiously, last time I checked, Wikipedia is not blocked in Russia. The reasons for this are probably more social than political; the government probably thinks that blocking Wikipedia will seem too heavy-handed (as if killing thousands of people in a neighboring country for no reason is not heavy-handed; but who would grasp Russia with the mind?). But it does try to create an "alternative" in the form of ruwiki dot ru (please never call it just "ruwiki", because this can be easily confused with the actual Russian Wikipedia). The links between the government and the website are actually not conclusively proven, but the site's content is so obviously edited in favor of the Russian government's narrative that there is doubt that there is at least some influence.

The site doesn't have much of a readers audience. According to Similarweb, Wikipedia is a top-20 website in Russia, and ruwiki dot ru is not even top-50. Surprisingly, it does have some editors, but not nearly as many as the Russian Wikipedia does. Some of them are likely paid.

Experience so far shows that a wiki encyclopedia that presents itself as "general", but actually exists for the sole purpose of promoting a certain political or business agenda, won't attract a sizable general audience of editor and readers. Wikipedia attracts editors because they like truth. Most readers come to Wikipedia for the same purpose. An encyclopedia that is politically biased is not so attractive to most people, and paying people to edit isn't scalable; Wikipedia wouldn't be as big as it is if editors were paid. (And yes, I'm quite sure that even Mr. Musk's money won't scale to something like the real Wikipedia.)

So forks of this kind don't scare me. I kind of wish that there were other kinds of forks or other serious Wikipedia competitors because it could make Wikipedia better. Amir E. Aharoni (talk) 20:52, 17 January 2026 (UTC)

The ruwiki.ru fork has effectively closed down all editing. See [2] (in Russian). To continue editing, users must provide papers proving their academic qualifications that are valid only for editing articles in the field of their education. Le Loy (talk) 03:44, 18 January 2026 (UTC)

I just found a new report that on March 22nd, it is not possible to edit freely. What about the license after the date? Ahri Boy (talk) 11:32, 1 April 2026 (UTC)

We changed as much as the internet has

Thank you Schiste for your contribution. I know you have been brewing on this idea for a while, so I can't say I'm surprised to read this, based on our previous conversations. I recognize the trends, and can agree with most of the big picture that you're sketching.

I'm pretty sure that the trends you're describing are the result of an intersection of several causes, both contributing and counteracting to the effect. There is one theme that I'm surprised that it doesn't play at least a supporting role in this piece: how we changed as a community. This is not always pleasant, but we should also reflect in how our social construct has changed.

While I don't really have solid numbers to back this up right now, there are hints to this hidden everywhere: we have also changed drastically as a community. My thesis would be: even without the AI revolution, we would have been in deep trouble regarding our editor recruitment. There are a few specific elements I would like to light out - there are possibly more: Approach to newcomers, Bureaucratization and Collaboration.

As a community (a community of communities, if you prefer) we have drastically changed our approach to welcoming newcomers. This can explain at least partially many of the trends we're observing: the people who are over represented in our editing community, are the type of people that got their foot in the door when we still welcomed new contributors. They are the type of people that were confident in being online, that were competent in writing semi-code and they (we!) set the tone for everyone who was to come after us.

When I think back of how the community welcomed me, and even more when I listen to the people who came before me how they were welcomed, that experience was nothing like how we currently treat our future potential colleagues. We had an exciting field to build on, experimentation and creative approaches were encouraged and complementarity was celebrated. All policies that we had, the software and our effective treatment was designed around the same principles: we invite you to make this edit, and we need your help!

At some point, this started to shift. In hindsight not very surprisingly: as Wikipedia became perceived as more valuable, as we started to grasp the impact we had on the world, we felt more and more a responsibility to protect that value. As a small group of Wikipedians was "fighting vandalism", they maybe got burned out a little too much, and I have heard many instances where our effective implementation of protecting content started to rub off negatively on potential colleagues. Some very tangible (paraphrased) examples I encountered:

• Recently an archivist told me how they tried to fix an article, but they were blocked after they posted something on the talkpage. It turned out, their organization's wifi was probably blocked for reasons nothing to do with them.
• A while ago I was at an in-person event, and I was unable to even create an account for 3 different people, because of our IP-blocking measures on mobile ranges and after that attempt, any device that tried creating an account from such a blocked IP. It took me about an hour to navigate the various exempt policies. People give up well before that.
• I was explaining to someone that Wikipedia is special because we are transparent how our pages come to be, and how the talk page works. Until they pulled out their mobile phone and I tried to show them the history tab and the talk page. After ten minutes I found the hidden link to history in the bottom of the page, but I still wouldn't know how they should find the talk page, or even categories. We created nice workarounds for ourselves in our own environments, but hide the valuable components from potential editors.
• In the Dutch language community, the phrase "bij twijfel niet inhalen" (leave the status quo unless you have a good reason) has become mainstream in place of "voel je vrij en ga je gang" (be bold, and go ahead), which is a paradigm shift. This results in many edits being reverted and rather than encouraging new contributors that their view is equally valuable, we're telling them that we know better.

