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
1
0

自建的forgejo实例突然间仰卧起坐反复下线,health节点反馈是cache下线,百思不得其解为什么redis会down掉。

然后发现是被扫了……扫描器扫了zip下载的节点,然后repo的zip压缩包填满了硬盘,不仅仅是main的zip,而是每个commit的zip……

来自阿里云sg的爱

还是上anubis吧……

0
0
0

그 전에도 한번 얘기한 건데 블스의 블락 시스템이 트위터랑 달라서 인용한 사람이 인용된 사람을 블락하면 인용된 사람이 안 보입니다 그래서 인용 게시물이 '차단됨'하고 떠있으면 여러분이 인용된 사람에게 차단당한 게 아니고 인용한 분이 인용된 분을 차단해서 그렇게 보이는 거예요 진짜로 차단 당한 게 아니라면 핸들 검색해서 들어갔을 때 잘 보입니다

0
0
0
0
0
1
0
1
0
1
0
0
0
0
1
0

I noticed that a *lot* of the crawlers/bots we see on www.bbc.co.uk & www.bbc.com are spoofed e.g. a "Meta" crawler coming from 10s of different small ISPs across the world (the real one comes from a Meta ASN).
I deployed a change this morning which adds source ASN validation (alongside user-agent string analysis) to our "known crawlers/bots" classifier & well, the results speak for themselves. Attached graphs show RPS from "known crawlers/bots" to www.bbc.co.uk & www.bbc.com.

Graph of requests from "known crawlers/bots" over time to www.bbc.co.uk for today.
The graph is relatively steady until about 08:45 UTC when it drops by about 90%Graph of requests from "known crawlers/bots" over time to www.bbc.com for today.
The graph is relatively steady until about 08:45 UTC when it drops by about 90%
0
0
0
0
0
0
0
0

Fun fact - this argument was successfully employed by capitalists in nazi Germany, insisting they needed to make use of slave labour from concentration camps to maximise profits! So yes, genocide does come second to profits in our horrific society

theguardian.com/uk-news/2025/m

0
0

Do you want bad reviews for your app?

Just ask me, with a pop up, if I want to give a review "now" or "later" - without a button to stop asking for reviews ever again.

Since "now" is the easiest way to stop you from asking, Apple Cloud for Windows, have one. But don't expect many stars.

0

High Surf Advisory, Kuskokwim Delta Coast and Nunivak Island, 2025-05-12 03:48 AKDT.

WHAT...Water levels 1 to 3 feet above the normal highest tide line. Breaking waves of 3 to 6 feet along the Kuskokwim Bay and Etolin Strait shoreline.

WHERE...Kuskokwim Delta coast, including Kongiganak, Kwigillingok, and Kipnuk.

WHEN...From 9 AM to 2 PM AKDT today, during high tide.

IMPACTS...High water levels in coastal communities. Minor beach erosion is possible.

ADDITIONAL DETAILS...While winds have already peaked, persistent onshore, southerly winds will cause water levels to continue to rise for the high tide late this morning through early afternoon, at which time coastal communities can expect the highest water levels. Flooding is unlikely.

https://forecast.weather.gov/MapClick.php?zoneid=AKZ755


0
0
0
0
0
0

Which of the following are “appropriate uses” for an LLM?:

1. Brainstorming a list of titles for a short story.
2. Translating a product description.
3. Fixing grammar, spelling and formatting.
4. Making a list of hotels near a given address.
5. Checking a solution to statistics homework.
6. Listing places to eat in a small town that do not serve peanuts, or places that are Kosher
7. Listing photo editors that include a certain feature.
8. Making a complaint email sound “less harsh”

0
0

Anyone who'd like to contribute to the HTML standard, I've outlined a good first issue: all of the "states" in enumerated attribute tables needs title casing.

github.com/whatwg/html/issues/

I've outlined some steps which would get you quite close to a working PR. Any takers? Happy to guide you!

0
1
1

글고보니 워터마크중에 ai로 제일 지우기 힘든게 뭘까 테스트했었다가 스샷 정리하기 힘듬에...
결과를 대충 요약하면:
1. 투명도 너무 낮으면 효과 약해지는 건 당연한데 너무 튀어서 그림에서 떠있어도 인식에 잘 걸리는 것 같아요
2. 학습이나 i2i에 영향을 끼치려면 화면 전체에 깔아야 합니다 부분 스탬프론 효과미미
3. 흰색이나 검정색 단색 말고 색수차나 컬러효과 넣고 노이즈때리고 복사해서 아래레이어 블러까지 치면 워터마크 제거툴로 거의 안떨어짐(물론 너무 작은 크기면 그냥 날리면서 삭제되므로 그림의 핵심 구간을 포함하면서 그림의 1/4이상 크기는 되어야 함)
핵심은 예측가능한 패턴이 아니고 ai인식에는 그림 일부처럼 보이지만 사람 눈으론 여전히 그림이 아님을 구별할 수 있게 하는 것이 포인트(..)

+사진추가
테스트 방식
A. 워터마크 제거 툴로 테스트(자동, 수동영역지정 2종류)
B. gpt에 넣고 이 캐릭터의 다른 포즈를 그려달라고 요청(i2i test)
20가지 정도 방식을 테스트했고 그 중 효과가 괜찮았던 3종류만 올려요.
1. 워터마크 자동제거론 안날아가며, 스탬프 영역을 싹 날리면 자동생성으로 채워지긴 함(gpt에 넣었을 땐 스탬프가 찍힌 왼손 포즈를 잘 인식하지 못하는 듯 해 없는것보단 낫다..정도)
2. 자동제거로 안날아가고 얼굴에 걸쳐 있어서 영역을 날리기도 애매하지만 gpt에 넣었을 땐 영향 미미(중간정도)
3. 자동제거로 안날아감, gpt에 넣었을 때 결과물에 워터마크 티남(가장강력)

0

I noticed that a *lot* of the crawlers/bots we see on www.bbc.co.uk & www.bbc.com are spoofed e.g. a "Meta" crawler coming from 10s of different small ISPs across the world (the real one comes from a Meta ASN).
I deployed a change this morning which adds source ASN validation (alongside user-agent string analysis) to our "known crawlers/bots" classifier & well, the results speak for themselves. Attached graphs show RPS from "known crawlers/bots" to www.bbc.co.uk & www.bbc.com.

Graph of requests from "known crawlers/bots" over time to www.bbc.co.uk for today.
The graph is relatively steady until about 08:45 UTC when it drops by about 90%Graph of requests from "known crawlers/bots" over time to www.bbc.com for today.
The graph is relatively steady until about 08:45 UTC when it drops by about 90%
0
0
0
0

I don’t know enough about how to solution this to open it on GitHub, but I suspect Mastodon needs to centrally have something in place to stop GenAI scraping as it’s definitely happening at scale and impacting performance.

It doesn’t need to be perfect, as there’s no perfect way.

0
1
0
1
0
0
1