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.
Je recherche des testeurs pour ma première application iOS. Elle bloque 11,5 millions de numéros de téléphone liés au démarchage, à la publicité et au spam, basés sur le plan de numérotation de l'ARCEP.
L'application est open source !
Si vous souhaitez tester l'application : https://testflight.apple.com/join/CFCjF6d2
The Creature loves all the murals around town!
#Compton #LosAngeles #Mural #Art #Plushtodon #Merchtodon #Photo #Photography
ぺろっ、このIPアドレスはCloudflareさんわねってwhoisしたらMircosoftさんだったしこれはAWSさんわねって思ったのもMicrosoftさんだったし←
겜 상태 좀 에반데...
tzdata-legacy入れたらたぶんアメリカのタイムゾーンへのシンボリックリンク帰ってくるんだろうけど
have a UK visa now for 6 months, want to look for events i can join -- maybe give a talk
The great virtue proof assistants, more than verifying correctness, is the ability to organize mathematical knowledge so that it is recorded and can be inspected to any level of detail by just following links.
Before I used proof assistants, I recorded my thoughts on paper notebooks, like everybody else. Then whenever I wanted to write a paper based on these thoughts, I had to both organize the thoughts and check them again. This becomes a problem if these were thoughts of 5 years ago. In TypeTopology I have thoughts recorded from 5 years ago, or more, that I didn't have time to publish yet, but I won't need to check them again or reorganize them too much when I find the time to write a paper based on them.
I find proof assistants, in particular Agda, the one I use, to be perfect blackboards/notebooks.
Another great virtue of proof assistants is that it allows for collaboration with people who you don't need to trust based on their reputation. If their code passes the proof assistant, then it is correct. So when I look at a pull request in TypeTopology from a new contributor I haven't met before, I can just "referee" it for style, relevance etc., leaving the correctness to the proof assistant.
As you know, it is easier to write proofs you've come up with than to read somebody else's proofs. It is invaluable that the proof assistant checks somebody else's proof for me, and I can concentrate on their ideas instead.
I tried making artificial sunlight at home
Link: https://victorpoughon.fr/i-tried-making-artificial-sunlight-at-home/
Discussion: https://news.ycombinator.com/item?id=43497394
Evan Prodromou shared the below article:
Evan Prodromou @evanprodromou@socialwebfoundation.org
Mallory and I were lucky enough to participate in a study on New Paradigms in Trust and Safety: Navigating Defederation on Decentralized Social Media Platforms by the Carnegie Endowment. Published a few days ago, the paper goes into detail on the uses and misuses of server-to-server blocking, called defederation, on the Fediverse today. I’m looking forward to feedback from the moderator and user community on the Social Web.
everyone's complaining about spammers bypassing even the most advanced and irritating "are you a human" captchas, but i haven't seen anyone discussing this solution. i was watching this documentary last night, i was half asleep so i don't remember much about it. it was narrated by this guy who sounded kinda like indiana jones if he hadn't slept in two days.
anyway in the documentary they used a new kind of captcha called a voight kampff test. they ask you questions about how to keep a pet turtle or something and it can tell whether you're a human or not. it needs webcam access to see your eyes for some reason, but it seemed pretty reliable. it seemed a lot less annoying than selecting all the motorcycles at least
Evan Prodromou @evanprodromou@socialwebfoundation.org
Mallory and I were lucky enough to participate in a study on New Paradigms in Trust and Safety: Navigating Defederation on Decentralized Social Media Platforms by the Carnegie Endowment. Published a few days ago, the paper goes into detail on the uses and misuses of server-to-server blocking, called defederation, on the Fediverse today. I’m looking forward to feedback from the moderator and user community on the Social Web.
Central timeって言われたんだけどTZ=America/どこ
First-of-its-kind trial enables paralysed man to stand via stem cell injection
Link: https://www.nature.com/articles/d41586-025-00863-0?linkId=13622861
Discussion: https://news.ycombinator.com/item?id=43463779
ミュンヘン郊外に住んでた時はちょくちょく飛行機雲が見えてしやわせだったなあ
home screen keeps glitching what the hell
The internet of things is truly a wonder. 😝
Much of the Signal discussion is overlooking a key point: an entire apartment building was leveled to kill one guy.
How many civilians died? Is this a war crime?
The religious language—“prayer,” “God bless,” “Godspeed to our Warriors”—makes it all the more revolting, like a game of Battleship from the safety of a church basement.