When people do make it through that very first step, they encounter what you also touched on: our expanding bureaucratization. You're right: most of our bureaucracy is with 2005 in mind - I would phrase that as with an "exponentially growing community" in mind. A little while ago I started making a list of all the committees and boards that our movement has, after realizing once again that we're having a hard time recruiting volunteers for many of these positions. My count is well into the hundreds and I am procrastinating the completion of that count.

Our most limited resource is volunteer capacity, and we're spending much of it in very inefficient ways. If that means that these volunteers get energy, experience and personal benefits out of it - great! But I don't know if this is always the case. This is not limited to movement dynamics, but also includes our community processes. Eventually, there are just no more volunteers to be found. I observe a challenge to find suitable candidates for various committees, a public example being the Dutch Wikipedia arbitration committee that is now without any members. Think by yourself: if you were to start a new hobby, how excited would you be to be welcomed with talk about committee X, how organization Y is opposing our views etc?

It is really easy to add another committee, another layer of consultation and decision making - and in the short term that may be the safe thing to do. In the long term, that's suffocating, it burns out volunteers and changes community dynamics. While I am a big proponent of democratic and deliberative decision making, we're really not that good at doing this efficiently, of effectively at scale. We need to drastically rethink how we 'do' movement decision making. I don't have a ready-made solution but while centralizing it in the WMF may seem tempting at the short term, I don't think it would be sustainable either.

Finally, and I'm not sure if that is something we've gotten worse at, or that I've gotten better at recognizing, is that we're not that great at true collaboration. This is the case both for the way we edit our pages, and for the talk pages where we have our discussions. While there are certainly pockets of Wikipedia where there is great true collaboration, it gets harder for me to find really encouraging examples. In the editing process, I see more and more frequently examples where editors are flagging how others are wrong, and reverting them, rather than trying to help them to do things right. Consensus seeking is harder than it seems to have been in the past, as society is also less amenable to it. The number of controversial topics isn't exactly shrinking. And when we think about behavior that we tend to celebrate, it is often (and this is nothing new) that we celebrate people who write really deep articles all by themselves, or create lots of content - not the examples when people help other people shine and embed their knowledge more effectively in the bigger project.

We all know unfortunately examples where talkpages or on-wiki conversations get drawn out far too long until basically some people give up. There is always a subtle reward for community members that just keep hammering their points and have the most time available. This is more true in some topics than in others, but as we get spread out over a larger number of talkpages, I believe that this effect reinforces itself. I think we need to revisit how we do this 'truth seeking' mechanism (I'd be a little more humble than truth-finding, personally) that you admire. While I believe you're correct that this was an important element of Wikipedia's success, I'm not entirely convinced that it's not also broken in it's own right.

Anyway, just some thoughts that I hope are complementary to your writings. As someone who has been a very active participant in much of the movement myself, I also feel some responsibility for these shifts. I have first-hand observed the temptation of forming more committees, and have frequently advocating for another decision layer. If I carefully think back to it, I think at a micro-level, I may still have been right. But it is really hard to take a step back and look at these problems as a systemic issue. (P.S. I should mention briefly that I didn't want to suggest that there is no overlap between my points above and Schiste's essay -- with a long piece like this, there is even likely something I overlooked. These are just the elements I realized as top-of-mind after reading it and the comments already above (which likely also have some overlap).) Effeietsanders (talk) 10:23, 18 January 2026 (UTC)

Bring Back Knowledge Engine

I think some of the concerns here about the decline in Wikipedia's relevance are certainly true, but a lot of this talk is overlooking a really simple solution. The main driver of web traffic to Wikipedia right now is search, but since search engines now put their "AI summaries" at the top, a lot of people don't bother to look at an article before their question is (often inaccurately) answered. The solution is really not that complicated. Wikipedia needs its own search engine to break its traffic dependence on other search tools. Heck, you could even use AI to select excerpts of articles to create "ChatGPTesqe" responses that still use human content (something Google used to do a lot before). As the author here has noted, Wikipedia has unique access to uniquely trustworthy information, and a dedicated team of people who maintain it, we should use that. An open search engine that "anyone can edit" could be the next step of Wikipedia. The last effort at this did fail on account of a lack of transparency (and lobbying from donors), but we can and should learn from our mistakes. A new "Wikisearch" could be a project harnessing both sides of Wikipedia, foundation (to run backend ,and infra) and community (providing tooling, source selection, and quality assessment). Allan Nonymous (talk) 19:35, 21 January 2026 (UTC)

