Carlos

Water day

My biggest open-source release! NumKong — 2'000+ SIMD kernels for mixed-precision numerics, from Float6 to Float118. Started in 2023. Opened the PR in 2024. Finally, merged this week! RISC-V, Intel AMX & AVX-512, Apple SME & SVE, WASM Relaxed SIMD. 200'000 lines of code in a 5 MB binary. Same scale as OpenBLAS. Available for C 99, C++ 23, Python 3, Rust, Swift, GoLang, & JavaScript. Int4 dot products via nibble algebra. Ozaki Float64 GEMMs on Float32 tile hardware. 6-bit and 8-bit floats back-ported to 10-year-old CPUs. 5'300x faster Geospatial metrics than GeoPy. 200x faster Kabsch than BioPython. 0 ULP where OpenBLAS hits 56... and a lot more! pip install numkong Or pull it from NPM, Crates, GitHub... and let me know what breaks 🤗 Links & highlights ⬇️

We used to be a society.

I have been dreaming of this day for a long time. Arena is now a book publisher, and our first volume, "Silicon" is open for preorders. It's quite unlike anything you've seen: a coffee table book capturing the ecstatic beauty of silicon technology. arenamag.com/silicon

El "Efecto Cursor": Rápido hoy, inmanejable mañana. Un estudio de la Universidad Carnegie Mellon (publicado en arXiv) ha analizado miles de repositorios que usan Cursor AI. ¿El resultado? Sí, escribes código 3-5 veces más rápido el primer mes, pero a los dos meses la complejidad y los avisos de errores estallan un 41%. Al parecer la IA está llenando GitHub de "spaghetti code" automatizado. Básicamente, estamos pidiendo préstamos de deuda técnica a un interés del 300% que pagaremos en 2027. La velocidad es una ilusión si el código es un laberinto.

For the siege of the Third Castle in Ran (1985), Akira Kurosawa built a massive, real castle set on the slopes of Mount Fuji. He then burned it to the ground in a single take, as no miniatures could replicate that level of destruction.

@willreil this cutter works like a charm for pcbs, some passes and "clac" amazon.es/Olfa-PC-l-C%C3…

Other than the fact that they’re all joined together in a slab, I am really proud and happy with how they turned out. I think these look sick. The plan is to score some with a knife, CNC some, and use a Dremel.

You're amazing, @matthewvenn! Congratulations! A well-deserved recognition!👏👏👏

SQLite is a masterclass in a library design. It shows how projects of even huge complexity can be very simple and unobtrusive to use. You can learn a lot from them if you are library author, regardless of programming language.

#NT松戸 これ面白い。ベースボードに磁石で張り付いて電源だけ供給、ブロック間の信号は赤外線で送ってるらしい。

Here is the Rainbow Islands jammer board. It has a 68000 CPU and a Z80. I'm consigning it to storage with the other 2 boards up there in the loft. At least it allows me to play the MAME full version with a clear conscience.

It works 🚀 First demo of the new #386fpgacore running on real FPGA hardware (Sipeed Tang Console 138K). VGA output, 3DBench, Norton Commander, and Turbo C all running. Still slow and buggy — but a lot of programs already work.

Directed by David Lynch!

Emociona ver como resiste en España algún cine solitario como este de Teruel capital

MAKE DRAGONBALL GREAT AGAIN! 🔥 3 chip solution (SoC with 68k core, FLASH, DRAM), no WIFI, no internet, no GPU, just raw frame buffer... but 1 month battery lifetime and infinite fun! \o/ flickr.com/photos/micahdo…

Esta noche, 21:30 en directo, charlamos con @dfsantos1 y su libro soobre IA para Z80. Podéis comprar el libro aquí: amazon.es/dp/B0GNZ9HSJ8

Es lógico que los gurús de las grandes empresas tecnológicas hayan adoptado una puesta en escena que siempre ha correspondido a líderes religiosos: papas, imanes, lamas, sumos sacerdotes. También aspiran a controlar nuestras vidas y nuestras conciencias.

Having a 1:1 80386 on #MiSTerFPGA would be absolutely awesome for retro computing. These projects push the limits of preservation and joy for all of us

Today's project: got my super basic Zig task scheduler running on @splinedrive's KianV softcore RISC-V CPU (programmed on an icesugar-pro dev board). Pretty cool!

AI models from OpenAI, Google, Anthropic and xAI can reproduce entire novels from memory. Researchers extracted 95.8% of Harry Potter from Claude nearly word for word. Gemini 2.5 Pro and Grok 3 didn't even require bypassing safeguards - they just kept writing. AI companies have long claimed their models "learn patterns" rather than store copies. A German court already ruled this constitutes copyright infringement. Anthropic paid $1.5bn in settlement. Cost of extracting a book? Between $2 and $120. "We don't store copies" - technically not, but 95.8% is a bit close? arxiv.org/pdf/2601.02671