rurban

4 ms·

61.
by rurban 2mo ago
The nuclear incident in a sub would be less problematic than stealing the nuclear missiles out of the sub. Like a Titan out of the Kursk sub.
62.
by rurban 2mo ago
It's not just Western shipping parts. You can buy cheap Indian parts, and have the very, same insane shipping costs as with "Western" parts. Ie a 2€ part from India will still cost 25€, whilst the 2€ part from China will cost
63.
by rurban 2mo ago
Just take away their printers, install order screens in the kitchen. No way to cheat then
64.
by rurban 2mo ago
He has at least some incredibly good products
65.
by rurban 2mo ago
Why did they do this on a plane Wifi and not on a city coffeeshop wifi? Silly
66.
by rurban 2mo ago
You cannot blame ethics on the HF escape, because this model had no safeguards. It was just trained on unethical (and stolen) sources. So this cannot be the reason to blame missing ethics. The models needed to solve a problem, and they did.
67.
by rurban 2mo ago
The dissed Steering Comittee answer: https://discourse.nixos.org/t/the-scs-role-in-the-disbandmen... (They disagree)
68.
by rurban 2mo ago
More probably worst use of NixOS. In embedded we care for every single kb, and don't have tolerance for this kind of space waste management system
69.
by rurban 2mo ago
62s for a single instruction! Wonder if an compilers cost tables knows that. But it's data dependent, and cost functions probably don't do that.
70.
by rurban 2mo ago
I had to write a gnome shell extension to see my various LLM token budgets. Green, orange, red for Claude, OpenAI, GitHub Business, Kimi, DeepSeek, and such. Very useful.
71.
by rurban 2mo ago
I had similar problems with my free movie festival ratings database, so eventually I had to get rid of the dynamic lookup, and I only dump now static sites, hosted on GitHub Pages. I could not fight the AI bots for free, the hosting CGI con
72.
by rurban 3mo ago
This was the first time for me. All tests failed, for the whole day. For 20m, ok, I can tolerate that. But for a whole day? Microsoft sucks big time.
73.
by rurban 3mo ago
I'd really need to marry this with qemu and osxcross, the ability to cross compile from x86_64-linux to darwin. So far I could only test apple compilation, but not runtime yet.
74.
by rurban 3mo ago
Thanks for the depth-anything.cpp port. Saves us a couple of kernel OOM's, causing connection disruptions.
75.
by rurban 3mo ago
USA above average? That cannot be true
76.
by rurban 3mo ago
In gdb? It's a shame gdb picked python
77.
by rurban 3mo ago
Certainly not. Every hand-optimized tight loop or lookup beats every compiler still. Compilers are extremely stupid, and esp. the language designers.
78.
by rurban 3mo ago
Because the Intel syntax is limited, and pretty broken by binutils. I had to change my C compiler rcc to emit AT&T GAS, instead of Intel syntax, even if Intel syntax looks easier on the eye. GNU Assembler (GAS) ≥2.45 on x86-64: call and
79.
by rurban 3mo ago
Well, the middle ages. Thanks we got around this mess
80.
Include "Build.h"(includebuild.com)
2 points by rurban 3mo ago|0 comments
81.
by rurban 3mo ago
Not really fine tuned. But the base was Qwen. We are analyzing images
82.
by rurban 3mo ago
For LLM inference in production we use vLLM (with our own model).
83.
by rurban 3mo ago
Did they train on sama? Highly unlikely
84.
by rurban 3mo ago
Happy with oh-my-pi (omp)
85.
by rurban 3mo ago
No openmp, no openacc, though that would all be trivial. For linking, my own linker still has a lot of regressions, but is at least 10x faster than shelling out to gcc/ld. Similar to tcc. No linker scripts, no LTO via the native linker
86.
by rurban 3mo ago
No way. No open prompts. But I use omp /collab, so I prompt it from my phone. Constantly
87.
by rurban 3mo ago
It compiles the kernel and my big projects, but I haven't released it yet. Feature wise I don't do bitint, decimal and FloatX. No auto-vectorization, but better than most simd/neon projects, which don't use __attribute__
88.
by rurban 3mo ago
Of course not. I don't care for gcc's broken optimizer to remove dead code without any warning. gcc has too many bugs for my taste, sorry. And I don't want to wait 20m on gcc compilation, when tcc can do it in 1m.
89.
by rurban 3mo ago
The proof is gcc's torture test suite. Have a look
90.
by rurban 3mo ago
I had to reverse engineer Chinese SDK's for various embedded chips. Their libs were full of this shit. It works in some cases, but rarely in all. In embedded you can thankfully throw this shit out and replace it.

More ›