Eating shit, the Firefox way.

blog.mozilla.org/en/firefox/ai…

#mozilla #firefox #AI

Introducing AI, the Firefox way: A look at what we're working on and how you can help shape it

AI Window is a user-controlled space we’re building in Firefox that lets you chat with an AI assistant and get help while you browse.

^{Kristina Bravo (The Mozilla Blog)}

New @Vivaldi release out now for your computer.

Tired of your AI browser yet? 🤖

vivaldi.com/blog/vivaldi-on-de…

#Vivaldi #AI #Browser #Windows #Macos #Linux #EU

Vivaldi 7.7: less friction, more control | Vivaldi Browser

This release brings together features that have been carefully refined based on what you’ve told us matters most: seamless access to your tabs across desktops…

^{Jon von Tetzchner (Vivaldi Technologies)}

Well, well, well.

"Judge Schwager said while issuing her ruling that she was astonished that OpenAI had not taken heed of what she called a clear legal situation."

Next up: books, illustrations, blogposts, news articles and related content, code... 🧹

#ai #copyright

dw.com/en/openai-loses-song-ly…

OpenAI loses song lyrics copyright case in German court

OpenAI lost a copyright infringement case in a lower German court for using popular song lyrics in its ChatGPT language model without paying royalties.

^{Mark Hallam (Deutsche Welle)}

When Python runs faster than C++ - you know something’s broken in the Matrix.

Meet Codon - a high-performance Python compiler that turns your code into native machine code, running 10–100× faster than CPython.

✅ No GIL.
✅ Real multithreading.
✅ Even GPU support.
✅ Same syntax, same vibe -
just turbo Python. 🐍💨

If you ever wished Python wasn’t so chill - your wish is granted.

github.com/exaloop/codon

#Python #DevOps #AI #ML #HPC #Performance #Coding

Got #qwen3coder 30B #LLM to work on #Framework Desktop. I tested it with e.g. "Write an Angular login page to oauth idp server. Use best practices. use HttpOnly cookies." and few others. It thinks few seconds and writes the code about in 30 seconds. The codes looked okay. I'm satisfied what i got.

And it's quiet - even when thinking! The monster multi-GPU AI machines are so outdated with a price of one 5090.

I followed github.com/pablo-ross/strix-ha… to install #llamacpp. Some changes from it:
- Skipped kernel update due to newer Ubuntu 25.10.
- Got sudo group error from distrobox. Removed --group-add sudo from distrobox create.
- Tweaked run parameters. I hacked them with trial and error. Looks like i can increase "context size" a lot.
Current command to run Qwen is:
```
distrobox enter llama-rocm-7rc-rocwmma -- ~/llama.cpp/build/bin/llama-cli -m ~/models/qwen3-coder-30B-A3B/BF16/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002.gguf --no-mmap -ngl 99 --ctx_size 16384 -n 20000
```
#homelab #AI

strix-halo-gmktec-evo-x2/ROADMAP.md at main · pablo-ross/strix-halo-gmktec-evo-x2

Configuration and documentation for optimizing Ubuntu 24.04 on AMD Ryzen AI Max+ 395 with Radeon 8060S for LLM inference using llama.cpp with ROCm 7 RC and rocWMMA. - pablo-ross/strix-halo-gmktec-e...

^GitHub

Moonshot AI's Kimi K2 Thinking Model Surpasses OpenAI's GPT-5

techlife.blog/posts/moonshot-k…

#AI #OpenAI #Kimi #KimiK2 #MoonshotAI #AIBenchmarks

Moonshot AI's Kimi K2 Thinking Model Surpasses OpenAI's GPT-5

Moonshot AI's Kimi K2 Thinking model outperforms OpenAI's GPT-5, sparking debate on AI dominance.

^TechLife

“And worse of all, my mental model of the code is completely gone, and with it my ownership.”

#AI #VibeCoding mastodon.social/@lobsters/1155…