Dagger: A shell for the container age
Link: https://dagger.io/blog/dagger-shell
Discussion: https://news.ycombinator.com/item?id=43486881
so sometimes when you're reading a dictionary definition of a word, you need to look up another definition. e.g. if you don't know what "smoothly" means, and the dictionary defines it as "having the property of being smooth", and you also don't know what "smooth" means, you have to look that up too.
but what if you also didn't know what "property" meant? you'd have to look up three words (smoothly, smooth, property) to understand one word (smoothly).
what if you only knew the definitions of the 10,000 most commonly used english words? which definition in the dictionary would require you to look up the most words to understand it?
It is intuitively plausible that formalization can help detect errors in mathematical papers. I discovered today that more old-fashioned numerics can also achieve a similar goal. In a paper I just uploaded at https://arxiv.org/abs/2503.20170 , I claimed an upper bound on a certain number-theoretic function t(N), and gave a proof. But, inspired by a nascent crowdsourced effort at https://terrytao.wordpress.com/2025/03/26/decomposing-a-factorial-into-large-factors/ to get enough estimates on this quantity for medium values of N to verify some explicit conjectures of Guy and Selfridge, I decided to plot this upper bound against the known values of t(N), and found a few places where the upper bound was actually *smaller* than the known value. This was of course distressing, but by isolating the smallest counterexample and numerically verifying key intermediate claims in my proof, I found where the error was (I was "off by one" in a certain estimate involving the floor function). This will be patched in the next revision of the paper of course, but is another example of how computer assistance (this time in the form of traditional numerics) can help detect and fix errors in papers.
Seeking: things you do while anxious to help be less anxious
跟風玩了一下AI吉卜力圖，想了想這對於人類會有什麼負面影響。如果產圖如此容易，是不是會降低人類對真實技術的學習，例如未來的人都不學繪畫了，反正下指令給AI就畫得出來。然後人類就不會（或很少）有新的風格創作產出，大家用的永遠都是末代人類創作的那幾套風格。會變成這樣嗎？
It is intuitively plausible that formalization can help detect errors in mathematical papers. I discovered today that more old-fashioned numerics can also achieve a similar goal. In a paper I just uploaded at https://arxiv.org/abs/2503.20170 , I claimed an upper bound on a certain number-theoretic function t(N), and gave a proof. But, inspired by a nascent crowdsourced effort at https://terrytao.wordpress.com/2025/03/26/decomposing-a-factorial-into-large-factors/ to get enough estimates on this quantity for medium values of N to verify some explicit conjectures of Guy and Selfridge, I decided to plot this upper bound against the known values of t(N), and found a few places where the upper bound was actually *smaller* than the known value. This was of course distressing, but by isolating the smallest counterexample and numerically verifying key intermediate claims in my proof, I found where the error was (I was "off by one" in a certain estimate involving the floor function). This will be patched in the next revision of the paper of course, but is another example of how computer assistance (this time in the form of traditional numerics) can help detect and fix errors in papers.
今日も家sshdにYubiKey待たせちゃってるおじさん
これすき
ベイビーわるきゅーれ エブリデイ！
第1話 10年後も一緒に死体凍らせよ
https://tver.jp/episodes/eplt32pjtf
ごあんぜん？
Why not check out a #FediArtist and comment on their music (or even buy a release) for today's #MusiciansDay?
For the most exhaustive list of #FediArtists head to https://indieart.support
Fresh releases, reviews and lots more are at https://nham.co.uk
And a whole lot of great recommendations by Fedi folks can be found at the https://isitfairtrademusicfri.day archives. (Also keep your eyes peeled for a reboot of it in the very near future).
Go on, make a muso's day and buy/boost their stuff 🎶 💙
Ouch, this sent a chill down my spine. [Tl;dr: GitHub locked a repo apparently because one of its contributors was in a US-sanctioned region]:
1. https://mastodon.social/@organicmaps/114155428924741370
2. https://mastodon.social/@organicmaps/114233788700982882
(TLDR: Thoughts on trauma with a relevant open question at the end.)
I wanted to take a moment and talk about trauma responses and how we, as people, try to get our needs met in a state of trauma.
What happens when we try to create areas that are 1 ) not trauma inducing and 2 ) are trauma informed while still experiencing trauma?
Trauma, in general:
When we're in a traumatic situation, we learn a lot of maladaptive responses. By that I mean we learn things that help us push through and/or survive the situation, but that damage healthy relationships. This is one of the earlier things that people in trauma recovery learn: "Let go of what no longer serves you."
Examples of things like this: it is common an abusive relationship to be told to Document Everything. Paper trails, possible camera recordings, etc. to keep yourself safe. BUT! When you transition from "unsafe" to "safe", these same behaviors are typically really invasive to your new partner, even if they understand why you're doing what you're doing.
Multiple spaces:
One aspect of trauma recovery focuses on what happens when the traumatic event ends. You've left your abuser, you've started to having savings after a life of poverty, etc. The skills and processing that happens here is different from enduring the trauma while it is still happening.
