this post was submitted on 11 Aug 2026
822 points (97.2% liked)
Technology
87123 readers
4105 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
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
How feasible is it (legally, technically) to host a federated, read-only copy of Reddit data from before The Enshittening?
Why bother?
Its all long since been injested by every gen AI model around, just ask one.
The problem with AI is that while 80% of the time it will provide correct information, but 20% of time it will generate very convincingly sounded bullshit.
This is the danger of it.. You test it few times and then you start trusting it. Kind of like that farmer that destroyed his whole crop.
I agree. However I don't think answers sourced from reddit are any more accurate.
Yeeeesss..., but typically it is easier to determine how reliable the person is from the test of the context.
Since Gemini is fed by the reason directly or might be fun exercise to ask it about yourself and while in some questions it will answer correctly or just loves to invent new things.
This works especially well if the person owning the account did not worry as much about privacy. So you hear many true things and also a lot of bullshit.
sure. that's an advantage of reading off reddit.
an advantage of asking an LLM is that it can consider multiple reddit posts and stack overflow and whatever else.
you're right that there's an eternal problem of determining whether information is accurate or just confidently incorrect, but I don't think that applies exclusively to chat bots.
It depends how mad Temu Elon aka Steve Huffman gets about it. He incorrectly believes everything belongs to him.
You could make a zim file out of it and share the zim file. I imagine it would be quite large though.