On Sat, Jan 17, 2026 at 10:55 PM Leinonen Teemu via Wikimedia-l < [email protected]> wrote:
> I really enjoy this discussion. > > As it is the latency between human curiosity and insight(s), Wikipedia is > losing out to AI chatbots. > > The key in the AI chatbots is the UI, which feeds your curiosity by > imitating human-to-human dialogue. > > Would it be possible to build something like this to Wikipedia, too? I > would love to have ”chat” with Wikipedia. > > > - Teemu > > It is 100% possible. There are free and open source models we could build on, without creating a dependency on AI companies or doing anything that tries to replace human editing and authorship. I bet a lot of data scientists and AI researchers would love to help us. Building our own model would likely be a prohibitively expensive idea, but there are techniques for applying existing models to search a specific source of data i.e. only answer from Wikipedia. This is being done today by almost every existing company with a database they/their users want to chat with. Overall I would guess that the Foundation feels scared to propose this because many people dislike AI on principle for various reasons, and it might look like they are following a trend started by for-profit companies that are reusing Wikipedia content with zero attribution, to the Cunctator’s point about copyleft. (I have considered starting a civil class action suit by editors against OpenAI and other foundation model providers for violating Creative Commons terms.) A lot of the search problems that exist in Wikipedia are fairly basic, or at least don’t require AI chat to fix. For instance, we handle typos (like if you misspell Liechtenstein) very poorly, or otherwise searching for anything that isn’t the exact name of an article (if you search “Halti mountain” it will not show you the article for “Halti”, the tallest mountain in Finland, nor “List of mountains in Finland”). It makes me sad that we are the backbone of Google’s knowledge graph and AI models, but have made no serious attempt to do the obvious thing which is at least try and see if we could build such things based only on our content and with appropriate human control/safety in mind. This has been a good discussion in response to a usefully thought-provoking essay by Christophe. Many good points previously raised. Overall, I think Christophe was right to sound the alarm and demand greater urgency. I started editing as a teenager. I am now middle aged, and I want Wikipedia to be around and bigger/better when I am dust. We must replace ourselves with a larger, more thriving community if the project is to survive and grow. You might argue with Christophe’s data analysis, but it is not a good sign when I feel guilty that I can’t make time to do English Wikipedia admin things, because the admin pool has been shrinking on our largest project. Not all our problems are technological of course, nor do all of them require the WMF to fix. This is perhaps the one quibble I have with the essay—we are not so totally dependent on the Foundation. It just depends what we want to do, who is interested, and can we form a consensus. Steven Walling > - > > > *Lähettäjä: *Ori Livneh <[email protected]> > *Päivämäärä: *sunnuntaina, 18. tammikuuta 2026 klo 0.21 > *Vastaanottaja: *Wikimedia Mailing List <[email protected]> > *Aihe: *[Wikimedia-l] Re: Wikipedia at 25: A Wake-Up Call (essay) > > On Fri, Jan 16, 2026 at 1:35 AM Erik Moeller <[email protected]> wrote: > > A huge reason readers may prefer AI summaries, even if inaccurate, is > to get to an answer more quickly (the same reason Wikipedia itself > outperformed other information sources even when its quality was still > very uneven.) > > > Yes. It's latency. It was always latency. "Wiki" means quick. > > Wikipedia made the sum of human knowledge (or its arithmetic mean, anyway) > accessible near you: in your home, in your backpack, in your pocket, on > your person. It shortened the distance from question to answer by > abstracting the trip to the library or bookshelf, the time spent poring > over the table of contents, and the wait for the newspaper to arrive, for a > new textbook edition, or for the translation to appear in your language. > > For performance engineers like me, Wikipedia's "end-to-end latency" is the > time between a reader clicking a link and the article fully rendering on > their device. For many years, I believed the key to Wikipedia's continued > relevance was shaving milliseconds off this number by tuning Wikimedia's > code and infrastructure. > > But *true* end-to-end latency is not measured between server and browser, > but between curiosity and insight. And it turns out that network and code > latency contribute only modestly to that number. The milliseconds it takes > for Wikimedia's servers to transmit an article to your device are dwarfed > by the time you need to wrack your brain for the right terms to query, > locate the relevant section of the article, interpret its meaning, and > relate it to your question. > > Wikipedia improved on Britannica's end-to-end latency by several orders of > magnitude. Modern AI is now doing the same to Wikipedia. I can describe to > Gemini what I want to know using vague, imprecise, or even incorrect terms, > and it tells me what I might be thinking of, using my language: not merely > the language listed in my Babel userbox, but terms I understand that relate > to concepts I already know and are appropriate to my level of understanding. > > At its worst, AI generates hallucinated, sycophantic slop. But at its > best, it is an interface to human knowledge that is not merely > incrementally faster than browsing Wikipedia, but categorically faster. > > I think the key to ensuring the future knowledge infrastructure remains > free and open is to once again beat closed, commercial platforms on > latency, ideally by an order of magnitude or more. This is possible, if you > again consider true end-to-end latency and the invisible factors that > contribute to it, like the time it takes to distinguish truth from > falsehood, and information from manipulation. > > I'm not sure Wikimedia should lead the charge. Even if their relevance is > fading somewhat, the projects are an immense trove of value for humanity. > Any rash effort to remake them from within is likely to destroy more value > than it creates. But there is plenty of room out there for new things. > > I'm glad to see you experimenting in this space, Eric. > _______________________________________________ > Wikimedia-l mailing list -- [email protected], guidelines > at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and > https://meta.wikimedia.org/wiki/Wikimedia-l > Public archives at > https://lists.wikimedia.org/hyperkitty/list/[email protected]/message/UOAYYQLZNSVXQBOGLV2OSVX2V5SY2NUD/ > To unsubscribe send an email to [email protected]
_______________________________________________ Wikimedia-l mailing list -- [email protected], guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/wiki/Wikimedia-l Public archives at https://lists.wikimedia.org/hyperkitty/list/[email protected]/message/S7I3QKJIMIW5LZIXOU5MIDKD4PUOULUO/ To unsubscribe send an email to [email protected]
