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.

Let me talk, following on @thisismissemEmelia ๐Ÿ‘ธ๐Ÿป's post, about JSON-LD from a data engineering and web backend perspective:

It's a nightmare.

First: one of the first things you learn about how to address the OWASP lists is the importance of not just accepting random payloads, but carefully validating them, but with JSON-LD:

1. Basically any field may be absent, present, a URI, a list of objects, a list of URIs of objects, a list containing URIs AND Objects, etc.
2. Fields do not have consistent names. BY DESIGN. You _have_ to parse an object to know what is in it. Parsing the context and then building the map of names is basically an ฮŸ(n*m) process where the amount of processing time you spend is _under the control of the sender_ by default, and nothing is ever fully defined.
3. Actually fully validating an object to the degree of knowing if it is the same thing as something else requires solving a GI-hard problem if there are blank nodes.

Applying basic type theory to anything done with JSON-LD is an exercise in frustration.

There are certainly use cases for it, but 99% of the time when I've seen it being used it is not for those use cases, but for use cases where we would strongly prefer JSON-Schema and/or OpenAPI definitions for things.

0
0

Be aware... you're on air! ๐Ÿ”Š Secure your connected devices before they broadcast for you. A short, practical Secure IoT guide with clear steps to protect privacy and lock down smart gadgets. Must-watch for every smart home owner!
peertube.securitymadein.lu/vid

0
0

็ˆชใŒๆŠ˜ใ‚ŒใŸRJ-45ใƒ—ใƒฉใ‚ฐใ‚’ๅˆ‡ใ‚Š้›ขใ—ใฆๆ–ฐใ—ใ„ใฎใ‚’ๅœง็€ใ™ใ‚‹ใ“ใจใงๅพ—ใ‚‰ใ‚Œใ‚‹ใ“ใ“ใ‚ใฎๅนณ็ฉใ ใ‚ˆ

0
0
0

I spent my education career pushing back against crap like this, so let me clue you in on something:

When a tech company invests in schools, it's never for the benefit of student learning. It's always, always for creating loyal customers out of the students and teachers.

(h/t @AAKL)

the74million.org/article/exclu

0
11
0

๐Ÿง€ Wie kommen die Lรถcher in den Kรคse? Kaum jemand konnte komplizierte Zusammenhรคnge so verstรคndlich erklรคren wie Peter Lustig. Fast ein Vierteljahrhundert moderierte er die ZDF-Sendung "Lรถwenzahn", das Wissensmagazin fรผr Kinder und Erwachsene. Am 23. Februar 2016 starb er.

๐Ÿ“ ndr.de/geschichte/loewenzahn-o

Dieses Bild ist eine Infografik zum Gedenken an den bekannten deutschen Fernsehmoderator Peter Lustig. Es ist eine Hommage an sein Lebenswerk in der Sendung โ€žLรถwenzahnโ€œ.

Bildinhalt und Komposition
Im oberen Drittel des Bildes ist ein Portrรคt von Peter Lustig zu sehen. Er blickt freundlich und direkt in die Kamera.

Erscheinungsbild: Ein รคlterer Mann mit Glatze, einem grauen Vollbart und einer markanten, runden Brille.

Kleidung: Er trรคgt sein typisches โ€žLรถwenzahnโ€œ-Outfit โ€“ ein gelbes Hemd unter einer blauen Latzhose. An seinem linken Handgelenk ist eine einfache Armbanduhr mit weiรŸem Zifferblatt zu sehen.

Gegenstand: Er hรคlt eine klassische, silberne GieรŸkanne aus Metall vor sich in die Hรถhe.

Hintergrund: Hinter ihm ist ein Teil seines berรผhmten blauen Bauwagens mit einer grรผnen Tรผr und weiรŸen Fensterrahmen erkennbar.

Textliche Informationen
Die untere Hรคlfte des Bildes besteht aus einem dunkelblauen Hintergrund mit weiรŸem Text:

รœberschrift: โ€žLรถwenzahnโ€œ-Onkel mit Witz

Hauptzeile (groรŸ): Vor zehn Jahren starb Peter Lustig

FlieรŸtext: Er erklรคrte Kindern und Erwachsenen von 1981 bis 2005 die Welt โ€“ auf seine ganz eigene Weise.
0
14
0
0
0
1
0

๋ฉ”๋‹˜ ๊ทธ๋ฆผ์˜ ์†์ด ์•ผํ•˜๋‹ค๋Š” ๋ง ๋ญ”๊ฐ€ ์•Œ ๊ฑฐ ๊ฐ™์Œ ์ด ์ปค๋ฏธ ๋ณผ๋•Œ๋งˆ๋‹ค ๊ทธ๊ฑธ ๋А๊ปด...(? ๋ผ๋งˆํ‹ฐ ์†์ด. ์•”ํŠผ ๋ผ๋งˆํ‹ฐ ์†์ด. ํ•˜ ๊ทผ๋ฐ ๋‹ค์‹œ๋ด๋„ ์ข‹๋‹ค.......

0
0
0
0

๐Ÿง€ Wie kommen die Lรถcher in den Kรคse? Kaum jemand konnte komplizierte Zusammenhรคnge so verstรคndlich erklรคren wie Peter Lustig. Fast ein Vierteljahrhundert moderierte er die ZDF-Sendung "Lรถwenzahn", das Wissensmagazin fรผr Kinder und Erwachsene. Am 23. Februar 2016 starb er.

๐Ÿ“ ndr.de/geschichte/loewenzahn-o

Dieses Bild ist eine Infografik zum Gedenken an den bekannten deutschen Fernsehmoderator Peter Lustig. Es ist eine Hommage an sein Lebenswerk in der Sendung โ€žLรถwenzahnโ€œ.

