GoKawiil
Tech News
Developer Nicholas Carlini built a custom port of the Stockfish chess engine running on a 200 MHz Intel Pentium Pro from November 1995, estimating its playing strength at roughly 3300 Elo versus Kasparov's near-2800 rating in 1996. He connected the vintage machine to the internet via a Realtek Ethernet card so the public can play against it directly.
An independent developer spent ten months building Chessbook-1, a 199-million-parameter chess model designed to mimic how real Lichess players move, think and manage time at any rating or time control. It now powers training games on Chessbook and was benchmarked against existing human-like engines Maia-3 and Allie 2.0 using September 2026 Lichess positions drawn after each model's training cutoff.
A developer created a set of Claude Code skills that convert a single chess game into a readable post-mortem, combining Stockfish engine analysis with the player's own recorded thoughts during the match. In a demonstration, Claude transcribed a French audio recording of the player's in-game reasoning via whisper.cpp, aligned it with move timestamps from the PGN, and produced a narrated video explaining specific decisions, such as bishop placement choices, verified against engine lines. All underlying files—transcript, engine analysis, annotated PGN, and storyboard—are shared in the project's example folder.
A developer built an open-source model that mimics TypeSafe's proprietary Jev system, which scores a variable list of text options in a single pass rather than generating answers word-by-word. The project, released without official documentation from TypeSafe, includes working demos where the same attention-based scoring head selects Doom combat moves and chess moves from image inputs, along with checkpoints, code and a rendering pipeline to visualize decision traces.
The Stockfish team has released version 19 of its open-source chess engine, claiming an Elo gain of up to 44 points over version 18 and a win rate more than three times its loss rate in test matches. The update introduces a new SFNNv16 neural network architecture, retires a secondary network added in version 16.1, and switches to universal binaries that auto-detect CPU features instead of requiring manual selection between AVX2, AVX-512 and similar builds. The release also adds native support for RISC-V and LoongArch processors.