We shipped a stateful, Vulkan-trained decision model plus LRL — a small compiled language for chaining questions into branching decision programs — and scored the v6 checkpoint at 48.6% on a 2,540-rec...

The csg test level grew a fifth room, a second door, and a clock that finally ticks in the right decade. This week the BSP portal graph became the sound-propagation graph — room brushes are acoustic v...

Our local Jev-class decision model is mid-run on its biggest dataset yet — 517,687 gate-verified records spanning 66 task families. Here's the architecture, the family taxonomy, the measured hardware ...

A companion guide to Jia-Bin Huang's walkthrough of the Kimi team's Attention Residuals paper — replacing fixed residual accumulation with softmax attention over layer outputs — plus grounded blueprin...

A compact 4.8M parameter GRU-L3 character-level language model trained on a single RTX 3080 Ti via a custom Vulkan compute backend. Architecture, WSD-S scheduling research, a regression and recovery c...

A compact 4.8M parameter GRU-L3 character-level language model trained on a single RTX 3080 Ti via a custom Vulkan compute backend. Architecture, research findings, training telemetry, and the path fr...

A follow-along field guide to the JEPA video: 30 years of self-supervised learning from random-dot stereograms to action-conditioned world models, with fact-checked references, equations, and diagrams...

Authentication and onboarding for the C.A.T. operator ecosystem.

V-world network status and swarm telemetry.

A case study of Grim Fandango — and what a 1996 puzzle-dependency chart from a defunct adventure studio still has to teach a live AI-development industry about specification, verification, and design.

A compiler for agent instruction files: Markdown → PIR → compact Markdown, with a custom line-oriented intermediate language called PIR.

We scaled a custom Vulkan training pipeline from a toy benchmark to a real dual-stream GRU-L3 model. The harness now clears every incremental stage, but a long-sequence NaN and a stubborn CPU/GPU grad...

CatGame is a custom C++20 Vulkan engine and game project featuring behavioral intelligence research, Go RL training, VFX systems, and side-scroll gameplay.

How the dream graph evolved from a static knowledge map into a living 3D neural network — with pulse propagation, trace-mode telemetry, dictionary-driven engram compilation, and concept-as-memory toke...

SWE-bench scores don't tell you if a model can build a browser OS from scratch. Bijan Bowen's nine-task generative coding suite does. We formalize his methodology, prove it satisfies separability and ...

When Google poached Windsurf's CEO in a $2.4B acqui-hire, the 250-person team was left leaderless. Cognition called on Friday at 5PM. By Monday morning, they had a deal. Twelve months later, revenue i...

Most NPCs are scripts wearing a costume. The NPC Persona Builder treats character design as a behavioral engineering problem — combining Plutchik's emotion wheel, MBTI personality types, 55 cognitive ...

Every training sample carries a story — who made it, under what license, and how it was acquired. If we tag every sample with its provenance, we can audit what a model knows, pay the creators who taug...

AI companies are strip-mining the internet to feed their models. The research says they'll collapse under their own exhaust. The future belongs to small, bespoke models running on local NVIDIA hardwar...

A classic and influential work that laid the theoretical foundations for information theory and a timely text for contemporary informations theorists and p

The new tileset editor stores levels as a flat uint16_t grid. That's the same data layout modern GPU tilemap renderers use to draw 10,000 tiles in a single draw call. Here's how we're adapting that ar...

How six interlocking knowledge systems — built across four months in 2026 — created the behavioral substrate for the next generation of non-player character design.