Twit Sile

@JulianRoepcke Hätte afd kein Russland Problem, würden sie jetzt bei 50% liegen...

Ein russischer Oligarch will, dass seine Schwester wieder im schönen Europa shoppen kann. ➡️Also haut er den Außenminister des mit ihm finanziell eng verbundenen russischen Präsidenten an, er solle das mal in der EU regeln. ➡️Der trägt den Auftrag weiter an den von russischem Öl abhängigen ungarischen Außenminister, der verspricht, sich für eine Streichung der Schwester des Multimilliardärs von der Sanktionsliste einzusetzen. ➡️Das ganze wird aufgezeichnet und so zweifelsfrei bewiesen. Und jetzt kommen die Unterstützer der AfD ins Spiel. Ihre Schlussfolgerung: 🤡Korrupte EU, Russland-feindliche Berichterstattung, Ungarn macht alles richtig, das ist im Interesse des Volkes. ➡️Und die wundern sich, warum man sie entweder nicht ernst nehmen kann oder als Bedrohung für Deutschland sehen muss …

@willametteshark @AndrewPerpetua What? They have guns dude. They are still capable to kill without their vehicle

@AndrewPerpetua An airborne pilot is an imminent threat; a downed pilot is not. This was the ROE during the 20th century. I support Ukraine, am not a legal expert, and don't actually care- I just think it's disingenuous to compare the two.

Helicopters are flying artillery pieces. If a 2A65 Msta-B crew lost their gun and tried running away and got killed nobody would blink. But a helicopter pilot gets killed and people shed a tear. There is no difference. Morally or legally.

@alina_yurenko How did we get there? Having maven....

here we go again

Eine mutmaßliche Vergewaltigung, wochenlanges Schweigen, Behördenversagen. Ein Fall aus Berlin wirft wieder einmal die Frage auf: Warum fällt es vielen so schwer, offen über sexualisierte Gewalt zu sprechen, wenn die Täter migrantische Wurzeln haben? @BILD m.bild.de/politik/meinun…

@khmarbaise @gutsOfDarkness8 Don't bother. It's AI generated article

@gutsOfDarkness8 try-with-resources existing since JDK7 (2011) docs.oracle.com/javase/7/docs/… and with effectively final variables since JDK 9 (2017) openjdk.org/jeps/213

Cleanest way to handle multiple resources in Java – no leaks, no finally blocks: In 2025+ Java codebases, you can use try-with-resources with effectively final variables declared outside the try block - huge readability & safety win.Even better: create small, focused AutoCloseable helpers that encapsulate cleanup logic. Example below shows: - multiple resources - no explicit close() calls - custom wrapper for readability - zero chance of forgetting to close

@sivalabs I just always use the latest versions. No need for configurations :)

I use SDKMAN all the time and use ".sdkmanrc" file to automatically set project-specific Java, Maven, Gradle versions. Created a simple IntelliJ IDEA Plugin to get autocompletion in .sdkmanrc file. github.com/sivaprasadredd…

How European restrictions on free speech can be dangerous: Germany uses anti-Nazi law to investigate a writer with a long history of attacking Nazi ideology because he posted an image comparing Putin to Hitler - an idea very much worth expressing. thetimes.com/world/europe/a…

@0xlelouch_ Async is never faster. It allows you to handle more connections at once, but not faster. Compare sync code with async in low number of connections < 512. Async can't be faster. Why would it? It's often slightly slower

If async is faster, why not make the whole codebase async?

@Aku_700 But why?

Netflix Propaganda:- • White man is the villain • Black man is the hero • Gay guy is the voice of reason • Woman has the balls • Dad is an idiot • Mom is the breadwinner • Children are sexualized • Trans/non-binary character is exceptionally wise or persecuted • Straight white male authority figure is corrupt/incompetent/abusive • Traditional masculinity is mocked or pathologized • Christianity/religion is depicted as fanatical or hypocritical • Police/military/law enforcement portrayed as systematically racist/oppressive • Climate change/environmental activism is the heroic motivation • Interracial pairings (especially black man/white woman) emphasized as triumph • Traditional/nuclear family shown as dysfunctional • Non-traditional family setups idealized • Wealth/success tied to villainy unless redistributed (anti-capitalist undertones) • Trans ideology or grooming elements shoehorned into kids’ content Netflix is social engineering, not entertainment.

Vor 4 Jahren eskalierte 🇷🇺 seinen Krieg in Ukraine. Putin-Fans, Russlandversteher und "Realisten" wussten damals ganz genau, dass a) das niemals passieren würde b) Ukraine sich in 3 Tagen ergeben würde. Ich habe die schlimmsten Äußerungen von damals gesammelt. Viel Spaß! ⬇️

@allenholub And what do you do if AI is not able to fix a critical bug. You have no idea about the code. The client is pissed off and loses money.