This would face two main difficulties:

• it would be very costly and difficult and building a whole new Web search engine may also be perceived as outside the scope of Wikimedia
• if you built such a tool that doesn't mean it will be used much (and it probably won't be as good as DuckDuckGo & Google)

For Wikimedia internal search improvements related to your ideas, see this and this. Maybe you were only referring to a search of Wikimedia content, that's a bit unclear to me but if you were then you didn't consider that Wikipedia unlike the Web doesn't have info on nearly everything people are searching answers/info for. Prototyperspective (talk) 18:58, 22 January 2026 (UTC)

Support and discussion

(note: moving this comment by User:~2026-42250-8 from the essay page diff)--CristianCantoro (talk) 15:37, 22 January 2026 (UTC)

• I fully agree with some points, and disagree with others.
  • We are a volunteer community that works to document the world in the most complete, rigorous, trustworthy and accessible way, and share our works with the world.
  • We won't and shouldn't try to be a better information technology company, because we aren't as well funded and single-minded.
  • The WMF leadership has shown to be concerned with geopolitical powers. We just want to write and film stuff, not work to empower corporate bots. We must insist on handcrafted authorship.
  • A notoriously more complex issue is that most people have very minimal technical leadership skills. We discuss the exact first paragraph about world leaders, but readers don't pay any attention to more than four consecutive words. Bias and disrust are already massive. Adding more bots won't fix that.

— The preceding unsigned comment was added by ~2026-42250-8 (talk) CristianCantoro (talk) 15:37, 22 January 2026 (UTC)

Thoughts reading this (most of them trash maybe)

I dont have much time reading this. But consider me "grasped" some of your analysis. I almost agree with the problems you stated and great job identifying them, helping people realize them a bit, but somehow I do not agree from either the things "you choose to recognize, and acknowledge", which is okay, consider you came from "the Global North" yourself. I came from Indonesia, English is not my strongest of suit, and I have difficulty composing my thought in them, but I'll try. Here are my two cents:

1. IDK why every community themed talks or writings or analysis fail to see that a lot of people see Wikipedia as "a hobby", "a leisure", "a non-essential activity that I'll do when I have time". Every time I joined community events, seminars, talks, I saw Wikipedia represented as "a 1 billion dollar mission that needs volunteers" (volunteers or unpaid workers?). I lost the joy of editing when I saw unending one sided false consensus, struggling sessions at project chat or teahouse, facing people with ideological baggage different than mine and acting cocky, or facing people who always brought up "oh Wikipedia Indonesia article number is less than English Wikipedia", without even give a peek at why at the grassroot level. Too many standards, too many guidelines, but still we look at other people with goofy face and a googly eyes "you too can edit Wikipedia!".
2. The reality is no, not all people can, and could and should edit in Wikipedia. You need to realize editing cost. It costs your time, your mental strength, your patience, your money. Money? Yeah, how could we edit comfortably unless we have at least laptops? "Oh you could edit with phones...", meh, learning that many guidelines, plus that functionality, meh, I rather do something else. We also need internet packages to keep connected and to watch our articles if they are nominated for deletions, or our edits if they got rollbacked. And we need money for books to refer to. What do you say? Library? Yes we have library but the quality is different from each region, and internet source are somehow, low quality.
3. I believe Wikipedia articles reflects the condition and quality of the press. If the press is biased, then the article is biased by default. You dont believe? Sure, no prob, but see how quick and how deep the press covers something when bad thing happened with the subject. "Bad news is good news". Media is third party, independent, etc. etc. suuuure, when the subject disagreed with their portrayal, they could not say anything, the Wikipedia article are there and cannot be reverted or edited "because a third party independent, verifiable, trusted media say that". And then we send them Streisand effect article. "The more you tried to negate what is written, the more you will got 'exposed' by the editors!" Suuure. Stamp that with NPOV-washing!
4. I am sorry for this, but I am deeply disgusted and despised with the "performative savior syndrome" (Our decline is the diversity debt coming due.). I mean, look at the flow of article from non English Wikipedia to Enwiki! Enwiki article, however bad they are, are free flowing to Indonesian Wiki, through daily translation and collaboration challenge for the new editors, but when Indonesian subject trying to be added to Enwiki, a bunch of editors immediately "thats not SIGCOV NPOV BLAHBLAH WPTHISTHAT, delete." The standards are too high, and do not have ability to see the conditions of resource and media other country had. Well, thats your answer to how Global North controls Wikipedia. I stop translating article a few years ago, and try to rewrite the article with available resources, and I have always 50% chance that the article wont last for 1 week, someone replace them with enwiki version, with english sources, carrying that Western worldview, framework and perspective. There are outliers, I believe, I saw influx of Indonesian diplomats, and several academicians added to Enwiki, but when I see the editor and the sources they added, well I can only say they are smart in hiding their COI.
5. Dont blame yourself. The problem is systemic. Critical thinking culture is sliding down. We champion ourselves as "the grandeur of open minded community", but when facing different POV, we scream WPNPOV on them. Some Indonesians have a phrase for this "our minds are open till our brains fell out on the floor" (open mindet sampek utekke ngglundung). Books are expensive, quality writing is now hard to find, and if they are dubbed qualified, the idea might not be good. Good ideas that are hidden behind invisible sources are left untouched. People leave Wikipedia not because teacher doctrine that you couldnt cite them, but because people are laughing at how we established authority over subjects. How often consensus are decided through hidden power struggle among the perspectives and ideologies. If you want to solve the problem Make people enjoy editing again, make people enjoy reading again, make Wikipedia a truly comfortable (not merely 'safe' whatever that means) place to start knowing things, and to stop Global North control let the standards be organic and dynamic.

