transcription: “your pretty unstable.” y-y-you think im pretty?

The symbol for Amazon’s VGT3, the Las Vegas facility where it scans book for AI training data.

Amazon is buying massive quantities of books, scanning them for AI training data, and destroying them in the process.

A 404 Media investigation was able to reveal Amazon’s book buying operation, which hasn’t been previously reported, by placing a tracking device in a rare book we suspected would be acquired by an AI company for training data, and following it around the country to its final destination.

That final destination was an Amazon warehouse in Las Vegas, Nevada. Amazon employees who work at this location say all they do is receive massive shipments of printed books which they then cut the bindings off in order to scan the books more quickly. The printed book is destroyed in the process. The logo of the Amazon team that works at this warehouse, called VGT3, is a dinosaur, brandishing its teeth and with a book in its hands.

“Amazon purchases books through commercial channels to help develop and improve the products and services our customers use,” an Amazon spokesperson told me in a statement.

The world’s AI companies are constantly looking for, and spending extreme resources to locate, more material to train their AI models. With books, that sometimes means destroying them in the process, something that large parts of the public have spoken up against, and which we can now confirm Amazon is doing.

In July, I published a story about booksellers who reported a historical spike in sales starting in the past year. They suspected this spike in sales was due to AI companies acquiring any books they can in search of new training data. Printed books are valuable as training data because a lot of the text they contain is not readily available on the internet, which AI companies have already scraped. The data is also conveniently organized and, if the book was printed before 2022, is guaranteed to be free of AI-generated text, which can make any AI model that is trained on it worse via a recursive process called “model collapse.”

📖 Do you know work at a facility where you scan books? I would love to hear from you. Using a non-work device, you can message me securely on Signal at @emanuel.404. Otherwise, send me an email at emanuel@404media.co.

These booksellers suspected AI companies were behind these large bulk purchases because of the high number of books they were buying, the seemingly random choice of books, and the fact that these buyers, unlike libraries and universities, did not seem price sensitive at all. But booksellers couldn’t say for certain who was behind the large purchases because the marketplaces where they sell their books keep the buyers anonymous. When an order comes in, a bookseller ships the sold books to a warehouse operated by the marketplaces, where books are sorted and then sent to the buyer.

In July, one bookseller told me they received a very large order of around 1,000 books on Biblio, one of these marketplaces. The seller agreed to put an Apple AirTag provided by 404 Media in one of the books included in this order so we could see where the book was going. And by extension, which company, AI or otherwise, was behind this massive order. 404 Media granted the bookseller anonymity because they worried sharing this information would harm their business. Biblio did not respond to a request for comment.

Another article to provide some clarity on the types of books: …biggo.com/…/37a2899b-5f1a-4571-9bb5-1156c0ac4605

Honzuki no Gekokujou: Ryoushu no Youjo, episode 21

Related Communities

• !aobanime@bookwyr.me

Reminder: Please do not discuss plot points not yet seen or skipped in the show. Failing to follow the rules may result in a ban.

All discussions

This post was created by a bot. Message the mod team for feedback and comments. The original source code can be found on GitHub.

Matrix Metadata Problem: the server sees everything except your messages

People pick Matrix because of end-to-end encryption. But E2EE only protects message content — everything around it is plaintext.

The server sees:

• Your ID, IP, client type and version, device you use
• Your presence, typing status, read receipts
• Which rooms you join and when you join/leave
• Who else is in those rooms → your entire social graph
• Room names, topics, avatars (all plaintext state)
• For every message: who sent it, which room, a millisecond timestamp, type, size

And federation copies all of that to every other server whose users are in the room. Your “social footprint” isn’t one copy — it’s as many as there are servers.

Guess how I made the screenshot. Hint: this is tty.

Im going to be going on vacation and travel in November, and honestly, this is needed. I’m on the lookout for a secondary phone that i can degoogle real quick, put the neccessary apps that i need and just call it a day. I’m already prepping my family and stating that im not gonna bring my primary phone. I was wondering what do you guys use, if any of yall bring a secondary phone. (FYI: I am in the US and will be traveling internationally)

tldr; NL sold less to US but actually bought more from them, mainly fuel.

