How is an AI that is being used to search the web for information, that crawls through studies, forums, etc a terrible source? It basically isn’t a source at all it just gathers it from different sources. The problem lies in the alignment and sponsorification of it and i would agree if this is about chatgpt but that wasn’t the point.
It might gather information from all those sources (with or without consent), but what it returns is no more credible than a story from a granny in your local market.
ONLY if you prompt it to return links, and read the information in those links yourself, only then, you’ve read information from the source.
It has been already proven that LLMs are bad at summarising - the only thing techbros have been pushing it for. It’s bad at summarising, bad at coding, bad at math and fucking terrible at image making.
There’s a reason the output is called ‘Slop’. And rightfully so.
I understand where you are coming from but Perplexity for example does actually list the sources its taking the information from and it seems to be pretty consistent and on point. Do you mind elaborating why the consent is important though? Are you against piracy and free information as well? It just seems that you have an irrational hatred against this technology, please correct me if iam wrong, i know there is a real danger stemming from it, especially in how govs will use it but i don’t see why using it for your own is bad in any sense, especially because of the real efficiency gain.
There’s not much room for interpretation here.
If you think that’s wrong - you’re wrong.
How is an AI that is being used to search the web for information, that crawls through studies, forums, etc a terrible source? It basically isn’t a source at all it just gathers it from different sources. The problem lies in the alignment and sponsorification of it and i would agree if this is about chatgpt but that wasn’t the point.
It might gather information from all those sources (with or without consent), but what it returns is no more credible than a story from a granny in your local market.
ONLY if you prompt it to return links, and read the information in those links yourself, only then, you’ve read information from the source.
It has been already proven that LLMs are bad at summarising - the only thing techbros have been pushing it for. It’s bad at summarising, bad at coding, bad at math and fucking terrible at image making.
There’s a reason the output is called ‘Slop’. And rightfully so.
I understand where you are coming from but Perplexity for example does actually list the sources its taking the information from and it seems to be pretty consistent and on point. Do you mind elaborating why the consent is important though? Are you against piracy and free information as well? It just seems that you have an irrational hatred against this technology, please correct me if iam wrong, i know there is a real danger stemming from it, especially in how govs will use it but i don’t see why using it for your own is bad in any sense, especially because of the real efficiency gain.
You really did read my comment explaining the shortcomings of statistical engines and called my ‘hate’ “irrational”, huh.
You should try some of that Natural Intelligence.
I’m done feeding the sealion.