<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en"><title>Blog · TRINITY</title><subtitle>Every article is published on this domain before anywhere else, with its receipts and with what it does not settle stated in the text.</subtitle><link rel="self" type="application/atom+xml" href="https://t27.ai/blog/feed.xml"/><link rel="alternate" type="text/html" href="https://t27.ai/blog/"/><id>https://t27.ai/blog/</id><updated>2026-08-12T00:00:00Z</updated><author><name>Dmitrii Vasilev</name><email>admin@t27.ai</email></author><entry><title>Sixteen defects fixed, zero of them reached the consumer</title><link rel="alternate" type="text/html" href="https://t27.ai/blog/a-repair-reaches-only-the-copy-it-lands-in/"/><id>https://t27.ai/blog/a-repair-reaches-only-the-copy-it-lands-in/</id><updated>2026-08-12T00:00:00Z</updated><published>2026-08-12T00:00:00Z</published><summary type="text">Two repositories carried the same six files under the same names. Repairing one changed nothing downstream, and no instrument in either could report why — because each copy compiles entirely on its own.</summary><category term="Zig"/><category term="Modularity"/><category term="CI"/><category term="Verification"/></entry><entry><title>A green CI does not mean the library builds for you</title><link rel="alternate" type="text/html" href="https://t27.ai/blog/green-ci-does-not-mean-usable/"/><id>https://t27.ai/blog/green-ci-does-not-mean-usable/</id><updated>2026-08-11T00:00:00Z</updated><published>2026-08-11T00:00:00Z</published><summary type="text">A Zig package with passing CI could not be compiled by any consumer. Lazy analysis means a test run proves only what the tests happen to touch — one line exposed five hidden errors in one package, and sixteen in the package under it.</summary><category term="Zig"/><category term="CI"/><category term="Verification"/><category term="Static analysis"/></entry><entry><title>The scale field does not need eight bits, and we were not the first to say so</title><link rel="alternate" type="text/html" href="https://t27.ai/blog/scale-field-width-already-published/"/><id>https://t27.ai/blog/scale-field-width-already-published/</id><updated>2026-08-11T00:00:00Z</updated><published>2026-08-11T00:00:00Z</published><summary type="text">Four bits cover the shared scale on every checkpoint measured -- bit-identical to E8M0, zero blocks truncated out of 20,462,464. The observation was published first by someone else; what is ours is the proof around it, and the constant does not survive contact with activations.</summary><category term="Number formats"/><category term="MXFP4"/><category term="Quantisation"/><category term="Prior art"/></entry><entry><title>Gigabit Ethernet on Artix-7 without a vendor toolchain</title><link rel="alternate" type="text/html" href="https://t27.ai/blog/open-gigabit-ethernet-artix7/"/><id>https://t27.ai/blog/open-gigabit-ethernet-artix7/</id><updated>2026-08-09T00:00:00Z</updated><published>2026-08-09T00:00:00Z</published><summary type="text">RGMII at gigabit through Yosys, nextpnr-xilinx and Project X-Ray with no vendor tools — the six blockers that had to be patched, the one still open, and what the workaround costs.</summary><category term="FPGA"/><category term="Open source"/><category term="Ethernet"/><category term="Artix-7"/><category term="openXC7"/></entry></feed>
