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.
구트현X(-_-)가 그나마 있던 영어 고객지원 계정을 삭제하고, 도움말 센터와 유료이용자 지원 계정만 남길거라고. 물론 욕을 퍼먹고 있음.
arstechnica.com/tech-policy/...
외국 유저들은 X고객 센터에서 답변을 받기까지 6개월 이상 기다려야 했고, 계정 정지 후에도 계속 요금이 청구되었다고 분노.
Disgruntled users roast X for ...
よつ葉乳業応援ついったらんど投稿の人を見に行くと、けっこうな確率で日本大好き排外主義自然派なんだけどがなんだあれ
よつ葉乳業、異物混入でバター回収しててそれ自体はたまにはあることだしよつ葉もきちんと対応してるようなので良いんだが、ついったらんどで「小林製薬・すき屋と同じ状況だから応援しなきゃ」って言われてるらしいのなんだこれ
Supposedly OpenAI is creating a Twitter-like social media product.
My thesis that commercial social media platforms and AI platforms are merging has another data point: now it's three of them: OpenAI, Twitter/X/Grok and what Meta is doing.
「正義の反対は別の正義」の出典はどこか知らないけど、この言説を広めたのはドラえもんだろう。いい大人がいつまでもドラえもんで得た知識のまま学びをやめてしまうと言うこの国の学問軽視と幼稚さも表していると思うよ。ドラえも〜ん、いい大人がいつまでも「正義の反対は別の正義」と言ってるよ〜。
＞日本では『正義』『公正』を個人の努力や気持ちの問題に帰着させる傾向が強いので、『正義』の反対は悪ではなく『別の正義』みたいな屁理屈（へりくつ）が横行しやすい。
「正義」の反対は「別の正義」なのか 「屁理屈」の横行に抗う哲学者
https://digital.asahi.com/articles/AST4J1651T4JUPQJ002M.html?iref=comtop_7_06
My definition of a successful day in #DevOps is when nothing explodes.
CVE-2022-38392이 여러모로 흥미로웠었어요.
구형 노트북 하드의 공진주파수 문제가 취약점이라니..
So the solution to AI scams is... more AI. We have all seen the increase in AI generated scams. The solution to that is more AI? Maybe the reason more ads are being blocked is that there is actually just a lot more scams due to AI making it easier to make?
I still believe that the move from contextual ads to surveillance based ads was and is a massive mistake and the way to fix that is to ban surveillance based ads.
Contextual ads were also a lot more manual, most of the time, meaning scams were harder. Time to learn from mistakes and get rid of surveillance based ads?
トランプはアメリカのもっとも競争力のある産業輸出を殺そうとしている。教育だ。
という記事。確かにね。
https://www.washingtonpost.com/opinions/2025/04/15/trump-higher-education-colleges-trade-war/
(출근전철에서)
바퀴를 새로 만들지 마시오 – 바쁜 현대인을 위한 5분 판타지 – 브릿G
https://britg.kr/novel-group/novel-post/?np_id=539657&novel_post_id=208801
회귀, 영원회귀 – 바쁜 현대인을 위한 5분 판타지 – 브릿G
https://britg.kr/novel-group/novel-post/?np_id=539650&novel_post_id=208801
오늘은 짧은 이야기 두편입니다. 오늘도 좋은 하루 되세요. :)
(All) Databases Are Just Files. Postgres Too https://lobste.rs/s/bezwfp #databases
http://tselai.com/all-databases-are-just-files
マスク外して出社してる。目が少しアレだが、大丈夫そうだ！！！
X, '지원' 계정을 폐쇄하고 사용자들에게 고객 센터 또는 유료 구독자를 위한 프리미엄 지원 계정에서 도움을 받도록 안내할 것이라고발표
다만 이런 발표는 사용자들로부터 부정적인 반응만 나왔는데 지금도 엑스의 지원은 형편없고 유료 구독자를 위한 프리미엄 지원 계정 또한 응답이 매우 느리고 고객센터 웹 페이지로 보내는 조치만 취해서 별 도움은 되지 못한다고 지적
RE: https://bsky.app/profile/did:plc:wld6fad6xsm4tz4kfkoikun2/post/3lmxgwo2j2q2n
[!긴급 공유 및 연대 요청!]
미아리 재개발 강제 집행에 항의하기 위해서 성북구청 앞에 천막 농성장을 설치했으나 경찰과 대치 상황에 있습니다.