I write this in burned out and tired. I lost the joy of editing. I am lost in the middle of community struggles. You could revert this, delete this, shame me for this, insulting me in the replies, I dont care. My love for Wikipedia withered away long time ago, and now Im merely a robot keeping this site alive. Thanks for listening me rant. Hope the next birthday will be better. ~2026-47483-7 (talk) 15:45, 22 January 2026 (UTC)

You say English isn't your strongest skill. Despite minor dialect or syntax differences, you expressed so many ideas that I can relate to so deeply.

My biggest challenges haven't been with cross-cultural or cross-language-wiki issues.

But so often I feel that when I go to a new project and read the beginner guides and policies and then try to contribute, senior editors come down hard with highly specialized jargon and criticisms like an old cranky rich white man slapping the hand of a poor child -- the condescension and patronizing nature of their responses seem to express a "don't even think about trying to help here, you'll just make more work for the REAL volunteers!" so I have given up after the swift rebukes and it's taken a few years each time to get up the courage to try on another project.

Even in the USA, having a phone or tablet costs money, having internet costs money, giving up a HUGE amount of time reading horrendously long documents that MIGHT actually help a new volunteer learn how to contribute costs money. Rich and middle class people can afford these privileges. The people being excluded aren't just in the "global south" - they are everywhere. Grayautumnday (talk) 18:34, 3 February 2026 (UTC)

Thoughts about article, perspective

This piece was really interesting to read and I think it gives everyone something to grapple with. Of course, the idea of Wikipedia becoming relegated to just an interesting site with issues maintaining the editing culture is really painful to think about, especially when you think about how true this place has stayed with such utter popularity.

It's difficult to think about how the way in which the majority of people consume information or just spend their time is shifting to such derivative content such as with LLMs. I remember as a child in the early 2010s (I'm only 22), I would always lookup websites on my own and go out and search for things that interested me, but as social media got more and more popular, I started to make my internet use mainly just consuming more derivative content. I think it's important to go out there and find new ways to show people how searching for the truth and the things that interest you is empowering and much more rewarding than scrolling Instagram. I got back into editing when I was watching Tor's Cabinet of Curiosities which was recommended to me on Youtube, so maybe it's safe to say there's a glimmer of hope and that we can make convincing ways to bring people out of content consumption and back into communities like this.

I would agree that this definitely prompts a change in funding and accounting for how people interact with the site, I just really worry that loosening the AI policy could become a bad choice if not done very sparingly or carefully. If changes are made to the site in such a way there might be a day where people talk about "the old days of Wikipedia being much better than now". We do need a new way of going about this era, but we have to be really conscious of how we can lose sight of what makes this place great. On first thought though, I think the idea of charging access to the data is a great way to go about it. ◀ Juniperol (bingus) ▶ --(talk) 07:16, 24 May 2026 (UTC)