Hi Luis,

Thanks for this - valuable thoughts.


> I would go a much different direction. After the 1st quarter, *our
> strategic process has to start by asking the hardest possible question:
> what if "reading an encyclopedia" is mostly over, like reading a print
> newspaper?* In other words, what if our long-term on-wiki readership
> graph looks like this one?
> https://www.pewresearch.org/journalism/fact-sheet/newspapers/
>
> *So, for purposes of strategic discussion: assume that within 2-3 years
> our on-wiki readership will be at 5% of peak. Or even braver: assume that
> on-wiki readership goes to essentially zero, with only editors reading it. 
> *I'm
> not saying this is the most likely outcome, just that approaching the
> question from that frame is much more likely to generate interesting and
> important ideas than trying to slowly tweak our decline.
>

To my mind, this is essentially the future we need to grapple with.

Wikipedia is just starting to experience what Encarta did for the
Encyclopedia Britannica, and what we did for Encarta. Crowdsourced humans
turned out to be cheaper and better than expert humans, now machines are
cheaper and better than crowdsourced humans.

When people have factual questions they are increasingly finding the
answers in AI-assisted browser search or direct in AI tools. These tools
have gone from 'fairly useful' to 'very useful' in just a few years. They
are well above the threshold of reliability most people have for everyday
use, while being somewhat more accessible and somewhat more helpful (I can
have a conversation with an LLM when I'm seeking to learn about a subject,
for instance... can't do that with a wiki page). Reliability and detail is
likely to improve further.

If this continues, Wikipedia is obsolescent - by the end of the decade it's
quite likely it'll be very big hobby project, falling into an awkward gap
between LLM-generated content and actual books,with minimal traffic. Large
parts of the Wikimedia ecosystem will also be obsolescent - in 2010 one
could usefully open up knowledge with a partnership with a university, but
there is less value to that if the LLMs have already read all the
scholarship that the academics use; and Wikipedia editing workshops will no
longer be a useful thing to do.

Our mission is to provide the sum of all knowledge to everyone, for free,
in a language they can understand. Part of our future mission could be to
ensure that 'the sum of all knowledge' is sufficiently broad, and
sufficiently broadly interpreted, and that breadth is incorporated in LLMs
of all shapes and sizes - knowledge equity again. Another part could be to
ensure there is equitable access to modern AI tools. Perhaps we need to
pivot to a core product of WikipediAI, our own ethically-sourced,
open-to-all, reliable learning LLM - though the magnitude of the costs
involved in creating a good LLM, the challenges of alignment and of
restricting use to the intended use cases would be significant.

Whatever the future, ideas to make Wikipedia a bit better (even with
AI-generated content) are not the solution. I am also somewhat sceptical
about e.g. Wikidata, LLMs appear to have little problem 'thinking'
conceptually with only natural language inputs, so projects to formalise
and connect concepts beyond natural language also seem a bit redundant to
me (though I confess this isnt' a very informed view)

Chris
_______________________________________________
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/6IDJ2FNG62CRBKUTFPOLWW5BYRLP4PZH/
To unsubscribe send an email to [email protected]

Reply via email to