If nobody uses Wikipedia they won't get any donations, and unfortunately they wasted the last two decades blowing literally hundreds of millions of dollars on random community and outreach programs instead of building an endowment in case something exactly like this happened.
No really, it was in the news a few years ago but nothing changed as far as I know.
I'll go out on a limb and say we _need_ Wikipedia and it's okay that traffic falls.
Physical print encyclopedias got replaced by Wikipedia, but AI isn't a replacement (can't ever see how either). While AI is a method of easier access for the end user, the purpose of Wikipedia stands on its own.
I've always scoffed at the Wikimedia Foundation's warchest and continuously increasing annual spending. I say now is the time to save money. Become self sustaining through investments so it can live for 1000 years.
To me, it is an existence for the common good and should be governed as such.
Wikipedia is the best thing ever. Anyone in the world can write anything they want about any subject, so you know you are getting the best possible information.
The Wikimedia Foundation is unlikely to last for 1000 years as an organisation because it doesn't exist within a social / economic system that will last for 1000 years. The US government is already actively plotting against it. Sure, they can try hopping from one country to another, but it won't be sustainable for that long.
It's not even certain if Wikipedia itself can exist for such a long period, given fragility of technological civilisation and data storage.
LLM's have definitely replaced 90% of what I used to look up on a Wikipedia, simply because they integrate from so many more additional sources.
But at the same time I continue to contribute edits to Wikipedia. Because it's the source of so much data. To me, it doesn't matter if the information I contribute gets consumed on Wikipedia or consumed via LLM. Either way, it's helping people.
Wikipedia isn't going away, even if its website stops being the primary way most people get information from it.
>It’s the dishonesty of Wikipedia that bothers me. The implication is that donations are urgently needed to keep the website running. In reality they have $300m in the bank and revenue is growing every year[0]. Even Wikipedia says only 43% of donations are used for site operations[1], and that includes all of their sites, not just Wikipedia. Fully 12% of the money they collect from you is. . . used to ask you for more money[1]
AI seems obvious, but social video? Are they saying people watch TikToks instead of reading Wikipedia, or people who used to look things up don’t bother anymore because of TikTok?
traffic falling means wikipedia will be cheaper to run. since they don’t rely on ads, it’d likely not affecting their revenues either (assuming those who don’t use it anymore weren’5 those givîng to it)
It's all right. Wikipedia was a magical device for its time, and it's still a great aggregator of information. It will probably last forever as such a link aggregator. Read-time curation is obviously far better than write-time curation, but the former used to be very hard. Now we have the former for cheap so it makes sense for Wikipedia to be Yet Another Source into the read-time curator. And the existence of a source database like Wikipedia makes many of these tools work a lot better.
People rightfully get upset about individual editors having specific agendas on Wikipedia and I get it. Often that is the case. But the chat interface for LLMs allows for a back and forth where you can force them to look past some text to get closer to a truth.
For my part, I think it's nice to be part of making that base substrate of human knowledge in an open way, and some kinds of fixes to Wikipedia articles are very easy. So what little I do, I'll keep doing. Makes me happy to help.
Some of the fruit is really low-hanging, take a look at this garbage someone added to an article:
How is it not conflict of interest when Google's AI summary (which is sometimes hilariously wrong) takes a click away from websites that pay for ads? Specially if it was trained on those websites
The sites paying for ads are usually offering products and services, not information. If you search "Plumbers near me", there will be no reason to have an AI summary, so they can show the ads. If you are searching "what temperature does copper melt", no one was buying ads on that anyway.
My personal traffic to Wikipedia fell after around 2019, when activist editors took over, and the site ceased to be trustworthy for a lot of important topics.
There are multiple new start ups and existing competition that's impacting them. They are facing competition.
The curious thing is that big LLM folks put together RAG systems which act much like wikipedia. But it's more than that. They built dictionaries, book repos(borderline illegal), news repos, and data knowledge base. These are bigger than wikipedia. Better because you dont have anonymous partisans.
Wikipedia is at a point where they have purged multiple perspectives and it has left an unreliable systemic bias in wikipedia. They are dealing with this problem and competitors are popping up because of these problems.
(looks outside at the street) I'm wondering why would traffic be less because of AI search?? SO I asked AI. Its answer? "By analyzing data and demand in real-time, AI is able to predict peak travel times and adjust train, subway, and bus routes as needed, reducing overcrowding and, in the case of buses, traffic congestion."
In the past, you would type a question into Google. To get the full, detailed answer you had to click on the Wikipedia search result.
With ChatGPT or Google AI Mode, you get all answers directly in the chat. And you can even ask follow-up questions. There is no need to click on a link.
From the data I have seen, 40% of searches on Google used to lead to a click to another website. In ChatGPT and Google AI Mode, this number is lower than 5%. One study (with a small N) even came to the conclusion that the number is 0%.
Looking how editors can fabricate stuff OR argue like little kids whether original source from origin country should be mistrusted vs encyclopedia britanica is enough for me to know Wikipedia is no longer what it was and lost its purpose.
So maybe they shouldn't depend on just search traffic. Their increased spend would suggest they got enough people to try and figure out alternative methods to attract editors and/or traffic.
I stopped contributing money, and then stopped trusting and using Wikipedia after I saw how wrong they are on a subject I am well-informed on personally (ARM architecture), and how fast they revert properly-cited 100% factually correct changes (literally citing the arm architecture manual).
