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.
「2025年回顧」「AI検索は優越的地位の濫用？」「Sora 2よりヤバイ？ X（旧Twitter）の『画像を編集』機能」など、週刊出版ニュースまとめ＆コラム #696（2025年12月21日～2026年1月3日） https://hon.jp/news/1.0/0/58050
Forgot to note it back before the holidays, but the esq.social Mastodon instancejust had its third birthday. 🎉
In that time, the instance has processed 597 million messages.
I'm responsible for, like, half at most.
ベネズエラに関心も知識もない人たちが「ベネズエラ人は喜んでいる！」とやるのはお決まりだが、最近は「ベネズエラ人は喜んでいる！AIもそう言ってるぞ！」という、より愚かなパターンもあるのだな……。
トランプによるベネズエラの攻撃と大統領夫妻の拉致。「国際法よりも国益を重視」といった見出しで報道されていることも多いけれど、どちらかと言えばアメリカにとっての「国益」の内実が変化したと言うべきなのでは。
冷戦後、曲がりなりにも自由主義（リベラリズム）の価値観を資本主義的な自由経済秩序とセットにして国際社会に拡大し、そうして築き上げた国際秩序の盟主であることがアメリカの「国益」だったし、アフガンやイランへの侵攻や支配も、そうした建て前のもとでなされていた。
それに対して、そうしてアメリカが自国の利益のために作り上げた国際秩序は、むしろアメリカを弱らせていると主張して、トランプは大統領になった。MAGA派の多くは国境をまたいだ軍事侵攻を支持しないとされていたが、結局は西半球を帝国主義的に手中に収め、石油の利権を得るというむき出しの「利益」を隠しもせず、トランプはベネズエラに軍事攻撃をした。アメリカはずっと自国の「国益」のことしか考えていないし、無責任な暴力の行使によって、世界に戦争と混乱を引き起こし続けている。
I wrote Zig bindings to quickjs-ng with 96% API coverage (~240 exported C decls) with unit tests, examples, and doc strings on all functions in less than 6 total hours with AI assistance. I never want to hear that AI isn't faster ever again. https://github.com/mitchellh/zig-quickjs-ng
This isn't slop. I worked for those 6 hours.
I was reviewing everything it outputted, updating my AGENTS.md to course correct future work, ensuring the output was idiomatic Zig, writing my own tests on the side to verify its work (while it worked), and more. My work was split across ~40 separate Amp threads (not one mega session, which doesn't work anyways unless you're orchestrating).
I have a ton of experience writing bindings to libraries for various languages, especially Zig. I have never achieved this much coverage in so little time with such high quality (e.g. test coverage). My usual approach is to get bind just-enough of the surface area to do my actual work and move on. This time I thought I'd draw the whole owl, because it's a new world. And I'm very happy with the result.
Anyone with experience writing bindings knows that you do some small surface area, then the rest of the coverage is annoying repetition. That's why I usually stopped. Well, LLMs/agents are really, really good at annoying repetition and pattern matching. So going from 5% API coverage to 95% is... cake.
There is probably some corners that are kind of nasty still, but I've been re-reviewing every line of code manually and there is nothing major. Definitely some areas that can just use a nicer Zig interfaces over the C API, but that's about it.
I plan on writing a longer form blog showcasing my threads, but you can at least see the final AGENTS.md I produced in the linked repo.
I will repeat that I was not sitting back at all during those 6 hours. While agents were working, I was working, just on separate -- but related -- tasks. I know for a fact that I could not have completed this amount of work in 6 hours fully manually (based on the experience that I've written something like 30+ bindings to C libraries in the past decade, probably more).
I wrote Zig bindings to quickjs-ng with 96% API coverage (~240 exported C decls) with unit tests, examples, and doc strings on all functions in less than 6 total hours with AI assistance. I never want to hear that AI isn't faster ever again. https://github.com/mitchellh/zig-quickjs-ng
This isn't slop. I worked for those 6 hours.
I was reviewing everything it outputted, updating my AGENTS.md to course correct future work, ensuring the output was idiomatic Zig, writing my own tests on the side to verify its work (while it worked), and more. My work was split across ~40 separate Amp threads (not one mega session, which doesn't work anyways unless you're orchestrating).
I have a ton of experience writing bindings to libraries for various languages, especially Zig. I have never achieved this much coverage in so little time with such high quality (e.g. test coverage). My usual approach is to get bind just-enough of the surface area to do my actual work and move on. This time I thought I'd draw the whole owl, because it's a new world. And I'm very happy with the result.
