What is Hackers' Pub?

Hackers' Pub is a place for software engineers to share their knowledge and experience with each other. It's also an ActivityPub-enabled social network, so you can follow your favorite hackers in the fediverse and get their latest posts in your feed.

0
9
0
0

Large-scale online deanonymization with LLMs

https://arxiv.org/abs/2602.16800

It's over, anons.

arXiv logo

Large-scale online deanonymization with LLMs

We show that large language models can be used to perform at-scale deanonymization. With full Internet access, our agent can re-identify Hacker News users and Anthropic Interviewer participants at high precision, given pseudonymous online profiles and conversations alone, matching what would take hours for a dedicated human investigator. We then design attacks for the closed-world setting. Given two databases of pseudonymous individuals, each containing unstructured text written by or about that individual, we implement a scalable attack pipeline that uses LLMs to: (1) extract identity-relevant features, (2) search for candidate matches via semantic embeddings, and (3) reason over top candidates to verify matches and reduce false positives. Compared to classical deanonymization work (e.g., on the Netflix prize) that required structured data, our approach works directly on raw user content across arbitrary platforms. We construct three datasets with known ground-truth data to evaluate our attacks. The first links Hacker News to LinkedIn profiles, using cross-platform references that appear in the profiles. Our second dataset matches users across Reddit movie discussion communities; and the third splits a single user's Reddit history in time to create two pseudonymous profiles to be matched. In each setting, LLM-based methods substantially outperform classical baselines, achieving up to 68% recall at 90% precision compared to near 0% for the best non-LLM method. Our results show that the practical obscurity protecting pseudonymous users online no longer holds and that threat models for online privacy need to be reconsidered.

arxiv.org · arXiv.org

0
0
0
5
0
0
0
0

はじめての百合スタディーズ 3장은 BL과 백합의 비교 논의, 그리고 백합 내의 성애와 우정에 대한 논의가 있었다. 민감하면서도 중요한 주제가 많아서 앞으로도 몇 번 다시 읽어보면 좋을 부분이 많았다.

0

Google Mail / "GMail" has no fewer than nine settings for "Smart Features and Personalization", e.g., Gemini. I have these all turned off in order to prevent Gemini ingesting my emails. To do this, you now have to opt out of benign machine-learning or even pattern-matching-based features, like sorting promotional emails from regular ones or spellcheck.

Despite this, as of today, Google is offering a Gemini smart compose feature on all my emails, despite me having turned off smart compose.

"Get help writing email… Gemini can help you write a new email or refine one you've written"
0
0
0
0
1

My wife - a former elementary teacher - and her friend - an art historian - just launched their Kickstarter and I'm genuinely proud of them.

They made a children's art book where a single line runs through twelve real masterpieces - kids follow it and actually look at the art. A lovely way to build art appreciation without it feeling like a lesson.

Reposting here or on other sites appreciated:

kickstarter.com/projects/littl

0
4
0

フリマサイトで気になっていた服をやっぱり買おうと決めて、購入ボタンを押す前になんとなく商品写真を見返していたら、タグの洗濯マークがドライ洗い以外全て斜線だった。
諦めました。日常的に着る服が洗濯機使えないのは厳しい。残念。
だが買う前に気がついてよかった。

1
0
0
1

My wife - a former elementary teacher - and her friend - an art historian - just launched their Kickstarter and I'm genuinely proud of them.

They made a children's art book where a single line runs through twelve real masterpieces - kids follow it and actually look at the art. A lovely way to build art appreciation without it feeling like a lesson.

Reposting here or on other sites appreciated:

kickstarter.com/projects/littl

0
4
0

New video footage shows a US Tomahawk missile hitting an IRGC facility in Minab, Iran, on Feb 28, showing for the first time that the US struck the area. The footage also shows smoke already rising from the vicinity of the girls’ school, where 175 people were reportedly killed, including children.

The footage appears to contradict President Donald Trump’s claim it was an Iranian missile that hit the school on Feb. 28.
bellingcat.com/news/2026/03/08

0
14
0
1
0
0
0
0
0
0
1
0
0

クロードのチャットとクロードコードは連携しないらしい
チャットで外見のモックとmodelの設計までやってくれたんだけど、早く開発を進めろとして無茶苦茶急かしてくる
設計を慎重にやりたいけど「設計は俺がやる!」と言われてるみたいだ
ちゃんと言えばやらせてくれるんだろうけど

0
0
0

Do druhého roku vstoupila stávka popelářů v britském Birminghamu a řešení sporu o platy je stále v nedohlednu. Radnice vedená labouristy trvá na svých podmínkách a svoz odpadu se snaží zajistit alespoň částečně jinak.
ℹ️ Stranu ale problém připraví o 40 % příspěvků jednoho z největších odborářských svazů v zemi a mohl by jim i prohrát květnové komunální volby.

0
0
1

RE: social.bau-ha.us/@raganwald/11

Think this through.

Marco is walking around, allowing himself to be photographed, wearing shoes that are two sizes too big.

He can afford to buy those shoes in the correct size and save his boss and himself the embarrassment.

But Rubio suspects, fears, knows in his black tiny heart, that gave him shoes two sizes too big on purpose, in order publicly to humiliate him, in order to force him to broadcast his blind self-annihilating loyalty before all the world. It is literally The Emperor Has No Clothes, except with the humiliation transfered to the courtiers.

0
0
0
0
0
16
0
0
0
0

Google Mail / "GMail" has no fewer than nine settings for "Smart Features and Personalization", e.g., Gemini. I have these all turned off in order to prevent Gemini ingesting my emails. To do this, you now have to opt out of benign machine-learning or even pattern-matching-based features, like sorting promotional emails from regular ones or spellcheck.

Despite this, as of today, Google is offering a Gemini smart compose feature on all my emails, despite me having turned off smart compose.

"Get help writing email… Gemini can help you write a new email or refine one you've written"
0
0
0
0
0