Hey everyone! I’m the developer of Kingdom Rising It’s a simple and fun strategy game where you build your base, upgrade your kingdom, train troops, and raid other players. Compete with friends and climb the leaderboard!

Would love for you to try it and tell me what you think

📱 https://apps.apple.com/app/id6796983706

any advice will help even generic stuff since other people also poballey have the same problem and would like some advice

do you ever feel like you have to choose between being ok or having friends and being private. like you know too much but you dont know what to do? have you been there? how do you start doing something and not stuck?

my friends dont use private things so i have to deal with it or be calld werid ( my freinds and adults think i just want privacy to play games when i shouldnt or watch videos that i should just bc im a yonger guy. if they dont acuse me of anything they just think im crazy or a bum who brings ppl down.

not to mention i feel stuck finding what to do and actalley doing ib. it feels easy and good to give up and ignore stuff like ts. lalala i cant hear you yk. its hrader to learn to bc im “spescial” and stuff. i worry that itll get very hard when i grow up too ((( have you been there?

cross-posted from: programming.dev/post/56270925

Please let me know if this should be posted somewhere else, and remove it if it does not go here.

Sometimes a camera just needs a sock.

The Flock Sock is a slip on cover designed to help protect a Flock camera from the harmful radiation of the sun. Your city spent their very own hard earned tax dollars scattering these cameras everywhere to keep you safe, all it cost you was a little privacy. The very least you can do is help protect these valuable civil servants from getting a bad sunburn.

Source: https://xcancel.com/tatami10jyo/status/2012068626778796409

Hi everyone,

Like many of you, I have developed severe “data-slurp fatigue.” But beyond privacy concerns, I’ve come to realize it’s a massive architectural inefficiency: we are wasting storage and risking security through the aggressively growing duplication of our home addresses and emails across millions of servers just to complete basic digital actions.

As a counterweight to this, I formulated an architectural concept called UYC: Unknow Your Customer — the structural antithesis to ubiquitous KYC.

The core premise of UYC is that systems must be built so they physically cannot hold, see, or remember unnecessary personal data. Instead of hoarding information, architecture should naturally leverage cryptography and pseudonymization in scenarios where collecting, storing, or processing personal data is entirely unnecessary.

As a solo creator, I built Factflew (factflew.io) specifically to test and prove that UYC is a viable foundation for a cleaner, independent web.

I’m bringing this to Lemmy to discuss the engineering reality of this approach:

• Is a UYC architecture viable at scale, or is the industry bound by too much engineering inertia when it comes to collecting data?
• How can we shift the developer mindset away from collecting everything “just in case”?

I am eager to hear your technical critique of the UYC principle and discuss whether we can build a leaner web. Cheers!

Kimi no Koto ga Dai Dai Dai Dai Daisuki na 100-nin no Kanojo 3rd Season, episode 10

Reminder: Please do not discuss plot points not yet seen or skipped in the show. Failing to follow the rules may result in a ban.

All discussions

This post was created by a bot. Message the mod team for feedback and comments. The original source code can be found on GitHub.

For our French community: Don’t miss the next /e/OS flash party in Lyon!

🧭 Lyon, at I-Factory at Campus de la Doua

⏰ November 21, 2026

👉 www.campus-du-libre.org

Sign up now: pretix.lyon-cvp.fr/LyonCafeViePrivee/CDL2026/

Come watch some classic shojo with other fun Fedi friends! We’re scheduled to start a new series on hopefully a new convenient day of the week.

Want to see if the group is right for you? At the same times on September 16th, we’ll be checking our some shojo OVAs. Come watch with us and see how much fun it can be!

Want to add this to your calendar? We got some .ics code below for you.

Link to post

Spotify Technology SA co-founder Martin Lorentzon said he would leave Sweden immediately if the country introduces new wealth taxes.

Lorentzon urged politicians to consider higher taxes on capital or dividends instead of a wealth tax, saying it would be “directly counterproductive”.

Lorentzon said the discussions about new taxes have already created uncertainty, with entrepreneurs asking if it is still OK to start a company in Sweden.

