lukeigel

I let an agent grind on this since March. I got scooped by the teenager slqnt on single player HL2, so here's multiplayer. I more or less let it run in loops for months!

Pangram reports as 100% AI generated. Makes sense for a README, but a tad bit funny given that their students must hand-write code

How does this feel different from Vercel AI SDK's newer local debugger? Richer in general?

Matheus has been an amazing addition to the Jmail team. If you're interested in contributing, we open sourced all the Reducto-parsed data at https://data.jmail.world

David Oks at it again.

Bring your own key and use Claude. We found that it's most willing to run deep research queries here.

Correct. https://jmail.world/about explains it all.

https://jmail.world/thread/07ff1467c0f2bb976664ecafc5829aa4?...

Many Yahoo emails do show you the original source, and the original is just an EML file. These were files directly exported from Epstein's Yahoo account. Bloomberg used these EMLs to confirm that the Yahoo emails are real (https://www.bloomberg.com/news/newsletters/2025-09-12/epstei...).

We don't tamper with these EMLs, so we currently don't make the EML accessible if the team had to redact any contents of that email.

See this for example.

https://jmail.world/thread/97d4a52d1df3948368770068262d2aab?...

We can fix examples where there are no redactions yet no EML download is available

Also, that says "Verified by Drop Site News", not Sponsored by. That's because Drop Site redacted these real Yahoo emails and gave them to Jmail. The original Yahoo dataset, which the DOJ and House Oversight Committee did not release, is stewarded by DDoSecrets (https://ddosecrets.org/article/epstein-emails).

This Yahoo dataset, which we helped release after launching the first Jmail, also proved Epstein's connection to Iran-Contra (!!). Now immortalized on Wikipedia (https://en.wikipedia.org/wiki/Jeffrey_Epstein#Financial_trou...)

Yeah, those are mailing lists that people signed him up for after he died, including with a joke name.

He was a very famous figure in August 2019, and normal people spammed his inbox with emails like the below

https://jmail.world/thread/4accfb5f3ed84656e9762740081a4579?...

These are all real emails! We can do a better job making this clear to the user.

Jmail maintainer and co-creator here. Very excited to see that someone finally made Jemini good!

Our development process has been interesting. Although just Riley and I first made Jmail, it's been really gratifying to see companies, journalists, and fellow developers like Diego rise to the occasion to make this entire suite of apps as high quality and extensive as possible.

Very proud of David Oks.

Doing gods work! I saw it yesterday on Twitter and now I've been relying on it all of this morning.

Sorry about that. That Attachments tab shows you literal attachments found in the .eml files from the YAHOO dataset. We have released very few .emls that have attachments though. Expect more and more over time.

Releasing YAHOO responsibly has been time consuming, and we're relying on Drop Site to tackle the redactions. See this post for context.

https://news.ycombinator.com/item?id=46347272

Yes we'd love to. Letting the dust settle first

Yes! We used our friends at Reducto (https://reducto.ai/) for all document extraction and parsing (one of the best companies I've ever referred to YC ;) )

We did an initial parsing pass of all four DOJ document batches on Friday. This takes a raw PDF and returns chunks containing typed blocks—each with a type (Title, Text, Figure, etc.), bounding boxes, content, and confidence scores. For PDFs that were just scans of photographs (which was like 90% of new content in Friday's release), it gave in depth descriptions of those! You can type search terms like "door" at https://www.jmail.world/photos to see what I mean.

For apps like Jmail and JFlights we use their structured extraction endpoint instead—you define a schema (e.g. {from, to, subject, date, body} for emails or {departure_airport, arrival_airport, passengers[], date} for flights) and it pulls those fields directly into JSON.

The JFlights example served as the best ad for Reducto and how doc parsing technology can speed up hours of journalistic investigations like this.

See for yourself. Given this document

https://www.jmail.world/drive/HOUSE_OVERSIGHT_002031

It inferred and enriched multiple flight cards on JFlights (https://www.jmail.world/flights). I was really shook when I first saw this.

Another person made an oddly beautiful ASCII ui for the text messages. All seem to be from HOUSE_OVERSIGHT (we have those plus DOJ, YAHOO. No dedicated text UI from us)

https://michelcrypt4d4mus.github.io/epstein_text_messages/

He also shouted us out last month which was very kind of him

Re: the DOJ emails prefixed with "EFTA", I have no idea how over-redacted they are. They definitely seem dubious though.

Re: the DDoSecrets emails though (YAHOO dataset), I have more to share.

Drop Site News agreed to give us access to the Yahoo dataset discovered by DDoSecrets, but on the condition that we help redact it. It's a completely unfiltered dataset. It's literally just .eml files for jeeprojects@yahoo.com. It includes many attached documents. There is no illegal imagery, but it has photos of Epstein's extended family (nephews, nieces, etc) and headshots of many models that Epstein's executive assistant would send to him. I was quite shocked that this thing existed.

We built some internal redaction tools that the Drop Site team is now using to comb through all of this. We've released 5 batches of the Yahoo mail now, with the 1k+ Amazon receipts being the most recent.