How could I trust them on things I do not know, if I know for a fact they are unrepentantly wrong about things I do know?
That little AI generated summary that seems to show up in every Google search is riddled with errors. I've learned that as often as not it's just making things up or confusing two different topics. I would not trust that thing under any circumstances. Its no better than a wild guess from a stranger on the street.
You have to keep in mind that the original google search, by summarizing search results based on links between pages, essentially destroyed the web of interconnected links that existed at that time, so that the original pagerank algorithm doesn't even really work anymore. And the search results Google returns are worse than they were years ago. And you know, I use it, it is still super convenient, and the long term effects took some time to become apparent.
Now it is again feeding on and regurgitating Wikipedia but again in a way that will end up destroying the thing it is summarizing. Aggregators are parasitic on the thing they derive information from.
68 comments
[ 2.5 ms ] story [ 78.8 ms ] threadI always assumed the need for metastatic growth was limited to VC-backed and ad-revenue dependent companies.
No really, it was in the news a few years ago but nothing changed as far as I know.
Physical print encyclopedias got replaced by Wikipedia, but AI isn't a replacement (can't ever see how either). While AI is a method of easier access for the end user, the purpose of Wikipedia stands on its own.
I've always scoffed at the Wikimedia Foundation's warchest and continuously increasing annual spending. I say now is the time to save money. Become self sustaining through investments so it can live for 1000 years.
To me, it is an existence for the common good and should be governed as such.
WHAT should the Wikimedia foundation invest in, that's viable for a thousand years?
That requires a Wall St/hedge fund and/or Buffett mindset.
The Wikimedia foundation is none of those, and they're not big enough to make even a ripple in the investment landscape.
It's not even certain if Wikipedia itself can exist for such a long period, given fragility of technological civilisation and data storage.
https://www.usatoday.com/story/news/politics/2025/08/27/wiki...
the article you linked in no way supports that sentence.
(there's just a committee looking into claims that edits are used for propaganda in general.)
665 ChatGPT-User
396 Bingbot
296 Googlebot
037 PerplexityBot
Fascinating.
But at the same time I continue to contribute edits to Wikipedia. Because it's the source of so much data. To me, it doesn't matter if the information I contribute gets consumed on Wikipedia or consumed via LLM. Either way, it's helping people.
Wikipedia isn't going away, even if its website stops being the primary way most people get information from it.
https://news.ycombinator.com/item?id=34106982
>2022
>It’s the dishonesty of Wikipedia that bothers me. The implication is that donations are urgently needed to keep the website running. In reality they have $300m in the bank and revenue is growing every year[0]. Even Wikipedia says only 43% of donations are used for site operations[1], and that includes all of their sites, not just Wikipedia. Fully 12% of the money they collect from you is. . . used to ask you for more money[1]
gift link: https://www.wsj.com/health/wellness/anti-depressants-lifesty...
> “younger generations are seeking information on social video platforms rather than the open web.”
YouTube I'll buy it, but the search function in Instagram and tiktok is beyond useless!
People rightfully get upset about individual editors having specific agendas on Wikipedia and I get it. Often that is the case. But the chat interface for LLMs allows for a back and forth where you can force them to look past some text to get closer to a truth.
For my part, I think it's nice to be part of making that base substrate of human knowledge in an open way, and some kinds of fixes to Wikipedia articles are very easy. So what little I do, I'll keep doing. Makes me happy to help.
Some of the fruit is really low-hanging, take a look at this garbage someone added to an article:
https://en.wikipedia.org/w/index.php?title=Salvadoran_gang_c...
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5022797
https://media4.manhattan-institute.org/wp-content/uploads/is...
https://www.researchgate.net/publication/241750096_Is_Wikipe...
This is the kind of capitalistic behavior that is repugnant to our idea of how things should work.
This is not what the commons is for - taking the work of creators, repackaging it, and using platform capability to re-sell it.
At this point, I am coming around to the argument that governments should make their own local/national AI.
The curious thing is that big LLM folks put together RAG systems which act much like wikipedia. But it's more than that. They built dictionaries, book repos(borderline illegal), news repos, and data knowledge base. These are bigger than wikipedia. Better because you dont have anonymous partisans.
Wikipedia is at a point where they have purged multiple perspectives and it has left an unreliable systemic bias in wikipedia. They are dealing with this problem and competitors are popping up because of these problems.
Larry put out his theses on his user page: https://en.wikipedia.org/wiki/User:Larry_Sanger/Nine_Theses
Originally being heavily censored, vandalized, and deleted. It does seem to have been allowed, despite it being in user space.
Every one of those theses is correct.
With ChatGPT or Google AI Mode, you get all answers directly in the chat. And you can even ask follow-up questions. There is no need to click on a link.
From the data I have seen, 40% of searches on Google used to lead to a click to another website. In ChatGPT and Google AI Mode, this number is lower than 5%. One study (with a small N) even came to the conclusion that the number is 0%.
The further aways ppl stay from it the better.
How could I trust them on things I do not know, if I know for a fact they are unrepentantly wrong about things I do know?
Now it is again feeding on and regurgitating Wikipedia but again in a way that will end up destroying the thing it is summarizing. Aggregators are parasitic on the thing they derive information from.