Anyone with experience writing bindings knows that you do some small surface area, then the rest of the coverage is annoying repetition. That's why I usually stopped. Well, LLMs/agents are really, really good at annoying repetition and pattern matching. So going from 5% API coverage to 95% is... cake.
There is probably some corners that are kind of nasty still, but I've been re-reviewing every line of code manually and there is nothing major. Definitely some areas that can just use a nicer Zig interfaces over the C API, but that's about it.
I plan on writing a longer form blog showcasing my threads, but you can at least see the final AGENTS.md I produced in the linked repo.
“ #rockinonsonic #ロキソニ #KNEECAP” / “KNEECAP Live at rockin'on sonic in Japan” (1 user) https://www.youtube.com/watch?v=d7q5HELztVM
During the holidays I've been doing Duolingo since I don't have graduate school reading and homework to do.
I've been competing with a stranger in the league system they have. We were running neck and neck for 1st and 2nd place, about 2000 points ahead of the next player. We've traded the top spot 3 or 4 times over the last 48 hours.
She must have spent hours studying today, because she's now 1000 points ahead of me. Gotta respect the hustle.
I wrote Zig bindings to quickjs-ng with 96% API coverage (~240 exported C decls) with unit tests, examples, and doc strings on all functions in less than 6 total hours with AI assistance. I never want to hear that AI isn't faster ever again. https://github.com/mitchellh/zig-quickjs-ng
This isn't slop. I worked for those 6 hours.
I was reviewing everything it outputted, updating my AGENTS.md to course correct future work, ensuring the output was idiomatic Zig, writing my own tests on the side to verify its work (while it worked), and more. My work was split across ~40 separate Amp threads (not one mega session, which doesn't work anyways unless you're orchestrating).
I have a ton of experience writing bindings to libraries for various languages, especially Zig. I have never achieved this much coverage in so little time with such high quality (e.g. test coverage). My usual approach is to get bind just-enough of the surface area to do my actual work and move on. This time I thought I'd draw the whole owl, because it's a new world. And I'm very happy with the result.
Anyone with experience writing bindings knows that you do some small surface area, then the rest of the coverage is annoying repetition. That's why I usually stopped. Well, LLMs/agents are really, really good at annoying repetition and pattern matching. So going from 5% API coverage to 95% is... cake.
There is probably some corners that are kind of nasty still, but I've been re-reviewing every line of code manually and there is nothing major. Definitely some areas that can just use a nicer Zig interfaces over the C API, but that's about it.
I plan on writing a longer form blog showcasing my threads, but you can at least see the final AGENTS.md I produced in the linked repo.
I will repeat that I was not sitting back at all during those 6 hours. While agents were working, I was working, just on separate -- but related -- tasks. I know for a fact that I could not have completed this amount of work in 6 hours fully manually (based on the experience that I've written something like 30+ bindings to C libraries in the past decade, probably more).
Where did the
I wrote Zig bindings to quickjs-ng with 96% API coverage (~240 exported C decls) with unit tests, examples, and doc strings on all functions in less than 6 total hours with AI assistance. I never want to hear that AI isn't faster ever again. https://github.com/mitchellh/zig-quickjs-ng
This isn't slop. I worked for those 6 hours.
I was reviewing everything it outputted, updating my AGENTS.md to course correct future work, ensuring the output was idiomatic Zig, writing my own tests on the side to verify its work (while it worked), and more. My work was split across ~40 separate Amp threads (not one mega session, which doesn't work anyways unless you're orchestrating).
I have a ton of experience writing bindings to libraries for various languages, especially Zig. I have never achieved this much coverage in so little time with such high quality (e.g. test coverage). My usual approach is to get bind just-enough of the surface area to do my actual work and move on. This time I thought I'd draw the whole owl, because it's a new world. And I'm very happy with the result.
Anyone with experience writing bindings knows that you do some small surface area, then the rest of the coverage is annoying repetition. That's why I usually stopped. Well, LLMs/agents are really, really good at annoying repetition and pattern matching. So going from 5% API coverage to 95% is... cake.
There is probably some corners that are kind of nasty still, but I've been re-reviewing every line of code manually and there is nothing major. Definitely some areas that can just use a nicer Zig interfaces over the C API, but that's about it.
I plan on writing a longer form blog showcasing my threads, but you can at least see the final AGENTS.md I produced in the linked repo.
