• Psychedelic Drugs Welcome Guest
    View threads about
    Posting RulesBluelight Rules
    PD's Best Threads Index
    Social ThreadSupport Bluelight
    Psychedelic Beginner's FAQ
  • PD Moderators: Esperighanto | JackARoe | Cheshire_Kat

Miscellaneous Looking for help creating a psychedelic archive

red22

Bluelighter
Joined
Nov 23, 2009
Messages
1,200
I've been collecting psychedelic-related content and I want to sort through it and make an archive out of it or contribute to an existing one. I've been having a lot of trouble focusing on this task, so I thought it would be easier if I could work with a like-minded person, in person. I can reside in your residence or you could provide me with a van or camp site. I can pay for my food.
 
I like the idea, at least the broad contours that you've shared.

One thing to consider when structuring your project is the role of LLMs like ChatGPT as digital librarians going forward. They are getting better and better and crawling through the depths of archived forum content online, and can synthesize that material or make it searchable. I think that the role of digital archives is going to change dramatically in the next few months as everyone gains access to low-cost automated teams of digital research librarians.
 
Doesn't Erowid already fill this role?

And IMO skilled human librarians are far superior to any kind of software that essentially exists in a world of text. LLMs can appear convincing because of their ability to synthesize "believable passages" from the material they are trained on, but their ability to conceptualize is entirely dependent on textual relationships. Humans conceptualize on the basis of our lived experiences in which we are immersed in all of our senses and in which we also interact with other beings of awesome biological complexity. From an pure information standpoint, we are "trained" on far far more raw information than any LLMs can even come close to.
 
Doesn't Erowid already fill this role?

And IMO skilled human librarians are far superior to any kind of software that essentially exists in a world of text. LLMs can appear convincing because of their ability to synthesize "believable passages" from the material they are trained on, but their ability to conceptualize is entirely dependent on textual relationships. Humans conceptualize on the basis of our lived experiences in which we are immersed in all of our senses and in which we also interact with other beings of awesome biological complexity. From an pure information standpoint, we are "trained" on far far more raw information than any LLMs can even come close to.
and psychonautwiki, plants of the gods, pihkal, tihkal... so my first thought was exactly the same, but maybe @red22 is up to something new, a new format, whatever actually... erowid for one was my bible when I was 16 (the year 2000) and it will always have *a special place in my heart*, but that doesn't mean that I don't check psychonautwiki...

so @red22 tell us more about your project
 
Doesn't Erowid already fill this role?

And IMO skilled human librarians are far superior to any kind of software that essentially exists in a world of text. LLMs can appear convincing because of their ability to synthesize "believable passages" from the material they are trained on, but their ability to conceptualize is entirely dependent on textual relationships. Humans conceptualize on the basis of our lived experiences in which we are immersed in all of our senses and in which we also interact with other beings of awesome biological complexity. From an pure information standpoint, we are "trained" on far far more raw information than any LLMs can even come close to.
Well put. To be clear, I mostly meant using them as super-powered search engines. The fact that they can in seconds scour the depths of Usenet drug user threads from the 90s in seconds to bring up not just results that are directly tied to my search terms, but things that are adjacent that I would never have thought to look for, then cite the sources for me to do my own follow up? That's insanely powerful from a research perspective.
 
Top