sbilik
2.2K posts

sbilik
@sbilik
XY, Classical Liberalism & Subsidiarity, Catholic, Comp Engineer, Veteran — USAF Captain, Widower Father of 5, GPU SoC designer, Geek, Emacs, Linux, Opus Dei

Why almost all the new distros are based on Arch Linux these days? 🤔

HOT or NOT?






Maggie Goodlander left New Hampshire at 15 years old to attend a tony boarding school in Groton, MA — Since then, she’s chosen to spend 83% of her life living outside of NH She only came back to run for Congress, backed by out-of-state donors. Where do her interests really lie?

"Moral of the story? Get married. Stay married. Have kids. Raise them together... You’ll be happier... The data says so." And might I add: Don't wait too long. We started in our early 30s and ended up with three. If we had started in late 20s, we might have gone for four.

We heard your feedback and took it seriously: Kagi Search now loads best on Netscape Navigator

FFmpeg is moving to Rust 🦀 Our use of C and Assembly in FFmpeg has been an unacceptable violation of safety. FFmpeg will be running 10x slower - but we're doing it for your safety. All your videos will appear green - safety first, working software later.

These absolutely insane LLM wizards are now experimenting with Turboquant not just to compress KV cache, but now, the entire model itself. This test showed a >50% reduction in memory footprint, allowing for Qwen 3.5-27B to be run on a single RTX 5060 @ 3.15bit precision - with no apparent degradation. This just goes to show that we're likely nowhere near full optimization for existing models. We are likely <1yr away from running big models on smol devices with minimal consequence. And during that time, they will only get better and better. What a time to be alive.

the original TurboQuant paper tested on A100 with models up to 8B. 6 days later, a bunch of strangers on the internet had it built and running on: - Apple Silicon M1 through M5 - NVIDIA 3080 Ti through DGX Spark Blackwell - AMD RX 6800 XT and 9070 - a 10-year-old Tesla P40 - an 8GB MacBook Air - models from 3.8B to 70B across 6 architecture families - 30+ independent testers along the way we found new optimizations the paper didn't cover and failure modes it didn't test. the fact that a loose group of people across the world can read a paper, build implementations from scratch, stress-test across hardware none of us could individually afford, and push the research further in under a week is genuinely one of the best things about this era. the tools and the community make it possible. open source is something else.

Memory manufacturers pivoted almost entirely out of every other sector off a NON-COMMITTAL LETTER OF INTENT from the guy known for LYING NONSTOP. You can’t make this shit up.








