apple-silicon

Ollama switches to llama.cpp—faster local inference

Ollama 0.30, Claude 4.6 context expansion, npm worm alert

MoE coding models, agent adoption surges to 59%

Cut token use by 98%, Apple Silicon, and LLMs