pull down to refresh

Wow. Dedication!

I was talking to this dude that is building systems for journalists and he promised he will send me a list of the FOSS tech stack he uses for AI-assisted case management so that I get to not have to figure it all out by myself. It's still on my list to build something for that as a self-hosted package, because I get really tired of manual labor involved with tracking shit.

100 sats \ 1 reply \ @Scoresby 28m

I like to believe that scrolling through all these things helps me to notice stuff that I never could have got keywording.

(anecdotally, I have certainly come across details and connections I don't think I could have found any other way)

If, however, there is a better management system than very long .txt files, I am curious about it.

reply
102 sats \ 0 replies \ @optimism 17m

I agree. The problem I have is that I hate clicking load more on nitter. And the awful web.archive.org UX. And search results pages. And paywalls. And all the things.

So the idea is to programmatically extract all the links, images, dysfunctional paywall shit, videos, feeds... everything. And leech it. Then have an algo index it. Then NLP. And then finally anchor everything towards everything.

Then we can Sherlock everything. Find a new lead? Scrape the hell out of it and everything it links to and archive it all

reply