I was reading this article about LLMs making bad citations. I found it pretty interesting, so I decided to try to replicate it with ChatGPT.
https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.php
I was reading this article about LLMs making bad citations. I found it pretty interesting, so I decided to try to replicate it with ChatGPT.
https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.php
Anyway, I took a paragraph out of the document and asked ChatGPT to identify the URL, publisher, publication date, and title. It failed. You can see the transcript here:
https://chatgpt.com/share/68573fa9-b340-800f-b9b4-7b74fdf0bf46
I tried it with a document I wrote, FEP 5711. It's an enhancement proposal for ActivityPub, adding some inverse relationships for important properties.
https://codeberg.org/fediverse/fep/src/branch/main/fep/5711/fep-5711.md
I was surprised to see that it had really no visibility of the FEPs. After a while, I realized that codeberg.org, the hosting service for FEPs, has ChatGPT blocked.
I guess there's maybe some justification like, those people are bad, and they don't deserve nice things like Fediverse Enhancement Proposals? Or, maybe, we have to take a principled stand against LLMs by not providing any training data for them? Such that, perhaps, people disappointed by not having good results in LLMs will return to using traditional search engines like Google or Bing, which are more ethical because reasons.
I understand the goal; many people don't want their code to be used by LLM code generators. But it also means that this document repository isn't visible for people who use LLMs like a search engine. Numbers vary, but afaict somewhere around 10% of people use LLMs as their primary search engine, and about 50% of people use LLMs some of the time for search.
@tobias @elshid @Codeberg well, that's hardcore, but the problem is that when people can't find your information on their chosen search engine, they don't go find another search engine. They don't even know they're missing your info.
Happy Trans Day of Visibility. Trans rights are human rights.
So, in my mind, "Would you share your Fediverse data with researchers?" implies that you have agency to share or not, and that you can consent or not.
But I guess you're reading it a different way.
@heluecht last question: have you ever participated as a research subject before in another part of your life?
@heluecht even for academic research?
@heluecht @jwildeboer but otherwise identical?
@heluecht @jwildeboer that seems like a meaningless distinction.
If you are sending an image or HTML text to specific people, there's no difference between a "direct message" and a "post".
What makes a direct message is that it's direct to specific accounts, not to "all followers" or "the public" or "a contact list".
That's all; it's a very easy concept.
@heluecht @jwildeboer I meant with the SocialCG, not just forum posts.
@heluecht @jwildeboer Are you still involved in development of Friendica? It would be great to see you get more involved in AP.
@heluecht @jwildeboer yes it does; you have to a) use HTTP Signature authorization and b) have access to the post.
@heluecht @jwildeboer Again, that's not true. The `followers` collection is public.
@heluecht @jwildeboer you don't need signatures. Just fetch the replies collection. The OP can use 'Add' or 'Accept' activities to indicate that an object has been added to the replies.
@heluecht @jwildeboer DMs are any content objects with only actors (no collections) in the addressing properties (to, cc, and so on).
He/him. Board member at CoSocial.ca.Director of Open Technology at Open Earth Foundation (OEF).Founder of Wikitravel, StatusNet, identi.ca, Fuzzy.ai. CTO of Breather, TRU LUV and MTTR.Creator of pump.io. Co-creator of GNU social.Co-chair of the Social Web Working Group at W3C. Co-author of ActivityStreams 2.0. Co-author of ActivityPub. Co-author of OStatus.Grad student in CS at Georgia Tech.This network has been my life's work. Thanks for making it.
bartle is a social network. It runs on GNU social, version 2.0.1-beta0, available under the GNU Affero General Public License.
All bartle content and data are available under the Creative Commons Attribution 3.0 license.