I think about this a lot, and heavily, at Nivenly 
(and it's projects, like Hachyderm 
) and others. Because the world is currently very trauma inducing and it is very difficult to ask people to unlearn patterns that hurt them, when they've been reinforced that those are the only patterns that are successful.
I do have more things I'd like to say, but in the now I want to pivot to a public discourse to hear what you have to say.
* How do you recognize when you're in a trauma free, but trauma informed, environment?
* What do you do when you want to participate in, organize/run, or create these environments?
* Or what are any general thoughts you have?
Scope broad: not just online, not just "events" like conferences, just a very open discussion 
(TLDR: Thoughts on trauma with a relevant open question at the end.)
I wanted to take a moment and talk about trauma responses and how we, as people, try to get our needs met in a state of trauma.
What happens when we try to create areas that are 1 ) not trauma inducing and 2 ) are trauma informed while still experiencing trauma?
Trauma, in general:
When we're in a traumatic situation, we learn a lot of maladaptive responses. By that I mean we learn things that help us push through and/or survive the situation, but that damage healthy relationships. This is one of the earlier things that people in trauma recovery learn: "Let go of what no longer serves you."
Examples of things like this: it is common an abusive relationship to be told to Document Everything. Paper trails, possible camera recordings, etc. to keep yourself safe. BUT! When you transition from "unsafe" to "safe", these same behaviors are typically really invasive to your new partner, even if they understand why you're doing what you're doing.
Multiple spaces:
One aspect of trauma recovery focuses on what happens when the traumatic event ends. You've left your abuser, you've started to having savings after a life of poverty, etc. The skills and processing that happens here is different from enduring the trauma while it is still happening.
I think about this a lot, and heavily, at Nivenly (and it's projects, like Hachyderm ) and others. Because the world is currently very trauma inducing and it is very difficult to ask people to unlearn patterns that hurt them, when they've been reinforced that those are the only patterns that are successful.
I do have more things I'd like to say, but in the now I want to pivot to a public discourse to hear what you have to say.
* How do you recognize when you're in a trauma free, but trauma informed, environment?
* What do you do when you want to participate in, organize/run, or create these environments?
* Or what are any general thoughts you have?
Scope broad: not just online, not just "events" like conferences, just a very open discussion
"In 2014, the all-party parliamentary group on modern languages estimated that the UK’s untapped linguistic potential was worth £48bn. It’s £8bn more than Reeves added to the tax bill last year, and if anything it has grown, since Brexit has re-oriented “global Britain” towards customers beyond Europe. Ask any native English speaker who trades internationally: if you’re buying, English will do, but if you’re selling, better to speak the client’s tongue."
Swahili? Mandarin? The UK is increasingly multilingual – yet our politicians won’t talk about it | Laura Spinney | The Guardian
https://www.theguardian.com/commentisfree/2025/mar/27/swahili-mandarin-uk-increasingly-multilingual
ねう
Adventures in #OpenBSD episode 0x819fa:
"I want my own secure cloud storage."
EASY.
Stefano Marinelli
Emelia 👸🏻
José A. Alonso
꧁ wakest ꧂
rclone mount with sftp and crypt enabled on it on the Macbookmount option is missing for rclone on fishlinux because of underlying FUSE implementation incompatibilites; THEREFORE do the logical thing and 
@gyptazy really nicely for a 
@BoxyBSD VM running FreeBSD
@OpenBSDAmsOpenBSD Amsterdam for initial setup of tailscale, then once BoxyBSD VM is in the Tailnet, can ssh / mosh in from home)doas and whatnotrclonetailscale file cp ( 
tailscale)sshfs using 
@soleneSolène 
's tutorials on the home boxNOTHING COULD BE SIMPLER. 
起動に時間がかかってるだけだった
Hunter Greene with 6 pitches of 100+ MPH in the first inning. Last time he had that many 100+ MPH pitches IN A GAME was May 19, 2024 at the Dodgers.
Set up by 
@KittyDas Metal Kitty 
, 
@bestiaexmachinaSciasm and 
@thomasTom 
, Musicians Day is a bi-annual event celebrating music by indie artists in the Fediverse. Musicians and listeners are encouraged to share their work or music they like.
As a federated music magazine with a sharp focus on Fedi music and Fair Music Platforms here are some of the things you will find on nham.co.uk:
El problema no es la privacidad.
El problema es la vigilancia.
とてもこまる
dockerが死んだかも
$ TZ=America/Los_Angeles date
Thu Mar 27 13:23:02 PDT 2025
A recurring lesson from both organizing and software engineering is the importance of asking people what they need instead of just trying to give it to them.
HDDの動作音っていいにゃあ
Glyph