Liquid AI uses Quantization-Aware Distillation to recover 97% of the accuracy lost in 4-bit quantization, enabling faster, high-quality model inference on edge devices.
A trio of high-profile open letters aired the deep rift in AI: Microsoft rallied for open weights and distillation, Anthropic warned of catastrophic misuse, and independent developers urged nuanced regulation that won’t crush innovation.
Hugging Face has released six Ettin reranker models of varying sizes, designed to significantly improve the accuracy of search and RAG systems at low cost through a 'retrieve-then-rerank' two-stage architecture.
Benchmarks show specialized document OCR keeps beating top GPT models on accuracy and cost; document parsing won't be swallowed by frontier models.