I keep reading posts about people who wrote applications entirely with "zero coding." They say they don't read the code and instead refine the application using increasingly precise prompts. I have some questions: * Is the application nontrivial? * Is it correct that when you find even a minor bug, you regenerate the entire application from a modified prompt? * If not, what's the granularity of the regeneration? * What is the underlying architecture? * Do you have to redeploy the entire app when you regenerate it? * What infrastructure do you use for the redeploy? Is it automated? * What's the downtime with a redeploy? * What happens if the new version modifies the database structure? * How do you guarantee database integrity across versions? * What do you do when that modification introduces different bugs or changes things that used to work? * Was the UI/UX developed interactively, with customers in the loop? * Is the UX unchanged after the regeneration when this is a simple bug fix? * How do you prevent specious UI changes? * What sort of testing do you have in place? * Is the testing automated? * Did you write the tests or did the LLM? * If it's the latter, how can you be sure that the tests are adequate if you don't read them? * Is the system secure against exploits after every regeneration? * How fault-tolerant is your system? If it's a web application, what downtime is expected? * Is it possible for multiple engineers to work simultaneously on this system? * How do you handle merge conflicts? * How do multiple engineers not get in each other's way? * Does your process work in a regulated environment? I have more once you've answered these.

@unclebobmartin Are they all working on the same problem? How do you coordinate, so there are no conflicts? What if they need to run the app? Each agent needs a separate running setup like database, Kafka etc

Starting with three claudes. One implementer. One planner. One reviewer. Using git worktrees instead of cloned repos.

@gunnarmorling 800% CPU. Is it also 8x times faster?

🏎️ Made some good progress over the weekend improving the performance of the #Hardwood parser for Apache Parquet; 11 files from the 2025 NYC taxi ride data set (~720 MB) can now be fully parsed in ~1.9 sec. Besides some decoder tweaks, I focused mostly on improving the parallelism of the parser at this time. Which, as it turns out, is a surprisingly tricky problem. I'm still not really happy with how things are, but they are much better now than before. A naive approach would be to just parse separate column chunks in parallel. This can help a little bit, but it falls short very quickly: Your file might not just have many columns to begin with, or they could have different lengths (one column is repeatable, while others are not). So I took a first stab at implementing page-level parallelism (The Parquet format organizes files in row groups which are made up of column chunks which are made up of pages), which allows to fan out the work on a more fine-grained level. Once you have identified the page boundaries within a chunk (Parquet supports indexes for that, but not all files have them), and you have parsed its dictionary (if the column uses dictionary encoding), you can distribute the work of parsing pages to multiple threads, which increases CPU utilization a lot. There's still a problem: there can be significant differences in terms of how CPU-intensive the decoding of a given page is, depending on its encoding type, and thus to the time it takes to parse a page; essentially, faster columns will wait on slower columns. The way I'm currently tackling this is via adaptive page pre-fetching: slower columns build up a deeper page pre-fetching queue over time, thus more threads can pick up their parsing tasks. Eventually, whenever a new page is needed when iterating through the Parquet file, that page should be decoded already, no matter its value or encoding type. This gets me to a CPU utilization of ~800%, which is a significant improvement over single-threaded parsing or basic column-level parallelism. In wall clock profiling, I'm still seeing decoder threads idling for about half of the time, but we're getting there, step by step 🤓. 👉 github.com/hardwood-hq/ha…

@maxandersen How does Maven do it?

Its 2026 and two of the most popular dev tools in Java space: Gradle and IntelliJ has no way to run a main class with a working terminal. What’s up with that ?

@starbuxman I wish the value classes arrived first allowing all those promised performance improvements.

I love this! They’ve already started prototyping* type classes in Java * this is nowhere near ready for inclusion in a JEP! mail.openjdk.org/pipermail/valh…

@piotr_minkowski How is it different from a simple @bean annotated method with if inside?

A new way to register beans programmatically beginning from Spring Framework 7 👇👇👇

Russians strapped an anti-tank mine to an African and turned him into a “kamikaze”. “You’re coal. Now you’ll run across the field. Go!”

Listen up, you piece of shit. For four years, Russians have been killing innocent Ukrainians, raping women, stealing, and destroying an entire country. You piece of trash, you haven't even condemned these war crimes, not once. And now you're all too happy to believe the Russians' latest lie because you want to keep buying cheap Russian raw materials. Fuck you.

Dabbling a bit with Rust again over the holidays. Gotta admit that not having something akin to Quarkus' dev mode feels like a massive step back in terms of a fast inner dev loop.

@gunnarmorling It would be nice however if PMs would use a bit of their brain and spend some time to think about what they want. In my experience they almost never do. Writing one sentence reqs and that's it. It's not only us...

One common mistake many software engineers keep doing is to wait for perfect requirements from PM. That's not how it works, in particular as you get more senior. It's your job to help craft requirements, identify what makes sense and what is doable (and what not).