A few thoughts on how we do redaction are here: https://www.jmail.world/about.

Unlike the DOJ, we've tried to minimize the ambiguity about what was redacted.

For example: all redacted images are replaced with a Gemini-generated description of that photograph.

Another example: we are aggressively redacting email addresses and phone numbers of normal people to avoid spamming them. Perhaps others would leave it all in, but Riley and I don't want to be responsible for these people's lives getting disrupted by this entire saga. For example, we redacted this guy's email but not his name: https://www.jmail.world/thread/4accfb5f3ed84656e9762740081a4...

Riley and I were not expecting this type of scope when we first dropped Jmail. Jmail is an interesting side project for us, and this new dataset requires full-time attention. Thankfully we have help though. We're happy to take on this responsibility given how helpful, thoughtful and careful both the Drop Site and DDoSecrets team has been here.

Thank you! I tried to optimize scrolling through the photos the most. Avoided any virtualization library, used Cloudflare for image transformations and for caching, etc. The masonry layout and polish you see here is all https://news.ycombinator.com/user?id=walz though

Specifically at https://www.jmail.world/photos

Tons of this week's DOJ drop is old photos. These photos of his Upper East Side mansion and his Island estate are well known

We found that Volume 2 and Volume 4 had the most never-before-seen stuff.

https://www.justice.gov/epstein/doj-disclosures/data-set-2-f... https://www.justice.gov/epstein/doj-disclosures/data-set-4-f...

Also, this morning they quietly released volumes 5-7. Will have to find out how much of this is new.

Yes, shoutout to our friend Aidan Dunlap for making an entire interface to see those Amazon orders! It's at https://www.jmail.world/jamazon

Jmail is the only place to see those Amazon order emails by the way! Those are from his Yahoo, which Bloomberg announced in September but Drop Site News actually let us release this month. It all came from https://ddosecrets.com/article/epstein-emails (redactions of the full dataset still taking place)

also, shoutout the Jason Liu (https://news.ycombinator.com/user?id=jxnlco) for discovering that one. His turbopuffer-based version of Jemini is coming soon!

Thank you!

The original site was on Railway and written in Pug! It crashed after Riley's tweet first went viral, then Riley did the heroic work of caching it all with Cloudflare after waking up to the site being down. After millions of unique visitors we racked up about $10 in costs.

This time we switched to Next.js 16 + Vercel, used Cloudflare R2 for asset hosting, and used Neon as the db. R2 has free egress, and Vercel + Next is cheap if cached correctly.

A special someone at Vercel gave us some tips on caching this one earlier today. We started by just using unstable_cache all over the place, and now we're migrating to ISR + full static pre-generation of as many pages as we can via generateStaticParams.

Thank you! The night before launching the original Jmail we had this hunch that jump-straight-to-end and "random page" would make the site work for people.

We found that Codex indeed refuses but Claude + Gemini are willing to RAG it

We have three datasets in Jmail now:

1. DOJ (The White House's docs that they were required by law to drop yesterday plus many court documents, videos, and other docs from many news cycles this year)

2. HOUSE_OVERSIGHT (the House Oversight Committee's releases. giant November drop that led to the original Jmail, then some photo drops this month)

3. Yahoo emails (originally sourced by DDoSecrets, then provided to us, redacted and verified by Drop Site News)

There is so much material in HOUSE_OVERSIGHT that never appears in DOJ, and vice versa. And then the Yahoo drop reveals even more new material. It feels like three odd slices of a giant dataset that keeps getting released.

re: people's complaints about yesterday's release having way too many redactions, I have no idea how much they over-redacted. I hear that they will release even more quite soon though.

In which app? Photos, email, etc? Tested across each.

Thanks! And it's a lot of info, yeah. ~90% of new data in yesterday's drop was photographs, which they redacted for us.

The House Oversight Committee's giant drop in November had tons of data we still didn't take advantage of even after doing the original Jmail, like flight logs.

For the Yahoo release, which is still ongoing, the folks at Drop Site News (see https://www.jmail.world/about) are handling the manual redaction which has been very time consuming, even with tons of AI to help in the background.

Last night we made a ton of new apps and we added "the Epstein" files which DOJ dropped yesterday.

Also since that post we worked with Drop Site News + DDoSecrets to post new Yahoo emails that no one has let the public see yet.

Yes, thanks Diego! Really excited about this.

I'm one of the co-creators of Jmail alongside Riley Walz. We launched a Gmail-like view of Epstein's inbox last month. It got millions of page views, tons of really amazing requests to collaborate on making more related data accessible, and even new Yahoo emails that no one else has allowed the public to see.

Yesterday's DOJ drop resulted in this very spontaneous rag-tag team of friends coming to my place in SF and each making their own app in the "Jmail" Suite. Riley and I are pretty shocked by how versatile this parody style is for visualizing Epstein's 20 year digital footprint.

It's been a ton of fun and we're working hard to polish each view here.