( º дº)<ｷｴｪｪｪｴｴｪｪｪ
Paint me like one of your French cats
完全に悪の組織だろう。もはや冗談では済まされなくなった。アメリカのお気持ちひとつで特殊部隊が襲撃してくるぞ。
＞3日、米国がベネズエラを攻撃した後、地図上のグリーンランドを星条旗の柄に色づけし、「もうすぐ」との文言を添えた画像をX（旧ツイッター）に投稿した。
米高官の妻、グリーンランドに星条旗の色つけ投稿 現地反発「無礼」
https://digital.asahi.com/articles/ASV146WWNV14UHBI001M.html?iref=comtop_7_05
『「トランプは米国がベネズエラを運営すると言った。どんな法的権限に基づいて？」
ルビオ国務長官「我々はベネズエラの人々により良い未来を望んでいる」
「もう一度尋ねる。法的権限は何？」
ルビオ「私たちは裁判所の命令を持っています」』
どんな裁判所が「植民地支配」を命令するんだ？
석유 회사 기업들의 이득을 위한 유혈 사태라고 보는것이 좀더 맞는것 같고요 근데 석유 회사들 이득을 왜 잘 챙겨주냐면 여러번 적었다시피 달러의 가치 유지와 연관이 되어있고... 그외에도 석유외에도 희토류와 금속 자원들을 미국과 관련 회사들이 컨트롤하고 싶으니까 그런것입죠 그리고 베네주엘라가 러시아 무기로 무장하고 (과거 쿠바처럼) 러시아 텃밭이 되가니까 이걸 봐줄수 없었다는 이야기도 있더라고요 여러가지가 겹친것 허허...
RE: https://bsky.app/profile/did:plc:7reki7xuobtaq6iuqquznqby/post/3mblldrhzuc2o
@buildItN0wLars https://vps-stuff.com new? Or too old?
Stop Forwarding Errors, Start Designing Them
Link: https://fast.github.io/blog/stop-forwarding-errors-start-designing-them/
Discussion: https://news.ycombinator.com/item?id=46491051
ございます
2,3時間で目が覚めるのを繰り返してた……
The two hardest problems in Computer Science are:
1. Getting up in the morning
2. Going to bed at a reasonable time
I wrote Zig bindings to quickjs-ng with 96% API coverage (~240 exported C decls) with unit tests, examples, and doc strings on all functions in less than 6 total hours with AI assistance. I never want to hear that AI isn't faster ever again. https://github.com/mitchellh/zig-quickjs-ng
This isn't slop. I worked for those 6 hours.
I was reviewing everything it outputted, updating my AGENTS.md to course correct future work, ensuring the output was idiomatic Zig, writing my own tests on the side to verify its work (while it worked), and more. My work was split across ~40 separate Amp threads (not one mega session, which doesn't work anyways unless you're orchestrating).
I have a ton of experience writing bindings to libraries for various languages, especially Zig. I have never achieved this much coverage in so little time with such high quality (e.g. test coverage). My usual approach is to get bind just-enough of the surface area to do my actual work and move on. This time I thought I'd draw the whole owl, because it's a new world. And I'm very happy with the result.
Anyone with experience writing bindings knows that you do some small surface area, then the rest of the coverage is annoying repetition. That's why I usually stopped. Well, LLMs/agents are really, really good at annoying repetition and pattern matching. So going from 5% API coverage to 95% is... cake.
There is probably some corners that are kind of nasty still, but I've been re-reviewing every line of code manually and there is nothing major. Definitely some areas that can just use a nicer Zig interfaces over the C API, but that's about it.
I plan on writing a longer form blog showcasing my threads, but you can at least see the final AGENTS.md I produced in the linked repo.
7만여 명 숨진 땅, 구호도 막은 이스라엘
h21.hani.co.kr/arti/world/w...
"외국 거주 유대인 지원과 반유대주의 대응 업무를 총괄하는 이스라엘 디아스포라부는 2025년 12월30일 자료를 내어 “가자지구에서 활동하는 37개 비정부기구(NGO)가 강화된 등록 요건을 충족하지 못했다”며 “1월1일부터 이들 단체는 가자지구에서 활동을 중단해야 한다”고 밝혔다. 액션에이드·국제구조위원회·국경없는의사회 등 지구촌을 대표하는 구호단체 상당수가 활동 중단 대상에 포함됐다."
7만여 명 숨진 땅, 구호도 막은 이스라엘
New blog post, first in a long time.
