491
Wikipedia Cofounder Larry Sanger Banned From Site for ‘Canvassing’
(www.404media.co)
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 23 Jun 2026
491 points (100.0% liked)
Technology
85675 readers
3424 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 3 years ago
MODERATORS
Not sure if troll or...
...ok, I'll bite. This is maybe the stupidest thing I've ever seen. You know that Wikipedia comprises one of the largest publicly-accessible sets of meticulously-curated natural-language data on the planet, right? And so when you're training up a new model, you're naturally going to start with that massive, freely-available repository?
Every major AI model has been trained, at least partially, on Wikipedia. This insane viewpoint is essentially saying that you don't think Wikipedia is reliable, but if some linear algebra chews it up for a few minutes, then it's ok. You're turning up your nose at tap water, but drinking your inside dog's urine.
Do you know how many living people are unable to get probable lies about themselves removed from wikipedia? Wikipedia is barely better than just taking random reddit comments as truth.
Gonna need a source for that, boss--
Wikipedia's intense scrutiny of sources and requirements for reliable citations are actually one of the reasons that Sanger started his malformed crusade.
If there's actual, provable lies about a notable person in the encyclopedia, then there should be actual, provable truths to combat it; and any Wikipedia editor can update the article in question to correct the record. If an edit war emerges, a community discussion can take place wherein the person in question can have their say. Wikipedia isn't the wild west, and any reasonable argument that it is died twenty years ago.
--but even if those two things weren't true--it's literally impossible to remove misinformation from AI models. I'm not saying that to be dramatic or overstate the problem. When a model is trained with misinformation, the misinformation becomes a part of the model; the entire corpus of everything it was trained on is baked into the neural network on a fundamental level, and humans can't manipulate it manually. Which means you can't remove any datapoint from the model without excising it from the training data and then retraining a whole new model.
So now not only are you drinking your dog's urine, you're claiming that the tap water is too yellow. Even if your assertion's true, your alternative is demonstrably worse.
No, not necessarily. Lots of false info spreads around, including serious academic publications. People who publish books and articles don't always do additional verification of the stuff they read elsewhere. And if nobody publishes something containing the correct version of the story, you as a WP editor don't really have a reliable source that you can use against the existing ones. I've seen this happen multiple times. Wikipedia is nominally meant just to convey what the sources say, not do active research or provide you with the capital T Truth.