
Machine Learning
BitCPM-CANN: 1.58-Bit Training Framework Opens New Path for Memory-Efficient AI on Domestic Hardware
5/26/2026

LLMs
Google's TurboQuant Compression Enables Faster LLM Inference on Modest Hardware
4/15/2026

Infrastructure
The Delicate Balance of Feature Growth and Memory Efficiency in libcurl
1/22/2026