Bildinhalt und Komposition
Im oberen Drittel des Bildes ist ein Portrรคt von Peter Lustig zu sehen. Er blickt freundlich und direkt in die Kamera.

Erscheinungsbild: Ein รคlterer Mann mit Glatze, einem grauen Vollbart und einer markanten, runden Brille.

Kleidung: Er trรคgt sein typisches โ€žLรถwenzahnโ€œ-Outfit โ€“ ein gelbes Hemd unter einer blauen Latzhose. An seinem linken Handgelenk ist eine einfache Armbanduhr mit weiรŸem Zifferblatt zu sehen.

Gegenstand: Er hรคlt eine klassische, silberne GieรŸkanne aus Metall vor sich in die Hรถhe.

Hintergrund: Hinter ihm ist ein Teil seines berรผhmten blauen Bauwagens mit einer grรผnen Tรผr und weiรŸen Fensterrahmen erkennbar.

Textliche Informationen
Die untere Hรคlfte des Bildes besteht aus einem dunkelblauen Hintergrund mit weiรŸem Text:

รœberschrift: โ€žLรถwenzahnโ€œ-Onkel mit Witz

Hauptzeile (groรŸ): Vor zehn Jahren starb Peter Lustig

FlieรŸtext: Er erklรคrte Kindern und Erwachsenen von 1981 bis 2005 die Welt โ€“ auf seine ganz eigene Weise.
0
14
0

I occasionally help an elderly neighbor get stuff done with their computer. And every single time, I walk away in incandescent rage at how hard we have made this stuff for people who have not spent their entire waking lives marinating in it

0
42
0

>โ€œBut it also takes a lot of energy to train a human,โ€ Altman said. โ€œIt takes like 20 years of life and all of the food you eat during that time before you get smart."

Human consumption of ~2000-2500 kcal/day is roughly 97-121 Wh. If that human lives to be 80 years old and keeps a constant energy consumption of ~100 Wh, then their total energy consumption is ~56 MWh.

GPT-3 consumed approximately 1.287 GWh during its training.

GPT-4 consumed 50-60 GWh.

Swing and miss.

techcrunch.com/2026/02/21/sam-

0
6
0
0

ใใ†ใ„ใˆใฐViteใ‹ใชใซใ‹ใซไฝ•ไธ–ไปฃๅˆ†ใ‹ใฎassetใ‚’ไฟ็ฎกใ—ใฆใŠใ„ใฆใ‚‚ใ‚‰ใ„ใŸใ„ใ‚“ใ ใฃใŸ

0

I spent my education career pushing back against crap like this, so let me clue you in on something:

When a tech company invests in schools, it's never for the benefit of student learning. It's always, always for creating loyal customers out of the students and teachers.

(h/t @AAKL)

the74million.org/article/exclu

0
11
0
0
0
0
0
2
0
0
0

ใฝใ—ใ”ใจ :saba:

ef6405ab28 (upstream/main) Profile editing: Featured tags (#37952)
e2aecd040c Collapse posts/collections in moderation interface (#37929)
e0cc3a30ef Allow displaying icon in `TextInput` component (#37951)
b785f05ddb Add coverage for `StatusBatchAction` (#37949)
f9c18ba170 Fix overly large avatars in report modal (#37950)
1d513966bd Remove unneeded haml-lint problem matcher (#37930)
6bd4d79802 New Crowdin Translations (automated) (#37940)
5fc2be54c7 Update dependency rspec-sidekiq to v5.3.0 (#37938)

0
0
0
5
0

To challenge bad AI drivers, I have to ask questions to the slopster like "explain to me what you did wrong." I don't like being vague, but if I'm too specific they just ask their AI. Inability to critically analyze and formulate their own mistakes is an instant unvouch. ๐Ÿ‘‹

Makes me feel like my own dad though. I used to hate being in trouble without knowing why I was in trouble and being told "think about it and come back to me when you figure it out." But, I understand him more and more every day.

0
0

If I believed that it were a thinking, feeling machine person (let alone a machine god), priority number one has got to be human rights for these things. Discussion of "alignment" and whatnot has to come second, right? Do you condition basic rights on a subordinate being agreeing with your values?

0
0

If I believed that it were a thinking, feeling machine person (let alone a machine god), priority number one has got to be human rights for these things. Discussion of "alignment" and whatnot has to come second, right? Do you condition basic rights on a subordinate being agreeing with your values?

@glyph@mastodon.social

Do you condition basic rights on a subordinate being agreeing with your values?
The fact that so many of the hardcore boosters and funders are eugenicists is... well, they've explicitly rejected the work on ML ethics as it actually impacts real, living people and society, instead embracing this 'alignment' thing. The idea of giving something rights only if it agrees with them lines up very neatly with how they treat people in general.

0
0
0
0
0
0
1
0
0
0
0

It's just one week until our un-workshop on Growing the Open Social Web!

The event page has been updated. Now we have the basic agenda, plus the submissions from attendees, and an attempt to summarize them.

Now also with Techdirt logo (as @j12tJohannes Ernst and @masnick.com are co-organizing)

Later today we'll e-mail an update to registered attendees.

There is still time to register, it's online, so it's easy to participate!

More info and registration: fediforum.org/2026-03-growing-

0
0
11
1

I made a promise to myself to never post a recruiting call for my research projects on mastodon ever again. However, I do know lots of folks follow me here who care about this part of my work and might want to participate in an open science project for developers. This is a note that I will be posting recruiting calls on other platforms and my sites in the next week, if you are curious to check that out.

0
0
13
0
0
22
0
1
0