시간이 되시는 분은 지금 즉시 성북구청으로 와주시길 바랍니다.
📍성북구청 앞
교통: 성신여대입구역(4호･우이신설선), 보문역(6호선･우이신설선)
レスラーが引退する、ということは惜しむべきことだけど、レスラーが退団することはまたどこかで出会える可能性があるのだよなぁ
昨日の午前中、多分mstdn.jpのサーバが落ちてたかなんかでマストドンにアクセスできなくて、ちょっと精神的に動揺した。このアカウントは自分の精神を保つ命綱のようなもので、ここに思ったことを吐き出せなくなるとちょっと困ったことになる。万が一のことを考えると他のどこかに避難所的なものを作っておいた方がいいのかもしれないけど、Xは治安が悪いのでもう半年以上アカウントを使ってないし、Blueskyは推し活アカウントとして運用してるんで、こういう場合どうしたらいいんだろう。
今日はきんよう……え？まだ木曜日!?!?
어떤 여자한테 정자기증으로 아이 갖기 권유 했는데 싫다는 대답 들었을 때 할 수 있는 행동 1. 수긍한다. 2. 아쉬워하며 다시 기회가 있길 바란다 3. 그녀와 연락한 수단을 타인도 이용 할 수 없도록 DM이란 기능을 제거한다 여기서 3번 고르는 심리가 이해가 안됨 진심 시헤백남재벌의 심리는 도통 ㅋㅋㅋㅋㅋㅋㅋㅋ
そういえば今月はじめの中須氏は生成AIによるものだったのか人の手によるものだったのか。
"Please consult with staff if you plan to shoplift"
Lower left: "I'm shoplifting"
Lower right: "I'm calling the cops"
didn’t take much for these guys to go from “big government is bad” to “what’s the big deal if you’re abducted by the government and trafficked to Mongolia”
著作権という法的な権利でいえば当たり前なんだよ。人間がジブリ風味に描いても、そのものズバリではない限りは権利侵害ではない。そりゃそうだろ。
流行の「ジブリ風」画像生成 文科省の見解「作風の類似のみなら著作権侵害に当たらない」 https://www.sankei.com/article/20250416-OELPYZPOSJE2JE672JBXJECDQA/
"We believe the Trump administration violated the First Amendment by attempting to shut down our operations simply because they disagreed with our independent journalism. They saw our reporting as not aligned with their political agenda." - VOA's Patsy Widakuswara, lead plaintiff in one of the lawsuits filed against USAGM's Kari Lake. https://foreignpress.org/foreign-correspondents-interviews/patsy-widakuswara-and-the-struggle-to-save-voa-trump-wants-to-dismantle-us
You Don’t Get to Cancel the Constitution: Trump, Musk, and the Unholy War on the First Amendment
Trump tried to muzzle journalists. Musk muzzled satire. Both slapped the First Amendment in the face while preaching patriotism. This is not about politics. This is about the Constitution, and we h…
jtwb768.wordpress.com · The Babblings of JT
Link author: 
The Babblings of JT@jtwb768.wordpress.com@jtwb768.wordpress.com
なんだかよくわからない。後でまた読む
CVE Program Funding Reinstated—What It Means And What To Do Next
https://www.forbes.com/sites/kateoflahertyuk/2025/04/16/cve-program-funding-cut-what-it-means-and-what-to-do-next/
ざっくり読みなおした
- CISAがMITREとの契約を11ヶ月延長した
- MITREは去年からCVE Foundationを準備していた
mstdn.jaws-ug.okinawa のサービス停止をGWごろに行う予定です。
8年やってきたけどメンテナンスコストが捻出できずに申し訳ない。
自前のサービスで自分からShutdownするのは初めての経験・・・
この8年近く運用してきて色々と勉強になった。
https://www.slideshare.net/slideshow/mastdonawsclosing/75622684?from_search=3
ネタで始めたmastodonが8年も続いたことがすごい事と自画自賛しておく。
ありがとうございました。
晴れてるのに要メンテごめんて $
PHTO 162153Z 14013KT 10SM BKN080 28/16 A3000 RMK AO2 SLP157 T02780161 $
#METAR_PHTO #metarbot
飲んだよー。一般的なホワイトビールの甘さはほとんど感じずずっきりと飲める感じですまるで水のように。そう、これはチェイサーに最適なビールの新定番です。#beer
I’m renewing my request for a good lawyer to handle writing terms of service for my tech/web business. If you’ve used a lawyer for this in the past, I’d appreciate a referral. Remote is fine as long as they are U.S. based. #IndieDev #Technology #law
日本酒風呂
ハイボール風呂
X-CHAT이라는 별도 앱으로 (아마 카톡이나 페북 메신저같은)기존 트위터 DM 기능이 분리된다고 합니다. 관련 대화와 공지가 어디서 진행되는지를 못 찾겠어요, 적어도 공식 계정은 아니네요...
