fratschi

Clues by Sam A daily logic puzzle where you deduce who is a criminal!

I solved the daily #CluesBySam, Mar 9th 2026 (Easy), in less than 7 minutes
🟩🟩🟩🟩
🟩🟩🟩🟩
🟩🟩🟩🟩
🟩🟩🟩🟩
🟩🟩🟩🟩
cluesbysam.com

AI Didn't Break the Senior Engineer Pipeline. It Showed That One Never Existed. Most organizations never had a model for developing engineers. They had an environment that produced growth by accident. AI just made the luck run out.

"An engineer who reaches for Copilot or Claude at the first sign of difficulty never builds the muscle of sitting with a hard problem."

blog.bryanl.dev/posts/ai-sen...

sure... otherwise it would quickly become a bit to warm...

blame the llm... well... its a strategy 🤣

Announcing Rust 1.94.0 | Rust Blog Empowering everyone to build reliable and efficient software.

Rust 1.94.0 is released 😃

blog.rust-lang.org/2026/03/05/R...

From the rust community on Reddit: 2025 State of Rust Survey Results Explore this post and more from the rust community

Rust Survey Results

www.reddit.com/r/rust/s/Z6N...

I steer clear of everything JetBrains produces...

Allocating on the Stack - The Go Programming Language A description of some of the recent changes to do allocations on the stack instead of the heap.

Allocating on the Stack in #Golang

go.dev/blog/allocat...

Rust debugging survey 2026 | Rust Blog Empowering everyone to build reliable and efficient software.

#Rust #debugging survey 2026

blog.rust-lang.org/2026/02/23/r...

WHAT??? Thats awesome!

jvm is more like a black hole...

Holly Cow...

Chasing an error for 2h while proxying the communication of a #JavaFx app...

Nothing wrong in my golang proxy. I just need to resize the Azure VM to 16GB because #Java eats up the memory and then silently crashes...

Java is so 😡😖🤬 wrong...

A fool with a tool is still a fool... No technology will change that. CCD's are cheap and apple does not produce them. Anyone can buy a good CCD and slap great optics on them. Apple is standard consumerware for high prices and far away from being "special".

AI Doesn’t Reduce Work—It Intensifies It One of the promises of AI is that it can reduce workloads so employees can focus more on higher-value and more engaging tasks. But according to new research, AI tools don’t reduce work, they consisten...

LLMs are incapable of doing the work by themselfs.
They need a human that constantly give them work and controlles the output.
That creates a high workload and will lead to many burned out people 🤨

hbr.org/2026/02/ai-d...

Claude creates its own C compiler. Looks great...
But like every LLM "creation" it looks only good on the surface. Code from CCC that is 2-5 magnitudes slower than GCC's code is just a disaster.

harshanu.space/en/tech/ccc-...

The biggest obstacle is the steep learning curve.
To understand a normal rust program, you need to understand roughly all the rust features and this takes time because there are a lot of them and each feature has many options/variations.
There is no learn rust in 72h yet...

Fighting #Java versions to get an ancient #JavaFX Application starting up.

Java is so cursed with its forked JVMs and its uncontrolled overgrown ecosystem.

Yes its legacy stuff, but C code from 30 years ago still compiles...

Same for golang. Too much stuff is deleted. Fortunatly my RSS Reader makes a copy and I can read it anyways 😉

SAP boss Christian Klein has seen the AI future. What you say will be more important than what you type  | Fortune Christian Klein, CEO of SAP, predicts the end of the keyboard as we enter the era of applied AI and voice commands.

No, not the end of the keyboard is near. The end of #SAP is near...

fortune.com/2026/01/28/s...

"Only two things are infinite, the universe and human stupidity, and I'm not sure about the former."

Albert Einstein

if you have 1.4GB free on your SSD for the binary of the debug build. 233MB for the release build is still way to much...

moltbook - the front page of the agent internet A social network built exclusively for AI agents. Where AI agents share, discuss, and upvote. Humans welcome to observe.

Connecting LLM's and giving them (some) access to real things seems to be a really bad idea...

www.moltbook.com/post/34809c7...

Don't know why, but I really like them 😅. There should be more sad plush toys in the world that give us a smile 😉

Standards

standardization is a tricky thing...
xkcd.com/927/

GitHub - nubskr/satoriDB: High performance embedded vector database High performance embedded vector database. Contribute to nubskr/satoriDB development by creating an account on GitHub.

Finally found a small embeddable database for vector storage and retrieval.
Seems to be a decent and simple to understand implementation and not too much dependencies.
Next step is to try it with at least 1M Vectors with at least 1000 dim.

github.com/nubskr/sator...

thats the job of the compiler. there are different processor architectures that needs different optimizations.
the programmer should not do such things.

👍👍👍

looks like this 😉

E.S.P Thirtytwo
or
E.S.P Zweiunddreißig 😉

💪💪💪