529
Trump had Epstein killed because he was going to "name names," Epstein's brother claims
(www.lgbtqnation.com)
this post was submitted on 24 Dec 2025
529 points (98.0% liked)
Politics
213 readers
216 users here now
Any politics anywhere in the world. Inevitably it'll be 99% US stuff, but that's not a rule.
This community works differently to how most politics communities work. It has strict rules designed to facilitate productive discussion. You can be rude, to a point, but you can't participate in bad faith:
- If you claim someone said something they didn't say, that's a temp ban.
- If you make a factual claim but then aren't interested in backing it up, that's a temp ban.
- If you're asked one or two reasonable questions about what you said, and you're still talking but you're pretending the questions didn't happen or rejecting the premise of answering them, that's a temp ban.
The idea is to make the discussion productive. Let's see how it works. Maybe this is a fool's errand but IDK how any set of moderation could be worse than lemmy.world.
Other misc rules:
- No unreliable sources.
- Keep it productive please.
- Self posts for discussion are fine. This includes videos or photos. No meme posts or screenshots please.
- No personal insults.
- No racism / transphobia / related bigotry.
founded 3 months ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
What? Use software to help, then go over it with a Couple pair of eyes to see what was missed.
Anything longer is just milking work or a complete lack of understanding of technology.
Did you do it word by word or something and spend a minute on each one deciding if it was private or not? Like seriously, this is the most wild claim I’ve seen in a WHILE.
The request specifically included staff emails, text messages, all plans for the site, etc. We had to pull paper records from off-site, scan 4'x3' pages through the plotter, etc. And OCR doesn't catch everything, so when you've got hundreds of pages on plan sets, some of them with random small hand-written notes etc on the pages it becomes a huge task.
And on other requests where the press gets involved, there may be a slight change in future requests from one requestor to another that requires us to look through a lot of the same data we previously searched for additional responsive documents. We had one particularly high-profile case where I produced 10,000 pages of responses for different reporters from about 1,500 pages of total data because different requests required different responses.
And little things like someone including PII as cursive writing in a JPEG in their email signature can be a killer because OCR can't find it so it requires require manual redaction. When that person has sent hundreds of emails to city staff and those emails end up showing up in reply chains, it can balloon really fast.
Then you'll get follow-up requests that snowball. For example, there are a few private permit expediters that turn in applications for developers and help them navigate through the city process. One who was the "applicant" for dozens of properties in town turned in an application for a high-profile person's house, and the reporters asked for all communications between that permit pusher and staff. Then when they saw all these other people completely unrelated to the case they were investigating showing up in emails, they asked for all emails between staff and those people, assuming they were associates of the high-profile person.
Ctrl f, or dictionary edit for a text file.
Seriously, there’s ways to use tech to make life easier, sounds like you’re just ignorant.
If you know how many to look for, makes life a hell of a lot easier lol. And someone or two double checks.
Ctrl F doesn't work on rolled sets of plans that haven't even been digitized. And OCR (the part that lets you search for text on scanned images) isn't super reliable on old scans, especially with hand-written blueprints and shit.
The thing with government records is they're messy, and getting them all to work together neatly isn't simple. Every 2-3 years there's some mandate from on high to change how everything is stored, and most of the old data, whether it's in an old binder or a Word Perfect doc from 1966 or anything else is only ever moved to a newer system when it's specifically needed - usually in an Open Records requests.
I did an estimate on what it would take to digitize od rolled construction documents for a city of 4,000 people I used to work in. Having the plotter running 8 hours a day would take about 2 years if that's all the employee did.
Is that what the DOJ is dealing with? 4x3s?
It sounds like your ability x 20 would have it done by the deadline.
I guarantee the evidence will include lots of document formats. It's not all emails.
And photographs are a huge problem because sometimes photos will have documents visible in the background, or other picturesin the pictures. Text on piece of paper on the table behind someone will not be easily recognized by software, but a human may be able to piece stuff together - especially if ther's context available from information gleaned elsewhere in the release.
It's not directly related, but I'll use an example from some epidemiological research I was doing for HHS a few years back on gonorrhea when we had to look out for HIPAA violations. To do that, or spatial data was going to be anonymized by limiting the precision to the census tract level.
The issue we ran into was sometimes that wasn't good enough. If you know a diagnosis was on January 14th and the patient lived in X census tract and was an African American female between the ages of 35 and 40 that is sufficiently anonymous unless there's only 3 african american women in the tract on the most recent census, and the appraisal district shows that it's all single-family homes and by using ownership records and some Google searching you can figure out that there's only one match for that age and race combination on that date, so you know that Jane Smith on 123 Maple Street has gonorrhea.
That's the kind of shit you have to look out for. Data that can be combined with other information to reveal something that you're trying not to reveal - like STI patients or SA victims. And not all of that information can be easily recognized autonomously. It massively increases the workload if an honest effort is to be made to redact what legally should be redacted while revealing the information the public is legally entitled to.
But honesty is an issue for this administration, so they're absolutely gonna redact a bunch of shit they shouldn't and will end up failing tonredact a bunch of stuff they should. The fact that the job is incredibly difficult doesn't excuse their evil and incompetence - it magnifies it.
Ok fair enough. I know it is when things like this cross your own area of expertise. Devil's in the details.
This administration campaigned on this, talked about this. I guess my point is not opposed to what you are saying other than - clearly they were not actually "working" on this
Absolutely. And they should have been working on this since January when the judicial seal on releaseing the files (for Maxwell appeals) was lifted. One month is nearly impossible. 11 months is plenty.
I never addressed OCR, that’s its own issue.
You’re just technologically illiterate and they have you doing redaction? Lmfao, no wonder it takes you so long.
If something is in text, copy it to another program like word or notepad, and then use those features to find how many issues need to be resolved. Stuff doesn’t need to be moved or updated, you’re just doing a shitty job and not using tech to make your life easier.
And yes, I know that blueprints can be drawn on a bloody napkin, that changes nothing.
You were saying to use ctrl F to magic the problem away. That requires OCR.
That's the whole problem - there's no one magic trick to solve the issue. It requires manual scanning and manual review.
Some archotectural or site plan building sets include 200-page rolls of Arch E sheets of mylar. For reference, laid out on the ground that would take up about 6-times the floor area of a 2-car garage. That's over 10 hours of just scanning.
Some of those pages contain 6pt font notes, and we're required to review it ALL manually because there's a bunch of information we're not allowed to release in an ORR (PII, certain critical public infrastructure details, security-related issues, email addresses, information hinting at the identity or residence of anyone related to a retired court officer, and more).
Uhh… because you specified emails and text messages. Which don’t require OCR. So it can be used in THOSE situations. I also never said it was this magic do all, I specified it was a TOOL to HELP.
If you can’t even read a comment before responding, I’m done. Cheers troll!
I made a post that was 3/4 about OCR-related issues (including why it's specifically an issue in emails with embedded images containing text). You made a flippant reposnse that didn't address OCR issues and called me out as ignorant and lazy.
I responded explaining some OCR issues again. You called me technologically illiterate for not doing a text search.
So I explained for a third time that you can't just trust ctrl-F, and now you're saying I'm the one not bothering to read what I'm replying to.