RE: https://bsky.app/profile/did:plc:a6qvfkbrohedqy3dt6k5mdv6/post/3lmxlhwvnoc2y
閣僚どうしの会談だと思ってたら大統領出てくるのどうなんだ
So, despite all our problems in the OSS community, I understand why people put code on forges (github, etc) and have been doing so for years. I understand the collaborative nature of development and the varied array of motivations one might have to create software either in public, or collaboratively with the public.
I … cannot figure out why anyone is doing this with "AI" models. Why do people upload weights to Ollama or Huggingface? What is (are) the incentive structure(s) there?
맙소사. 찾아봤는데.. 암호화폐쪽 저널리스트이며 유명 인플루언서이기도 한 티파니 퐁이란 사람한테 일론머스크가 팔로하고 멘션 나누다가 (이 과정에서 티파니퐁은 팔로어가 급증, X 수익이 2주만에 21000달러 발생) DM으로 "내 아이를 낳아줘" 이랬다고 함. 티파니 퐁은 거절하면 자기 수입이 줄어들까봐 걱정했고, 실제로 이 사실이 알려지면서 팔로는 취소되고 유료 계정 수입도 줄어들었다고.
www.independent.co.uk/news/world/a...
RE: https://bsky.app/profile/did:plc:de27rm6eyuf5ez6gmvjdmilq/post/3lmxmse4bx22q
Musk asked cypto influencer Ti...
don't call it that
這是騷擾私訊嗎 
참, 정말, 너무나도 공교롭게도, 월스트리트 저널에서 이런 보도가 나오자, DM 기능을 삭제하겠다는 이야기가 나옴.
참, 정말, 너무나도 공교롭게도, 우리는 이 비슷한 상황을 이미 '마음에 들어요' 기능 비공개 업데이트 때 경험해 봄.
RE: https://bsky.app/profile/did:plc:nruu4ywzsygrdnz6z2ynmgbd/post/3lmwqo33cok2f
Mental model mismatches can either be the easiest or hardest thing to overcome when teaching. In the easy case, the person you're teaching gets a win for free where something they thought was a problem turns out not to be a problem at all. In the hard case, you need to walk backwards to try and find all the assumptions and beliefs that the person might not even know they have, which can be painful, embarrassing, and hard to unlearn.
This is one of the biggest and saddest problems with using LLMs to learn. The LLM is not capable of doing what a human can, "ah, the way you have formulated the question shows me that there is some mismatch between your mental model of the thing and the thing itself, so before I try and answer that question we need to back up." Since the LLMs are explicitly trained to produce text that statistically resembles an answer to the question posed, and they are additionally conditioned to be complimentary and congenial, they not only routinely fail this basic part of teaching, but actively make it worse by reinforcing the unstated beliefs behind the framing of the question.
Now when I am teaching someone something and I know they use LLMs, I have an additional step where I have to back up and ask them to show me what the LLM told them, because now they have two problems - the original mental model mismatch, and some garbage that was presented to them as a highly confident true answer.
"카카오톡처럼 전화번호를 등록한 사람만 사용 가능하다."
<-이거라면 최악일듯. 하지만 진짜 최악은
"앞으로 트위터는 X챗(명칭이 언급되는 중)에 전화번호를 등록해야 사용 가능하고, 전화번호를 등록하면 X챗 이용자들에게 알림이 감"
<-이렇게 되면 그날 블스 서버 터지는 날.
RE: https://bsky.app/profile/did:plc:a6qvfkbrohedqy3dt6k5mdv6/post/3lmxn5fmqxs2y
1. DM을 유료 유저 기능으로 전환한다 2. DM 내용은 AI가 학습할 수 없다 (설마?) 3. 앞으로 DM은 일론 머스크를 비롯한 "검증된" 사람들에게만 허용된다. 같은 상상력이 막 자극되는데요..
SparklingOutlaw🍉
GENKI
Olivier Forget
