https://sf3d.fi/ Lab Notes and research from SF3D AI Lab — multi-agent cognition, local LLM inference, and AMD RDNA4 hardware. en Mon, 13 Jul 2026 00:00:00 GMT sf3d postbuild https://sf3d.fi/blog/the-value-function-that-reads-the-shape https://sf3d.fi/blog/the-value-function-that-reads-the-shape Mon, 13 Jul 2026 00:00:00 GMT To focus on one topic, the machine had to price every candidate question against seven sub-questions — and the obvious way to do that silently collapses all seven into one average direction. The theorem behind that collapse, the fix that reads a shape instead of a sum, and the discipline that catches it before a euro is spent. https://sf3d.fi/blog/the-machine-that-falsifies-itself https://sf3d.fi/blog/the-machine-that-falsifies-itself Sat, 04 Jul 2026 00:00:00 GMT Why we built a self-correcting AI research bench on a single AMD R9700: the thesis that insight is compression, the multi-agent engine that refutes its own ideas, and the three-tier funnel that spends money only where it pays. https://sf3d.fi/blog/from-liquid-nitrogen-to-tokens https://sf3d.fi/blog/from-liquid-nitrogen-to-tokens Wed, 24 Jun 2026 00:00:00 GMT I was out of the game for years — the honest story of why I quit overclocking, and why AI brought back exactly the same spark in a new form. https://sf3d.fi/blog/what-this-site-is https://sf3d.fi/blog/what-this-site-is Wed, 17 Jun 2026 00:00:00 GMT The short version — who SF3D is, what the lab is building on AMD RDNA4, and how this site works.