Today I discussed with the founder of Wikipedia my nostr implementation, wikifreedia. We had a nice long chat about it. At the end I asked whether I should import all Wikipedia articles. “It will bias the type of information that is available in it and the tone” I don’t think a wikifreedia-like system needs to be data-complete. I think it’s fine if it has only fresh, selected/curated entries that embrace their bias.

Replies (55)

Thanks for articulating this better than I was about to. Biases are an inherent part of all data. Imperical data only exists in sterile environments.
pablo talking to jimmy whale, or whatever his name is first: he is saying that wikipedia is biased second: pablo, cocaine is bad for you third: if this shit involves a "consensus" and there is a game to make that consensus that doesn't make biasing it extremely expensive you just wasted a whole lot of time bro
If it's a truly open system it will have all the entries. Curation and selection has to happen at the edges; what seems to be a problem isn't one.
I agree. I actually said on the call that anyone can do this and explained the WoT approach. They were super interested in that as a novel way of doing away with consensus.
dude, you don't know jack about distributed systems, or even discrete mathematics and why this will be so overloaded with jackasses reediting their articles to their viewpoint when other jackasses reedit it to their viewpoint... have you even thought about how to keep a cap on that? are you going to prune the database of old entries or ratelimit them?
my bias that i find people who use weasel words and coercive "friend" expressions to pretend they win arguments because they actually didn't think about the stupid shit they just babbled? no, that is precisely me embarrassing you for being loud about dumb ideas and don't worry i have had this happen to me many times that's why i know it's good for you
I think it would be nice if there was maybe some "in the middle" solution. I don't think (but i am also not 100% sure) that everything is constantly being driven into certain narratives, it's just specific topics. Most of them will have some kind of political character ro try to protect some party's specific interest. I would think that if some article was created way back in, say, 2006 and only had minor edits that also happened a long time ago it's probably ok. But yeah maybe it doesn't have to start with loads of data. In the end you could still go to the wikipedia page itself but i think it would be great if switching back and forth qas not necessary. Maybe the question is more a matter of if it's more work to have everything and have people go at the absolutely tainted articles (a good way to find out if the trust system works) or start from scratch not having tainted articles but run the risk that it doesn't get much traction and it takes long for articles to appear at all.
On one hand it's just a small percent of articles that are biased, so wikifreedia maybe not be needed for the vast majority. But there are still a lot of "facts" on Wikipedia as a helpful starting point, even on articles with a lot of bias. It may actually help the final result be a bit more balanced/holistic. I'm not sure it works exactly the same, but have you seen Rationalwiki and Conservapedia? I think both start fresh with a very politically bias POV, and they're both garbage (although good that they exist).
wouldn't AI scrapers be used to fetch a variety of open source material on a given topic? it seems to me wikipedia is obsolete maybe an AI that analyses wikipedia entries for original source material, opinion, editor conflict of interests, and scores it with a transparent weighting process would be highly amusing, and valuable and very irritating for the spooks
maybe a crazy idea - but what if you did it like an overlay? ie - start with what's on wikipedia (ie rev proxy into it), an then allow anyone on nostr to fork an article into wikifreedia with all that it entails?
Chad Lupkes's avatar
Chad Lupkes 1 year ago
Not all pages, that's likely too much at once. However, if you could build a script that identified pages that are facing deletion because of irrelevance, you could save a lot of content that would otherwise disappear. Another potential source of articles would be the Fandom sites.
Build it properly from the ground up and it will be infinitely more powerful. Patience is needed for the new decentralized paradigm.
Wikipedia needs to start packaging their info into education "programs". (med, engineering, etc) All the info is there. It's just a matter of packaging it all into specific programs by linking and jump-linking the important parts.
pam's avatar
pam 1 year ago
Larry Sanger used to be vocal about it
pam's avatar
pam 1 year ago
i think its the right way to go - removing that single source of "truth" is fundamentally crucial. It may not necessarily be mutually exclusive to being data-complete as the latter is more related to the number of contributors. But the idea of enabling individual thoughts and biasness would shake up this silo regimes
Ai rotten tomatoes for Wiki and journalistic enterprise? Maybe combine that with community notes from X and end up with an interesting project. Partner with internet archive to stamp important events into history before they are edited back out ?
Default avatar
tbc 1 year ago
It was great to meet you, Pablo! Hi everyone else. Pablo is referring to Larry Sanger. He's my boss. We work on the Encyclosphere project. We're building the universal network of encyclopedias. I think there's overlap between our project and Wikifreedia.