In the world of ever-growing AI slop, every authentic record becomes priceless.
Paper books, vinyl audio records, old CDs, verified digital archives...
I was thinking it is worth creating a database of hashes for every human-generated piece of content, so authenticity can be checked. We already have the training data corpora, might as well ensure we know it is human-generated.
Automation is inevitable. Growing AI is inevitable. Maybe it will become better, maybe not, but people will become accustomed to it. The better AI is, the more people will believe it. At some point it will replace critical thinking, just as newspaper were shaping popular opinions.
I have always been afraid that Google might be manipulating popular opinion, erasing, rewriting history by changing links it suggests. That is why I like archive.org; I can always go back to see state of web pages before. No rewriting history.
That is why I also keep my own repository of links https://github.com/rumca-js/RSS-Link-Database so that I always could return to links, that are important to me. I do not trust Google, or any big tech giant.
I recalling my experience with newspaper archives in college. It's thirty been years since I was in college, but I assume that libraries still have some kind of copy of newspaper articles onsite, or do they?
Back then there were microfiche scans. I assume they're not making new microfiche scans, but do they have scans of old newspapers available on a computer? If you want to do primary source research for a history term paper how do you do it?
Those microfiches have been digitised and OCR'd while newspapers from the 90's onwards were created digitally and as such are archived in some processed source form - PDF or similar.
Here's a list of newspaper archives containing such material:
6 comments
[ 0.18 ms ] story [ 33.8 ms ] threadPaper books, vinyl audio records, old CDs, verified digital archives...
I was thinking it is worth creating a database of hashes for every human-generated piece of content, so authenticity can be checked. We already have the training data corpora, might as well ensure we know it is human-generated.
I have always been afraid that Google might be manipulating popular opinion, erasing, rewriting history by changing links it suggests. That is why I like archive.org; I can always go back to see state of web pages before. No rewriting history.
That is why I also keep my own repository of links https://github.com/rumca-js/RSS-Link-Database so that I always could return to links, that are important to me. I do not trust Google, or any big tech giant.
Back then there were microfiche scans. I assume they're not making new microfiche scans, but do they have scans of old newspapers available on a computer? If you want to do primary source research for a history term paper how do you do it?
Here's a list of newspaper archives containing such material:
https://en.wikipedia.org/wiki/Wikipedia:List_of_online_newsp...