Why I chose to go towards Machine Learning research
https://nilio.ca/post?title=Why%20I%20chose%20to%20go%20towards%20Machine%20Learning%20research
ʕ•͡ω•ʔ 奇跡的に座れた
Members of the community formerly known as "Empty Spaces" read this thread
https://bsky.app/profile/did:plc:tjut7ymmni3tu4h777yj5xqj/post/3mbcpse5rr22e
followed by this thread
https://bsky.app/profile/did:plc:tjut7ymmni3tu4h777yj5xqj/post/3mbct6xufrs2e
EDIT: Confusing/enlightening bonus thread
https://bsky.app/profile/jana-aych-ess.bsky.social/post/3ma4vfitpbk27
お仕事はじめてきまーす！
it's important to craft things with care in subtle ways to reward people who are attentive
하...일어나야지..
North Dakota law lists fake critical minerals based on coal lawyers' names
Link: https://bismarcktribune.com/news/local/government-politics/article_515812a0-d29a-4161-91f1-3e53003e2911.html
Discussion: https://news.ycombinator.com/item?id=46492161
Eurostar AI vulnerability: when a chatbot goes off the rails
Link: https://www.pentestpartners.com/security-blog/eurostar-ai-vulnerability-when-a-chatbot-goes-off-the-rails/
Discussion: https://news.ycombinator.com/item?id=46492063
쿠팡 물류센터, 방한복 없는데 냉동창고 '근무 지시'
n.news.naver.com/article/437/...
방한 물품 대여 창구가 이미 운영을 마친 시간대였기 때문입니다.
하지만 센터 측이 지급해준 건 핫팩 3개뿐이었습니다.
쿠팡 측은 "안전 교육을 실시했다"는 입장입니다.
[단독] 쿠팡 물류센터, 방한복 없는데 냉동창고 '근무...
돈은 내주머
NIRO
미치겠네 ㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋㅋ
트럼프, 베네수 부통령에 "옳은 일 안하면 더 큰 대가"
www.yonhapnewstv.co.kr/news/MYH2026...
"아마도 마두로보다 더 큰 대가를 치를 것"
쓰레기
트럼프, 베네수 부통령에 "옳은 일 안하면 더 큰 대가...
태양 만세!! \[T]/
주님이신 날아다니는 스파게티 괴물 님🍝께서는 악인에게나 선인에게나 당신의 해가 떠오르게 하신다.
2026-01-05T07:37:02+09:00
-40° 🥶
@catsaladCat 🐈🥗 (D.Burch) 
The one place where both commonly-used measures agree that it's really fucking cold.
(Kelvin looks on in distain).
The two hardest problems in Computer Science are:
1. Getting up in the morning
2. Going to bed at a reasonable time
애애앵
산업혁명기에 괜히 빈곤층과 열악한 노동에 종사하는 사람들이 늘어나는게 아닙니다. 기존에 준비하던 직업 혹은 종사하던 직업이 사라지면.. 새롭게 변한 산업에서 처음부터 적응하고 배워야하니 그렇게 내몰리는거죠.
Was Bekenner*innenschreiben angeht, ich habe vor über 30 Jahren ein ganzes Seminar zu forensischer Linguistik besucht und irgendwo steht da auch noch die Literatur dazu hier rum.
Ich sags mal so, die Analysen des BKA waren nicht der Rede wert und haltlos.
Und heutzutage kann ich mir noch mehr als damals vorstellen, dass da sich welche als "links" bekennen, die ganz andere Ziele verfolgen.
Stop Forwarding Errors, Start Designing Them https://lobste.rs/s/eaxaac #rust
https://fast.github.io/blog/stop-forwarding-errors-start-designing-them/
트럼프, 베네수 부통령에 "옳은 일 안하면 더 큰 대가"
www.yonhapnewstv.co.kr/news/MYH2026...
"아마도 마두로보다 더 큰 대가를 치를 것"
쓰레기
트럼프, 베네수 부통령에 "옳은 일 안하면 더 큰 대가...
おはようやで！
오늘 아침은 저속노화맨 사진 붙였다가 땡처리중인 레토르트 죽. 안 달아서 좋네.
SparklingOutlaw🍉
洪 民憙 (Hong Minhee)
洪 民憙 (Hong Minhee) 
Glyph
＠3/23(日)
東京オフ主催 
DIGIWB/화이트베어 
힐씨
pkg update
Miguel de Icaza ᯅ🍉
