Article · MarkTechPost ·
qwenGPU compiles a 4-lane SIMT GPU into Qwen3.5-2B as 45 untrained layers. The model writes every step of real CUDA launches itself, using Gated DeltaNet state as registers and memory.
A tool or model opens in the Heidelberg AI Curriculum: what it does, what it costs, how to start. A repo opens on GitHub.
ArticleMarkTechPost
NVIDIA released Auto Ontology, an Apache 2.0 library that generates a semantic layer from enterprise metadata. It grounds agent SQL in certified definitions and serves context through MCP, CLI and API.
ArticleMarkTechPost
Nace AI open-sources Drex Max v1, a 27B multimodal decision model scoring 67.52 that runs on Apple Silicon.
ArticleMarkTechPost
DeepSeek released dsh-superagent, an MIT-licensed research prototype that turns GitHub and Feishu requests into parallel worker sessions. In DeepSeek’s own simulation, p95 task completion fell to 15.7 minutes from 29.5 minutes.
ArticleMarkTechPost
Factory released Droid 1.0, a model-independent coding agent harness that makes programmatic tool calling the default. On Factory’s own Terminal-Bench 4 runs, Droid scores 65.2% with Claude Opus 5.5 at $4.73 per passed task.