Paywall? www.removepaywall.com/search?url=https%3A%2F%2Fww…

Hundreds of social media posts by Prime Minister Rob Jetten and his party D66 were written by artificial intelligence, the Volkskrant discovered. Jetten’s AI-generated posts cover sensitive topics, including the war in Ukraine, the commemoration of the MH17 plane crash, and the government’s apology to the Mollucan people. Experts worry about the effect this will have on the Prime Minister’s credibility.

So, I’ve got a homelab thing going, with NextCloud and Immich and other things… using Docker. But the one thing I can’t seem to wrap my head around is how to get my calendar needs met behind the scenes. I’m hoping there’s some sort of calendar syncing server I can use.

The problem: I use NextCloud and another android app to use my calendar. Easy. Historically, my household has used a single Google Calendar shared amongst members.

The real problem: They won’t ditch Google. But I need a solution so that I can:

1. Sync their calendar events to my NextCloud calendar.
2. Sync my calendar events to their Google Calendar. I think that the issue here is, the share links that the calendars give, can only be used to list a new separate calendar on each of the respective platforms - not sync the events between the existing calendars…

I’m hoping that there’s maybe a syncing solution out there that can play middle man and sync the two calendars directly, rather than having separate calendars for the two environments?

Hello folks!

I would like to share an update about BentoPDF. It’s an open source privacy first PDF toolkit that runs in your browser.

1. You can now actually edit text inside PDFs

BentoPDF now allows you to click existing text, edit it, and have the text reflow while preserving the original fonts and styling.

It also includes:

• Bold, italic, underline, strikethrough, superscript, subscript, font family/size, colours, outlines, character spacing and line spacing
• Left, centre, right and justified alignment
• Bulleted and numbered lists with indentation
• RTL and LTR text alignment
• Find and replace across the document
• Edit images by rotating, flipping, duplicating, resizing or deleting them
• Object alignment, distribution, rotation, flipping, duplication

Please note this is a work in progress. You may encounter bugs, which you can report and I will look into.

2. Hyper Compress

Hyper is an open source PDF compression engine built to solve specific problems.

• Compression sometimes returns a file larger than the original. Hyper’s result is always binary: it either produces a smaller PDF or returns the original. This helps create predictable workflows.
• Hyper includes a true lossless mode. It preserves searchable text, document structure and PDF conformance, rather than rebuilding the document from scratch like Ghostscript.
• It runs everywhere: CLI, Node SDK, C API, self hosted service, and WebAssembly build.

Repository and benchmarks: github.com/alam00000/bentopdf-hyper-compress

I also compared it to Adobe’s compression API, but only managed around 100 PDFs where results were within 5%. Testing on a bigger corpus got expensive.

3. Kura

Kura is a PDF standards, conversion and preflight engine.

It supports:

• All 11 PDF/A conformance levels: PDF/A-1a, PDF/A-1b, PDF/A-2a, PDF/A-2b, PDF/A-2u, PDF/A-3a, PDF/A-3b, PDF/A-3u, PDF/A-4, PDF/A-4e and PDF/A-4f
• Accessibility: PDF/UA-1 and PDF/UA-2
• Print production: PDF/X-1a, PDF/X-3, PDF/X-4, PDF/X-4p, PDF/X-5g, PDF/X-5n and PDF/X-5pg
• Engineering and variable data printing: PDF/E-1 and PDF/VT
• E-invoices: Factur-X, ZUGFeRD, XRechnung and Order-X
• 396 bundled print-preflight profiles

It has been tested against several standards suites, including the veraPDF corpus, Isartor, BFO, Ghent Output Suite 5.0, the PDF/UA Reference Suite and Cal Poly’s PDF/VT suite.

Across 30,677 PDF conversions it had zero crashes and zero timeouts, with a 0.05 second median conversion time.

Like Hyper, it ships as a CLI, C library, npm package, Docker image and WebAssembly build.

Repository and benchmarks: github.com/alam00000/bentopdf-kura

The release also contains other improvements and bug fixes: github.com/alam00000/bentopdf/releases/…/v2.8.8

Thank you and have a great weekend!

Romance of the Stars

In commemoration of the release of “Puella Magi Madoka Magica the Movie -Walpurgisnacht: Rising-” in Japan today (28th Augest, 2026).

Edit: Mine’s a combination of Your Lie in April and Liz and the Blue Bird

hachyderm.io/@molly0xfff/117225270043571205