This is a noble gesture but you should know that this is a deliberate "drip" tactic that is utilized to soften the impact gradually, making the consumption and analysis of this data part of a broader social normalization process which works ironically against the interests of any substantial response that could occur.
Essentially we are helping them brain hack the masses into accepting the status quo.
I have heard it framed as a slow drip to keep the pressure on those who signed on from recinding their support. Like a one-way parking ramp - making it more and more painful to reverse course.
IIRC the slow drip was an explicit tactic used during the Snowden revelations and that seemed much more effective at getting everyone talking about it than Drake’s revelations or Congressional inquiries.
I would say these are 'tagged'; organized is a bit of a stretch, IMO.
This is some sort of searchable index that reminds me a lot of what the web was like prior to Google; the Altavista days. It's just a jumble of poorly formatted text that isn't really contextually aware and is largely useless for the volume of textual documents.
Then there's the whole 'I landed on an email that offers the text of an email that, I assume, the pertinent information is in the attachment listed, of which I cannot easily access.
--
I'd be more interested to see how you enhance this in v2.
Thanks for your feedback. Agreed it's not perfect and can be improved. That said, I wanted to put it out there in case it did help someone, even in it's imperfect state.
> It's just a jumble of poorly formatted text that isn't really contextually aware and is largely useless for the volume of textual documents.
I did a quick spot check and the lack of _clear_ date field is going to make contextualizing a bit trickier. It looks like most of the `email` have them but other types like `report` may have an unknown "first, created/circulated internally" date and a broader "the public can see it" date.
Nevertheless, it's only a matter of time before this gets loaded into a graph DB so the context becomes more apparent similar to what the journalists did for the panama papers.
Unfortunately it's several thousand raw text files, and determining a single relevant date, sender and message for each file is a bit more complex than I had time for this week. Next week we'll make the files more readable, ingest & display the images, amongst other usabillity improvements.
Yep, I've had a snoot-full of this project this week - but next week I'll start pulling in the images and making improvements. Thank you for the feedback!
This is a noble gesture but you should know that this is a deliberate "drip" tactic that is utilized to soften the impact gradually, making the consumption and analysis of this data part of a broader social normalization process which works ironically against the interests of any substantial response that could occur.
Essentially we are helping them brain hack the masses into accepting the status quo.
I have heard it framed as a slow drip to keep the pressure on those who signed on from recinding their support. Like a one-way parking ramp - making it more and more painful to reverse course.
Weird that you could argue for the opposite effect.
Slow drip and fight the quick news cycle forgetfulness by taking a front page spot for each drop.
IIRC the slow drip was an explicit tactic used during the Snowden revelations and that seemed much more effective at getting everyone talking about it than Drake’s revelations or Congressional inquiries.
Zeteo News published some searchable archive. It seems this links the original (?) PDFs and provides a Google-based search.
https://journaliststudio.google.com/pinpoint/search?collecti...
I would say these are 'tagged'; organized is a bit of a stretch, IMO.
This is some sort of searchable index that reminds me a lot of what the web was like prior to Google; the Altavista days. It's just a jumble of poorly formatted text that isn't really contextually aware and is largely useless for the volume of textual documents.
Then there's the whole 'I landed on an email that offers the text of an email that, I assume, the pertinent information is in the attachment listed, of which I cannot easily access.
--
I'd be more interested to see how you enhance this in v2.
Thanks for your feedback. Agreed it's not perfect and can be improved. That said, I wanted to put it out there in case it did help someone, even in it's imperfect state.
Marked improvements incoming...
> It's just a jumble of poorly formatted text that isn't really contextually aware and is largely useless for the volume of textual documents.
I did a quick spot check and the lack of _clear_ date field is going to make contextualizing a bit trickier. It looks like most of the `email` have them but other types like `report` may have an unknown "first, created/circulated internally" date and a broader "the public can see it" date.
Nevertheless, it's only a matter of time before this gets loaded into a graph DB so the context becomes more apparent similar to what the journalists did for the panama papers.
Great job!
Would be nice if the messages are formatted with just date, sender and message and the metadata is hidden in a popup or whatever.
Thank you for the feedback.
Unfortunately it's several thousand raw text files, and determining a single relevant date, sender and message for each file is a bit more complex than I had time for this week. Next week we'll make the files more readable, ingest & display the images, amongst other usabillity improvements.
How long until people build EpsteinGPT that does semantic search on this?
Looks great and terrifying, as expected.
Thanks for putting this together.
The stuff I am coming across in here is WILD.
If you could find a way to make some of this more easily searchable/readable/etc. (Minimize headings/convert links/fix line breaks etc)
BTW, the chat logs of other parties anonymous, but you can clearly tell these are with Steve Bannon:
https://searchepsteinfiles.com/file/text/HOUSE_OVERSIGHT_025... https://searchepsteinfiles.com/file/text/HOUSE_OVERSIGHT_027...
Nice finds!
Yep, I've had a snoot-full of this project this week - but next week I'll start pulling in the images and making improvements. Thank you for the feedback!
You need an AI to go through each one and assign a score of how damning it is so we can sort all files by that.
https://couriernewsroom.com/news/we-created-a-searchable-dat...
This is of course not what they're still trying to get released, so what is it based on?
The stuff at https://oversight.house.gov/release/oversight-committee-rele... ?